Published online before print April 28, 2009
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* CRIBI and Department of Biomedical Sciences, University of Padova, Italy; and
Novartis Vaccine and Diagnostics srl, Siena, Italy
2. Correspondence: C.R.I.B.I.-University of Padova, Via G. Colombo 3, 35121 Padova, Italy. E-mail: emanuele.papini{at}unipd.it
ABSTRACT
Hypervirulent MenB causing fatal human infections frequently display the oligomeric-coiled coil adhesin NadA, a 45-kDa intrinsic outer membrane protein implicated in binding to and invasion of respiratory epithelial cells. A recombinant soluble mutant lacking the 10-kDa COOH terminal membrane domain (NadA
351–405) also activates human monocytes/macrophages/DCs. As NadA is physiologically released during sepsis as part of OMVs, in this study, we tested the hypothesis that NadA+ OMVs have an enhanced or modified proinflammatory/proimmune action compared with NadA– OMVs. To do this we investigated the activity of purified free NadA351–405 and of OMVs from MenB and Escherichia coli strains, expressing or not full-length NadA. NadA351–405 stimulated monocytes and macrophages to secrete cytokines (IL-1β, TNF-
, IL-6, IL-12p40, IL-12p70, IL-10) and chemokines (IL-8, MIP-1, MCP-1, RANTES), and full-length NadA improved MenB OMV activity, preferentially on macrophages, and only increased cytokine release. NadA351–405 induced the lymphocyte costimulant CD80 in monocytes and macrophages, and NadA+ OMVs induced a wider set of molecules supporting antigen presentation (CD80, CD86, HLA-DR, and ICAM-1) more efficiently than NadA– OMVs only in macrophages. Moreover, membrane NadA effects, unlike NadA351–405 ones, were much less IFN-
-sensitive. The activity of NadA-positive E. coli OMVs was similar to that of control OMVs. NadA in MenB OMVs acted at adhesin concentrations 
106 times lower than those required to stimulate cells with free NadA351–405.
Key Words: chemokines • cytokines • cell activation • inflammation
Introduction
The gram– bacterium Neisseria meningitidis, found in the upper respiratory tract of a large part of nonsymptomatic people, can gain access to the bloodstream, where it proliferates, inducing fatal septic shock or meningitidis, especially in young patients [1
]. Infection with serogroup B strains, a principal cause of meningococcal casualties in developed countries [2
, 3
], could not be prevented with vaccines based on capsular polysaccharide-protein conjugates successfully applied to other serotypes, as their capsule is formed by the human self-antigen polysyalic acid [4
]. In addition, vaccines based on outer membrane preparations were strain-specific and therefore, offered limited protections. Consequently, a reverse vaccinology approach led to the identification of serogroup B new vaccine candidates, among which is the adhesin NadA [5
]. The rsNadA351–405, missing the outer membrane anchor sequence, is a good immunogen included in a five-component universal anti-meningococcal vaccine under trial at present [5
].
Functionally, NadA is an interesting molecule that likely contributes to the N. meningitidis parasite cycle and pathogenicity. In fact, the expression of this adhesin is linked to hypervirulence and has been proposed to improve epithelial cell colonization and invasion [6 ].
Interestingly, not only epithelial cells but also human DCs, monocytes, and macrophages express specific receptors for NadA and are activated by this meningococcal adhesin [7
, 8
]. In DCs, NadA351–405 up-regulates molecules involved in antigen presentation and the secretion of cytokines in an IFN--dependent way. In monocytes and macrophages, NadA351–405 stimulates cytokine and chemokine release with a certain tendency for chemokine production at the expenses of shock-related cytokines such as TNF- and IL-1. NadA is also a survival and differentiation signal for monocytes: After 3 days in the presence of the meningococcal adhesion, they show the characteristic of a special macrophage type up-regulating CD80, HLA-DR, the LPS coreceptor CD14, and the FcRIII IgG low-affinity receptor CD16. In conclusion, NadA is a good immunogen but may also exert a direct inflammatory and immunological stimulation in vivo.
Another approach to develop anti-meningococcal vaccines is based on the use of OMV preparations [9 ]. A limit of such vaccines is that they are effective against the same strains from which OMVs are obtained and therefore, fail to protect from other strains able to infect humans [9 ]. OMVs are released spontaneously and massively during meningococcal sepsis [10 ] and similarly to NadA, are immunogens active on APCs and other defensive cells. In a previous study, it was documented that OMVs from a meningococcal isolate used in a Norwegian vaccine induced the secretion of proinflammatory cytokines IL-8 and IL-10 in a whole blood model [11 ]. Similarly, OMVs used in a Cuban vaccine activated PMNs [10 ].
Originally, it was assumed that the proinflammatory activity of OMVs is a result of LPS, the most abundant and powerful gram-negative PAMP. However, experiments performed with a LPS-deficient N. meningitidis mutant and with mice hyporesponders to LPS suggested that other superficial components may be responsible for cytokine production and immune cell activation [12 , 13 ]. These observations opened the possibility of designing the composition of the outer membrane by genetic manipulation with the goal of improving their immune-modulatory activity.
In this study, we asked whether NadA, known to have an intrinsic activity on monocytes/macrophages/DCs when in a soluble form, can improve the stimulatory effects of N. meningitidis B outer membranes when present in these vesicles as an intrinsic membrane protein. This information may be relevant for improving OMV performance in vaccine formulations. Moreover, our study focuses on the actual significance of NadA as an agonist during host tissue infection by N. meningitidis, considering that the surface of the outer membrane is rich with other strong PAMPs, primarily of LPS.
