Neisserial Outer Membrane Vesicles Bind the Coinhibitory Receptor Carcinoembryonic Antigen-Related Cellular Adhesion Molecule 1 and Suppress CD4⁺ T Lymphocyte Function

Cell lines and tissue culture techniques.Primary human CD4⁺ T cells were negatively selected from peripheral blood mononuclear cells by using the EasySep immunomagnetic-based cell purification system (Stemcell Technologies, Vancouver, British Columbia). Purified primary lymphocytes and the Jurkat CD4⁺ human T-lymphocyte line (ATCC CRL-10915) were both maintained in RPMI 1640 medium (Invitrogen Life Technologies, Burlington, Ontario) supplemented with 10% heat-inactivated fetal bovine serum and 4 mM GlutaMAX (Invitrogen), referred to as RPMI-G. Cells were cultured at 37°C in 5% CO₂ humidified air. Where indicated, the CD4⁺ T cells were stimulated with the indicated concentrations of recombinant human interleukin-2 (IL-2; BD Pharmingen, Mississauga, Ontario) for 48 h prior to antibody or OMV challenge. COS-7 African green monkey kidney cells (ATCC CLR-1651) were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (Cansera; Etobicoke, Ontario), 100 units/ml penicillin-streptomycin, 1 mM GlutaMAX, 1 mM sodium pyruvate, and 1 mM nonessential amino acids (Invitrogen). Cells were cultured as described above.

Bacterial strains. N. meningitidis strain K454 (26) and N. lactamica Y92 1009 (34) were obtained from the Meningococcal Reference Unit, Manchester, United Kingdom. Neisseria gonorrhoeae MS11-derived strains constitutively expressing defined Opa variants (25) were generously provided by T. F. Meyer (Max-Planck-Institut für Infektionsbiologie, Berlin, Germany). Opa variants were expressed in the background of MS11 strain N279, which does not express pilin and has a deletion in the opaC₃₀ locus encoding this strain's only heparan sulfate proteoglycan (HSPG) receptor-specific Opa variant (25). N. meningitidis and N. lactamica were grown from frozen stocks on brain heart infusion agar plus 1% (vol/vol) horse serum (Difco Laboratories, West Molesey, Surrey, United Kingdom), and N. gonorrhoeae strains were grown from frozen stocks on GC agar (Difco, Oakville, Ontario) supplemented with 1% (vol/vol) IsoVitaleX enrichment (BBL; Becton Dickinson, Cockeysville, MD). All bacterial strains were cultured at 37°C in 5% CO₂ humidified air, and gonococcal strains were subcultured daily, using a binocular microscope to maintain the desired opacity phenotype. Opa expression and variant type were routinely confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10%), with resolved proteins either stained using Coomassie brilliant blue or subjected to immunoblot analysis using the Opa cross-specific monoclonal antibody 4B12/C11 (2), which was generously provided by M. Achtman (Max-