The proimmune effects of OMVs isolated from NadA– and NadA+ Escherichia coli and MenB and of rsNadA351–405 were analyzed, quantifying the overexpression of the principal molecules involved in antigen-dependent lymphocyte stimulation by flow cytofluorimetry. In addition, the secretion of a large cytokine/chemokine/growth factor panel was measured using a Bio-Plex multiplex bead-based suspension array system.
MATERIALS AND METHODS
NadA351–405 preparation
rsNadA was designed and purified as described previously [14
]. Briefly, the DNA sequence of nadA allele 3, cloned from the hypervirulent MenB 2996, encoding the deletion mutant NadA351–405 with no membrane anchor, was cloned into a pET21b vector (Novagen, San Diego, CA, USA). The protein secreted in the extracellular medium of the transformed E. coli BL21(DE3)-NadA351–405 strain was purified by Q Sepharose XL and Phenyl Sepharose 6 Fast Flow (Pharmacia, Piscataway, NJ, USA) chromatography. LPS contamination [tested by a Limulus test kit from Sigma Chemical Co. (St. Louis, MO, USA)] was ablated to <0.005 EU/µg protein by a further passage on a hydroxyl apatite ceramic column (HA Macro Prep). No E. coli antigens were detected by Western immunoblot analysis, with a rabbit polyclonal antibody raised against whole E. coli cells (Dako, Denmark). Purified NadA351–405 shows a single 35-kDa band after SDS-PAGE and silver staining, consistent with the predicted molecular weight, and is a homotrimer, as assessed by light-scattering analysis [15
]. Aliquots of protein solution (2 mg/ml in PBS, pH 7.4) were frozen in liquid nitrogen and stored at –80°C.
Bacterial strains
The mutated gene coding for NadA351–405 was cloned into the pET21b vector (Novagen) as described [15
]; then, the plasmid was transformed into E. coli BL21 (DE3; Novagen), used as an expression host; its control carried the pET21b plasmid with no insert (E. coli-pET). Tettelin et al. [16
] described the MC58 strain; the MC58nadA– strain was prepared as described by Capecchi et al. [15
]. To complement the NadA null mutant, the nadA wild-type gene was reinserted under the control of the Ptac promoter into the chromosome of MC58nadA–between the converging ORFs NMB1428 and NMB1429, through transformation of the MC58nadA– strain with the pCOMnadA plasmid, which was generated as follows. The Ptac promoter and ribosome-binding site were amplified from plasmid pMMB206 [17
] using the Tac1/Pind-R primer pair and cloned as a 234-bp KpnI-NsiI fragment adjacent to the chloramphenicol cassette into the pSLComCmr plasmid [18
] consisting of a chloramphenicol-resistance gene flanked by upstream and downstream regions for integration between the NMB1428 and NMB1429 ORFs, generating a pCom-pseudo-random binary sequence. The nadA gene was amplified from the MC58 genome with the NadAF6-NdeI/NadAR6-NsiI primer pair cloned as a 1089-bp NdeI/NsiI fragment downstream of the Ptac promoter generating pCOMnadA. Transformants were selected on chloramphenicol, and resistant colonies were analyzed by PCR for correct insertion by a double-homologous recombination event. The complementing strain was named MC58nadA–_C.
E. coli OMV preparation
OMVs were prepared as decribed [15
]. Briefly, E. coli strains were cultured at 37°C in Luria-Bertani broth supplemented with 100 µg ml–1 ampicillin for 5 h; bacteria were harvested, suspended in 1 ml, 10 mM Hepes buffer (pH 7.4), and sonicated on ice. The cell membranes were recovered by centrifugation at 13,000 rpm at 4°C for 30 min in a microcentrifuge. Cytoplasmic membranes were resuspended in 200 µl, 10 mM Hepes, and an equal volume of 2% Sarkosyl in 10 mM Hepes (pH 7.4) was added; then, cytoplasmic membranes were incubated at room temperature for 30 min under agitation. The outer membranes were then recovered by centrifugation at 13,000 rpm, 30 min at 4°C, washed three times with 10 mM Hepes, and resuspended in the same buffer. Aliquots were frozen and conserved at –20°C. LPS-LOS content of our OMV preparations, quantified by Limulus tests, was 6.5 (SE±1.5) EU/ng OMV proteins, regardless of the expression of NadA.
N. meningitidis OMV preparation
Bacteria were grown overnight in GC plates at 37°C. They were then harvested, resuspended in PBS, and inactivated for 3 h at 56°C. Then, they were washed and resuspended in 20 mM Tris-HCl (pH 7.5) in the presence of protease inhibitors and sonicated (Branson Sonifier 450) on ice. Samples were centrifuged at 5000 g for 20 min at 4°C to eliminate the unbroken cells and inclusion bodies and then at 50,000 g for 75 min at 4°C to eliminate cellular debris, and cell membranes were washed with 20 mM Tris-HCl (pH 7.5) and 1 M NaCl 10% (v/v) glycerol and centrifuged at 50,000 g for 75 min at 4°C. Pellets were resuspended in 20 mM Tris-HCl (pH 7.5), 2% (v/v) Sarkosyl, for 20 min to solubilize cytoplasmic membranes. Samples were centrifuged at 50,000 g for 10 min at 4°C, and supernatants were centrifuged at 75,000 g for 75 min at 4°C. OMVs were washed with 20 mM Tris-HCl (pH 7.5), centrifuged at 75,000 g for 75 min at 4°C, and then resuspended in 20 mM Tris-HCl (pH 7.5), 10% (v/v) glycerol. Aliquots were frozen and conserved at –20°C.
LPS-LOS content of our OMV preparations, quantified by Limulus tests, was 4.9 (SE=±0.2) EU/ng OMV proteins, regardless of the expression of NadA.
Cell preparation and culture conditions
PBMC were isolated from buffy coats of healthy donors by centrifugation over a Ficoll-Hypaque (Amersham Corp., Arlington Heights, IL, USA) step gradient and suspended in RPMI 1640 (Gibco-BRL, Grand Island, NY, USA), supplemented with antibiotic and 10% FCS. Residual T and B cells were removed from the monocyte fraction by plastic adherence. The purity of preparations (percentage of CD14-positive cells) and cell viability (trypan blue exclusion test) were, respectively, higher than 98%. All cells were kept at 37°C in a humidified atmosphere containing 5% (v/v) CO2, unless specified otherwise. Macrophages were differentiated from monocytes by treating plastic adherent cells with RPMI-1640 medium without serum for 1 h and by a further incubation for 6 days in the same medium as above [19
].
Monocytes and differentiated macrophages were treated or not for 24 h with different stimuli—NadA (up to 3 µM), LPS 0.2 µg/ml, IFN- (1000 U/ml), E. coli pET BL21-NadA OMVs (E. coli NadA+), or E. coli pET BL21 OMVs (E. coli NadA–)—and with N. meningitis MC58 OMVs (MenB NadA+), N. meningitidis MC58nadA–_C OMVs (MenB NadA++), or N. meningitis MC58nadA– OMVs (MenB NadA–) in RPMI 1640 containing 10% FCS. After 24 h, cells were harvested and analyzed. Culture supernatants were collected frozen in liquid nitrogen and conserved at –80°C for cytokine analysis.
Flow cytometry analysis
Monocytes or macrophages were harvested, washed, and incubated in PBS containing 1% FCS and 0.1% NaN3 (FACS buffer) for 30 min at 4°C with different mAb, PE-conjugated, and directed to CD80 (B7.1), CD86 (B7.2), MHC class II (HLA-DR), and ICAM-1 (CD54; BioLegend, San Diego, CA, USA). After washing, propidium iodide was added to exclude dead cells, and cell fluorescence intensities of the gated populations were measured with the FACSCalibur flow cytometer and analyzed with EXPO CellQuest software. Data were collected on 10,000–20,000 events.
Bio-Plex multiplex cytokine assays
All antibody pairs used, directed against different noncompeting epitopes of a given cytokine, were purchased from BioRad (Hercules, CA, USA). Calibration curves from recombinant cytokine standard were prepared with fourfold dilution steps in RPMI-1640 medium containing 10% FCS. All assays were conducted in 96-well sterile prewetted filter plates at room temperature and protected from light. A mixture containing 5000 microspheres/cytokine was incubated together with standard or sample in a final volume of 50 µl for 30 min under continuous shaking (300 rpm). After three washes by vacuum filtration with Bio-Plex washing buffer, a mixture of biotinylated antibodies diluted in Bio-Plex detection antibody diluent was added (25 µl to each well). After a 30-min incubation and washing, streptavidin-PE diluted in Bio-Plex assay buffer was added (50 µl/well). At the end of a 10-min incubation under continuous shaking and after washing, the fluorescence bead intensity was measured in a final volume of 125 µl Bio-Plex assay buffer. Data analysis was done with Bio-Plex manager software using a five-parametric curve fitting. The detection limit of the assay for all antigens was 1 pg/ml.
SDS-PAGE and Western blotting
E. coli OMVs, MenB OMVs, and NadA351–-405 were resolved by SDS-PAGE, blotted onto a nitrocellulose membrane, and probed with a polyclonal anti-NadA antibody. Secondary antibody was the AP goat anti-rabbit IgG (Chemicon, Billerica, MA, USA). Blots were developed with AP buffer [100 mM NaCl, 5 mM MgCl2, 100 mM Tris/Cl (pH 9.2)], supplemented with 1% v/v 5-bromo-4-chloro-3'-indolyphosphate p-toluidine salt and 1% v/v nitro-blue tetrazolium chloride (Sigma Chemical Co.).
ELISA assay
The day before the experiment, the wells of a polyvinyl chloride microtiter plate (Nunc, Rochester, NY, USA) were coated with different amounts of E. coli or MenB OMVs (0.5, 1, and 2 µg) and purified NadA351–405 (1, 10, and 100 pg, 1, 10 and 100 ng). The day of the experiment, wells were washed, blocked with 1% BSA-PBS, and incubated for 1 h with anti-NadA antibodies; then, after extensive washing, the plate was incubated with HRP anti-rabbit IgG (Chemicon) and developed with 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) solution (Chemicon).
Preparation of polyclonal anti-NadA antibody
Purified NadA351–405 (50 µg) were used to immunize rabbits. The immunization was performed s.c. together with CFA for the first dose and IFA for the further three doses (Days 18, 34, and 52). Antisera were taken on Day 68, and IgG were purified using Protein A beads (Amersham Corp.).
RESULTS
In a previous study, NadA351–405 activity was tested on monocytes and monocyte-derived macrophages as a single agonist [8
]. Here, we also treated these cells with NadA351–405 in the presence of IFN- or LPS and of IFN- and LPS. The rationale of this protocol is aimed at revealing NadA activity in the presence of an immune stimulant-like IFN- [7
] and the possible synergic effect of costimulation with LPS, the most abundant PAMP of Gram– outer membranes implied in local tissue inflammation and systemic shock induced by N. meningitidis [4
].
Monocytes and macrophages were also treated in the same way, in the absence or in the presence of IFN-, with OMVs from an E. coli mutant strain (E. coli NadA+), expressing high quantities of NadA [14
]; the hypervirulent MenB MC58 (MenB NadA+), expressing moderate but significant levels of NadA [15
]; the genetically altered version of strain MC58 (MenB NadA++; see Materials and Methods), expressing higher NadA levels comparable with the hypervirulent 2996 strain. Controls were OMVs from isogenic E. coli strains (E. coli NadA–) and the nadA gene knockout N. meningitidis MC58 strain (MenB NadA–). Stimuli were normalized based on OMV protein concentrations.
Pattern of cytokine/chemokine secretion by monocytes and macrophages stimulated with sNadA351–405 and with OMVs from E. coli and N. meningitidis B-expressing or not NadA
Up to 27 analytes were then quantified in the extracellular media using a suspension array assay (Bio-Plex), with the exception of IL-23, which was quantified with a conventional ELISA test. To facilitate data presentation, data are grouped into three categories: The inflammatory cytokines, IL-1, IL-1β, TNF-, IL-6; the regulatory cytokines, IL-12p40, IL-12p70, IL-23, IL-10, GM-CSF, G-CSF, FGF-β, VEGF; the chemokines, IL-8, MCP-1, MIP-1, IP-10, RANTES, eotaxin. IL-2, IL-3, IL-4, IL-5, IL-7, IL-13, IL-15, and IL-17 were below the detection limits (not shown). In the presence of IFN- costimulation, no significant further production of this cytokine was measured (not shown).
Inflammatory cytokines.
As already reported [8
], NadA351–405 alone was scantly effective in inducing IL-1, IL-1β, and TNF- production by monocytes and macrophages (Fig. 1
). However, we now find that all of these cytokines were strongly induced by NadA351–405 in monocytes when IFN- was present. Macrophages costimulated with NadA351–405, and IFN- induced no IL-1, a slight but significant quantity of IL-1β, and a good TNF- secretion corresponding to 10% of that of monocytes. IL-6, a cytokine induced in inflammation but with systemic anti-inflammatory effects, was stimulated by NadA351–405 in monocytes and macrophages, regardless of IFN- costimulation. The effects of LPS were not increased further by NadA costimulation, with the exception of IL-6 production in monocytes, where the action of the two agonists was additive.
 View larger version (24K): [in a new window] |
Figure 1. Induction of IL-1, IL-1β, TNF-, and IL-6 by NadA351–405 in monocytes and macrophages. Human adherent monocytes or monocyte-derived macrophages were incubated for 24 h in RPMI plus 10% FCS, with or without different stimuli: NadA351–405, E. coli LPS 0.2 µg/ml, IFN- (1000 U/ml) as indicated. Cytokines were determined with a Bio-Plex suspension array in collected cell supernatants. Values (ng/ml) are the mean from three experiments run in duplicate ± SE. Asterisks indicate the conditions in which NadA induced a significant action (*, P 0.05) compared with controls not treated with the adhesin.
|
(IFN--dependent) and of IL-6 (IFN--independent) compared with E. coli NadA– OMVs (Fig. 2
), and macrophages were stimulated equally by E. coli OMVs, independently of NadA presence and IFN- costimulation.
 View larger version (39K): [in a new window] |
Figure 2. Induction of IL-1β, TNF-, and IL-6 by NadA expressing OMVs from E. coli and MenB in monocytes and macrophages. Cells were incubated for 24 h in RPMI plus 10% FCS with E. coli pET BL21-NadA OMVs (E. coli NadA+ OMVs), E. coli pET BL21OMVs (E .coli NadA– OMVs), N. meningitidis MC58 wild-type OMVs (MenB NadA+), N. meningitdiis MC58nadA–_C OMVs (MenB NadA++), and N. meningitidis MC58nadA– OMVs (MenB NadA–), with or without IFN- (1000 U/ml) as indicated. The amounts of the secreted cytokines were determined with a Bio-Plex suspension assay in culture supernatants. Values (ng/ml) are the mean from three experiments run in duplicate ± SE. Asterisks point to conditions in which NadA+ vesicles induced an effect significantly different (*, P0.05) from that induced by NadA– ones.
|
only in macrophages and the release of IL-6 in macrophages and monocytes in an IFN--independent way when compared with NadA-negative MenB OMVs. MenB NadA+ OMVs, expressing a lower level of adhesin, showed a stimulation pattern similar to that of MenB NadA++ OMVs but were less (TNF-, IL-6) or not (IL-1β) effective.
Regulatory cytokines.
IL-12p40, the component of IL-12p70 and IL-23, was produced by monocytes and macrophages only when costimulated with NadA351–405 and IFN-, although in macrophages, its amount was 10–20 times less intense than in monocytes (Fig. 3
).
 View larger version (27K): [in a new window] |
Figure 3. Induction of IL-12p40, IL-12p70, and IL-10, by NadA351–405 in monocytes and macrophages. Cells were incubated for 1 day in RPMI plus 10% FCS, with or without NadA351–405, E. coli LPS 0.2 µg/ml, and IFN- (1000 U/ml) as indicated. Values corresponding to cytokine concentrations in cell supernatants (ng/ml), obtained with a Bio-Plex immune suspension array, are the mean from three experiments run in duplicate ± SE. Asterisks indicate the conditions in which the adhesin effects are significant compared with controls (*, P0.05).
|
351–405 plus IFN- in a way resembling that of IL-12p40 in monocytes and macrophages.
IL-10, the Th2-differentiating anti-inflammatory cytokine antagonist of IL-12p70 effects, was induced by NadA351–405 alone in monocytes and in macrophages, as macrophage secretions are most efficient (about tenfold that of monocytes). In both cases, IL-10 production was down-regulated by IFN-.
LPS effect on these cytokines was not influenced by costimulation with NadA, regardless of IFN- costimulation.
IL-23, a cytokine with an activity, in part, resembling that of IL-12p70 involved in nervous system autoimmunity and the growth factors GM-CSF, G-CSF, FGF-β, and VEGF, was not up-modulated by NadA in any costimulation conditions (not shown).
In conclusion, IFN- costimulation synergized NadA351–405 induced IL-12 secretion but antagonized the secretion of IL-10 as a result of the same agonist in monocytes and macrophages. However, the expression of NadA in E. coli OMVs did not change their IL-12 stimulatory activity in monocytes and macrophages, independently of IFN- presence (Fig. 4
), and determined a stronger release of IL-10 in macrophages in the absence of IFN- costimulation. On the contrary, IL-12 and IL-10 were induced more efficiently by NadA-positive MenB OMVs in macrophages. This action was improved by IFN- but was already significant with no costimulation.
 View larger version (25K): [in a new window] |
Figure 4. Induction of IL-12p40, IL-12p70, and IL-10 by NadA expressing OMVs from E. coli and MenB in monocytes and macrophages. Cells were incubated for 24 h in RPMI plus 10% FCS with E. coli pET BL21-NadA OMVs (E. coli NadA+ OMVs) or E. coli pET BL21 OMVs (E. coli NadA– OMVs) and with N. meningitidis MC58 wild-type OMVs (MenB NadA+), N. meningitidis MC58nadA–_C OMVs (MenB NadA++), or N. meningitidis MC58nadA– OMVs (MenB NadA–), with or without IFN- (1000 U/ml) as indicated. Bio-Plex immune suspension arrays were used to quantify cytokines in the culture supernatants. Values (ng/ml) are the mean from three experiments run in duplicate ± SE. Asterisks indicate the panels corresponding to conditions in which the activity of NadA-positive OMVs was significantly higher (*, P0.05) than that of NadA negative ones.
|
351–405 was a good chemokine inducer (Fig. 5
). In fact, PMN chemoattractant IL-8 was stimulated by NadA351–405 alone in monocytes (at very high levels) and macrophages (20 times less). Interestingly, these effects were inhibited almost totally by IFN- costimulation and down-regulated by LPS costimulation. Similarly, NadA351–405 induced a MIP-1 secretion by monocytes and macrophages partially inhibited by IFN-. In monocytes, LPS slightly synergizes the NadA effect, and the presence of IFN- had no inhibitory effect on this synergy. In macrophages, LPS determines, per se, a strong MIP-1 secretion, which was moderately increased further by NadA costimulation. IFN- inhibited LPS-induced effects, and NadA351–-405 determined no modulation in these conditions.
 View larger version (31K): [in a new window] |
Figure 5. Induction of IL-8, MIP-1, MCP-1, and RANTES by NadA351–405 in monocytes and macrophages. Cells were incubated for 24 h in RPMI plus 10% FCS with NadA351–405, E. coli LPS 0.2 µg/ml, and IFN- (1000 U/ml) as indicated. Chemokines were determined with Bio-Plex multiplex cytokine assay in culture supernatants. Values (ng/ml) are the mean from three experiments run in duplicate ± SE. Asterisks highlight the conditions in which the NadA effect is positive with respect to controls (*, P0.05).
|
351–405 in monocytes and macrophages, but IFN- had no inhibitory effect or a slight synergic action, respectively. LPS increases the secretion of MCP-1, and its effect seemed additive with that of NadA. IFN- abrogated NadA effects also in the presence of LPS costimulation.
As expected, IP-10 was strongly induced by IFN- in monocytes and macrophages. NadA inhibited IP-10 production partially by both cell types. The adhesin did not affect IP-10 secretion as a result of IFN-/LPS costimulation (not shown). Eotaxin was not induced by free adhesin (not shown).
In spite of strong IFN--independent chemokine production by monocytes and macrophages stimulated with sNadA, no significant activity was measured between NadA+ and NadA– OMVs from E. coli and N. meningitidis (not shown).
Expression of molecules related to antigen presentation on monocytes and macrophages stimulated with sNadA351–405 and with OMVs from E. coli and N. meningitidis B expressing NadA or not
The increased secretion by monocytes/macrophages of soluble mediators such as cytokines, chemokines, and growth factors, induced by the different forms of NadA studied here, is supposed to have a regulatory effect on important innate-defensive reactions such as, for example, inflammation or phagocyte recruitment. However, adaptive immune responses strictly require the induction of the main proteins responsible for antigen presentation to lymphocytes. Therefore, human monocytes and macrophages, stimulated as indicated in the previous paragraph with soluble or OMV-associated NadA, were analyzed after 24 h for their plasma membrane expression of HLA-DR, CD80, CD86, ICAM-1, CD14, and CD16. Flow cytofluorimetric data (Fig. 6
) show that costimulation of monocytes and macrophages with NadA351–405 and IFN- resulted in the selective increase of CD80, and no activity was detected when the single agonist was administered separately to cells. The intensity of CD80 in monocytes was about three times lower than the one induced in macrophages. In monocytes, the LPS stimulus was not synergized by the contemporaneous presence of NadA and IFN-, and in macrophages, the triple stimulation with NadA, LPS, and IFN- was more efficient than the stimulation by LPS in inducing CD80. NadA351–405 did not induce CD86 in monocytes and did not increase its basal expression in macrophages, regardless of any costimulation. Similarly, also, HLA-DR and ICAM-1 (not shown) levels were not modulated by NadA305–450 in any condition.
 View larger version (28K): [in a new window] |
Figure 6. Effect of NadA351–405 on CD80, CD86, and HLA-DR expression in monocytes and macrophages. Adherent human monocytes or differentiated macrophages were cultured in RPMI plus 10% FCS with NadA351–405, E. coli LPS, and IFN-, as indicated. After 24 h, cells were collected, and the indicated markers were stained with PE-labeled anti-CD-specific antibodies and analyzed by FACS. Data are expressed as MFI and are the mean ± SE of four experiments run in triplicate. Significant values with respect to control cells (not stimulated with the adhesion) are indicated by asterisks (*, P0.05).
|
, only after prolonged incubation with NadA during cell differentiation [8
]. These new data are consistent and suggest that IFN- accelerates CD80 expression. No NadA effects were seen on CD14 and CD16 (not shown).
Figure 7
shows the quantification of CD-marker expression in monocytes and macrophages after OMV stimulation. NadA– and NadA+ E. coli OMVs increased the basal expression of CD80 in monocytes and macrophages, independently of IFN-, as the action on macrophages was more intense. On the contrary, MenB NadA+ and MenB NadA++ OMVs determined an additional expression of CD80 in macrophages only, compared with MenB NadA– OMVs. Again, this effect was IFN--independent.
 View larger version (30K): [in a new window] |
Figure 7. Effect of NadA expressing OMVs from E. coli and MenB on the expression of antigen presentation-related CD markers in monocytes and macrophages. Cells were treated with E. coli pET BL21-NadA OMVs (E. coli NadA+ OMVs) or E. coli pET BL21 OMVs (E. coli NadA– OMVs), N. meningitidis MC58 wild-type OMVs (MenB NadA+), N. meningitidis MC58nadA–_C OMVs (MenB NadA++), or N. meningitidis MC58nadA– OMVs (MenB NadA–) as indicated. Cells were costimulated with (open bars) or without (solid bars) IFN- (1000 U/ml) in RPMI plus 10% FCS. After 24 h, the plasma membrane amounts of indicated CDs were quantified with PE-labeled anti-CD-specific antibodies and analyzed by FACS. Data, expressed as MFI, are the mean of four experiments run in triplicate. Bars are ± SE. Asterisks indicate the effects induced by NadA-positive OMVs significantly different (*, P0.05) from corresponding NadA-negative vesicles. ctr, Control.
|
but still evident also in the absence of this costimulant.
The basal expression of HLA-DR in monocytes, in the absence of IFN- costimulation, was increased by E. coli OMVs and not affected by MenB OMVs, independently of NadA expression. In the presence of IFN-, HLA-DR levels reached a maximal value in monocytes, which was not altered by E. coli OMVs, and reduced substantially by MenB OMVs, regardless of the expression of NadA. In macrophages, NadA+ and NadA– E. coli OMVs increase the basal HLA-DR expression, and only MenB NadA++ OMVs augmented HLA-DR significantly compared with MenB NadA– in a IFN--independent way.
In monocytes, ICAM-1 expression was not altered significantly by any OMV type in any stimulation condition. In macrophages, however, E. coli OMVs strongly increased ICAM-1 expression, independent of NadA presence or IFN- costimulation. MenB NadA+ and MenB NadA++ OMVs determined a higher ICAM-1 expression compared with MenB NadA– OMVs, independent of IFN- costimulation.
An overall view to data represented in Figure 7
evidences some trends: E. coli OMV preparations are, per se, more stimulatory than MenB OMVs; the presence of NadA in OMVs is irrelevant in the case of E. coli and responsible for a dose-dependent activation increase in the case of MenB membranes; a NadA-dependent effect is limited to macrophages and is largely IFN--independent.
Although this study is focused on cells involved in the direct engagement with N. meningitidis cells in the blood and tissue, we also tested the differential effect of NadA-positive and NadA-negative OMVs on Mo-DCs, the main APC responsible for the initiation of the immune response. In these cells, MenB NadA++ OMVs induced a 25% (SE=±4; n=3) higher increase of CD86 and a 30% higher increase of ICAM-1 (SE=±6; n= 3) compared with MenB NadA– OMVs independently on IFN- costimulation, and NadA– and NadA+ E. coli OMVs were, on the contrary, equally effective (data not shown). Mo-DCs appear therefore sensitive to the presence of NadA in MenB OMVs but significantly less than macrophages in which CD86 and ICAM-1 reached values that are doubled with respect to NadA-negative OMVs.
NadA expressed in meningococcal B outer membranes is much more effective than NadA351–405
The NadA content of MenB and E. coli OMVs was estimated semiquantitatively by Western blot and quantitatively by an ELISA assay. NadA expressed in the bacterial membranes migrates as a single 150-kDa band, corresponding to the SDS-resistant trimer (Fig. 8A
). On the contrary, NadA351–405 migrates as a 35-kDa monomer, as the resistance to SDS of its trimeric organization, although present in the folded state, is lost as a result of the removal of the membrane anchor. The NadA signal present in MenB NadA+ OMVs was weak but visible, and in MenB NadA++ OMVs, it was more intense. In E. coli, NadA+ OMV NadA appeared very expressed. This order of NadA expression (E. coli NadA+>MenB NadA++>MenB NadA+) was confirmed quantitatively by an ELISA assay, in which purified NadA351–405 was used as a standard. In fact, we determined the following NadA content normalized for total OMV protein concentration (±SE; n=5) of the outer membrane preparations used: E. coli NadA+ 
1050 ± 210 pg/µg protein; MenB NadA++ 135 ± 50 pg/µg protein; MenB NadA+ 10 ± 5 pg/µg protein (not shown).
 View larger version (20K): [in a new window] |
Figure 8. Increased stimulatory efficiency of OMVs as a result of NadA expression as a function of adhesin concentration. (A) Indicated amounts of OMVs (based on protein concentration), isolated from positive or negative strains, and of purified NadA351–405 were separated by SDS-PAGE, and NadA immune-reactivity was determined by Western blot using specific rabbit anti-NadA antibodies and secondary AP-conjugated anti-rabbit IgG antibodies, following transfer on nitrocellulose. (B) Adherent monocyte-derived macrophages were stimulated for 24 h in the presence ( ) or absence (•) of IFN- (1000 U/ml) with different concentrations of NadA351–405 and of OMVs from NadA-positive and -negative E. coli and MC58. Values of plasma membrane CD80 expression measured by FACS and IL-12p70/TNF- secretion determined by ELISA in cells stimulated with NadA– OMVs were subtracted from values obtained in corresponding NadA+ OMV-stimulated cells. Data are expressed as a function of NadA present in the assay at a given membrane concentration. NadA OMV content was determined separately by ELISA using NadA351–405 as standard (see text). Data are from a representative experiment from three.
|
351–405) or as part of OMVs, we stimulated macrophages with different concentrations of the isolated, truncated adhesin and of purified E. coli and MenB OMVs. The plot reported in Figure 8B
shows the increment of TNF- and IL-12p70 secretion and of CD80 expression with respect to controls (cells not treated with the adhesin in the case of sNadA stimulation) or with respect to E. coli NadA– and MenB NadA– OMVs (in the case of OMV stimulation), plus or minus IFN-. Effects were expressed as a function of NadA concentration in the assay. The improvement of cell activation was detectable at MenB OMV doses corresponding to adhesin concentrations as low as 30 pg/ml, and sNadA351–405 activity begins to be detectable at doses of 10–20 µg/ml and only in the presence of IFN-. Data confirmed that NadA did not induce an extra effect when expressed in E. coli OMVs, despite ten- to 100-fold higher expression compared with MenB OMVs.
To test whether the higher effect of NadA+ OMVs was a result of a parallel stimulation by meningo-specific PAMPs present in the outer membranes, we costimulated macrophages with NadA351–405 (3 µM) and with adhesion-negative Men B and E. coli OMVs (up to 5 µg/ml) and analyzed CD activation markers (CD80, CD86, HLA-DR, and ICAM-1): In no case did we observe a synergic effect (not shown).
DISCUSSION
NadA and OMVs are immune-modulines, in principle, able to influence inflammatory and immune cell responses during N. meningitidis tissue invasion but also good immunogens suitable for vaccine formulations. In this study, we tested the biological activity emerging from the contemporaneous presence of NadA and other outer membrane agonists present in OMVs. This was done using a virulent MenB (MC58), expressing relatively low but detectable levels of NadA and a genetically modified isogenic mutant expressing the same adhesin to higher levels (approximately tenfold compared with MC58 wild-type), similar to those observed in other hypervirulent strains such as 2996. In addition, we used E. coli OMVs from a transformed strain, strongly expressing NadA (100-fold compared with MC58 wild-type). The activities of these preparations were compared with those of OMVs from parental E. coli NadA– strain pET and with OMVs from a MC58 strain subjected to nadA gene knockout. The activity of free rsNadA351–405 was measured in parallel.
We decided to analyze a large panel (33 analytes) of soluble or membrane-expressed markers of cell activation in various costimulation conditions with the main immune and microbial-derived monocyte/macrophage agonists (IFN- and LPS). We followed this strategy, as it is now clear that different agonists may induce different activation patterns, distinguishable as a whole but likely to remain unobserved when limited parameters are measured.
A picture of unexpected activation properties emerged when comparing membrane-associated with free NadA. The analysis of the expression of membrane proteins necessary for antigen presentation and of the secretion of cytokines demonstrates that membrane NadA is optimally effective in activating human macrophages and less effective on monocytes. This effect is, however, seen only if the adhesin is expressed in the N. meningitidis outer membrane. In fact, high expression of NadA in the E. coli outer membranes is with scarce or no effect on any cell type or in the same case, results in an extra activity, which is IFN--dependent.
In agreement with previous data, we confirmed and extended the observation that purified sNadA activates monocytes and macrophages. The cytokine/chemokine pattern induced in these cells by free NadA351–405 differs sensibly from that induced by membrane full-length NadA when expressed in N. meningitidis OMVs. sNadA effects are regulated positively or negatively by IFN-, and the effects of NadA expressed in N. meningitidis OMVs are only partially or not at all dependent of IFN-.
Our observations suggest caution in extrapolating a physiological scenario from information obtained with single purified agonists, as the combination of several different signals, as this study demonstrates, may generate quite different results concerning activation patterns and cell specificity.
On the other hand, our data also suggest that NadA density on the outer membrane is one of the determinants of MenB efficacy in the interaction with the immune system. For example, the level of the adjuvant cytokine IL-12p70 is improved with respect to MenB NadA– OMVs by a factor of approximately two by MenB NadA+ OMVs and by a factor of approximately four by MenB NadA++ OMVs. Consistently, the main molecules involved in antigen presentation and interaction with T lymphocytes are up-regulated more intensely by MenB NadA++ OMVs compared with MenB NadA+ OMVs.
Such a difference in stimulatory activity cannot be ascribed to a differential endotoxin presence, as LPS and LOS amounts were comparable in adhesion-positive OMVs compared with adhesion-negative ones (see Materials and Methods). Moreover, the increased effect of NadA-bearing OMVs is unlikely the result of a more efficient cell association of these membranes mediated by the adhesin itself. In this case, in fact, we should have observed a general increase of the production of all cytokines and chemokines induced by control OMVs in monocytes and macrophages. Experimental evidences, on the contrary, show that only some mediators are induced selectively to higher levels and in macrophages, prevalently, while others are not modified. This is consistent with the observation that high expression of NadA in E. coli did not increase bacterial adhesion to monocytes [8 ].
Strikingly, the positive effects of membrane NadA were observed at adhesin doses that are much lower than the ones necessary when sNadA was used. Cells stimulated with NadA-positive MenB OMV preparations were sensitive to pmolar adhesin concentrations, and NadA351–405 was active in the µmolar range and requires in most cases, IFN- costimulation.
One possible explanation for such a difference between isolated NadA and membrane NadA is the higher avidity of the adhesin array, which is formed, thanks to a membrane anchor, an organization that has been well-documented for all oligomeric coiled-coil adhesins [20 ]. Stimulation of clustered NadA receptors on target cells may well account for stronger efficacy and also for the difference in effects, as a result of a difference in intracellular signaling.
However, increased avidity and NadA receptor clustering are not enough to explain the peculiar effects of membrane NadA, as the expression of the adhesin in E. coli outer membranes is with scarce effect. It is therefore possible that specific costimuli present in the N. meningitidis outer membrane, as for example, porin A, modifies the outcome of the biological effects of membrane NadA, compared with the action on the same cells by isolated NadA.
However, our data also show that the treatment of macrophages with NadA– meningococcal B OMVs, supposed to contain these specific PAMPs, did not synergize the activity of free NadA351–405. We conclude that the adhesin and other agonists on Men B OMVs must be organized in the same membrane structure to exert their synergic effect. The presence of these enhancing conditions, clustering and costimulation by meningo-specific PAMPs, appears therefore necessary and may also explain the IFN- independence of the effects induced by membrane NadA, compared with free NadA molecules.
Indeed, the cytokine/chemokine balance induced by NadA associated to OMVs is more similar to that of NadA351–405 plus IFN- (high cytokines/low chemokines) than to that of NadA351–405 alone (low cytokines/high chemokines). On the other hand, the fact that in most cases, NadA expression did not improve E. coli OMV stimulatory activity, and LPS effects are not enhanced further by NadA simultaneous treatment rules out the possibility that NadA acts synergistically with the principal Gram– endotoxin and also with LOS, the similar variant of N. meningitidis. This again is in agreement with the lack of synergy between adhesion-negative OMVs from E. coli and MenB and NadA 351–405.
The main unexpected finding of this study is that the expression of NadA on N. meningitidis OMVs turned out to selectively enhance macrophage functions, and it is far less relevant for monocyte functions, which are normally influenced by adhesion-negative outer membranes.
This suggests that NadA expression is with no effect on the intrinsic efficacy of OMVs to stimulate shock-related cytokines by circulating monocytes but on the contrary, may improve antigen-presenting activity of tissue macrophages. The up-modulation of molecules necessary for antigen presentation by NadA in OMVs is predicted to enhance macrophage N. meningitidis phagocytosis and killing by favoring the adjuvant action of CD4+ T Th1 lymphocytes. Our data reinforce the hypothesis that NadA, when inside the body as part of OMVs, can be recognized by innate immune cells, such as macrophages. Finally, our experiments suggest that full-length wild-type NadA may increase the efficacy of vaccine formulations based on N. meningitidis outer membranes.
ACKNOWLEDGMENTS
This research was supported by ex 60% 2008 funds, by Prin 2003, and by Progetto di Ateneo 2004 of the University of Padova (Italy). We thank the Centro Trasfusionale of the Hospital of Padova (Italy; ULSS 16) for providing buffy coats from human donors and Dr. M. Morandi and Dr. E. Ciccopiedi (Novartis Vaccine and Diagnostics, Siena, Italy) for NadA351–405 purification. We are also especially indebted to Dr. M. De Bernard (Department of Biology, University of Padova) for helpful suggestions and critical reading of this manuscript.
FOOTNOTES
Abbreviations: AP=alkaline phosphatase, DC=dendritic cell, EU=endotoxin unit, FGF-β=fibroblast growth factor β, IP-10=IFN-inducible protein 10, LOS=lipooligosaccharide, MenB=Neisseria meningitidis B strain, MFI=mean fluorescence intensity, Mo-DC=monocyte-derived DC, NadA=Neisseria meningitides adhesin A, OMV=outer membrane vesicle, ORF=open-reading frame, PAMP=pathogen-associated molecular pattern, PMN=polymorphonuclear neutrophil, s=soluble, VEGF=vascular endothelial growth factor
1. These authors contributed equally to this work. 
Received January 16, 2009; revised March 9, 2009; accepted March 12, 2009.
REFERENCES
and R-848 dependent activation of human monocyte-derived dendritic cells by Neisseria meningitidis adhesin A J. Immunol. 179,3904-3916, interleukin-1β (IL-1β), IL-8, macrophage inflammatory protein 1 (MIP-1), MIP-1β, and interferon-inducible protein 10 by human neutrophils stimulated with group B meningococcal outer membrane vesicles Infect. Immun. 68,6917-6923This article has been cited by other articles:
 |
|
 |
|
  T. N. Ellis and M. J. Kuehn Virulence and Immunomodulatory Roles of Bacterial Outer Membrane Vesicles Microbiol. Mol. Biol. Rev., March 1, 2010; 74(1): 81 - 94. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
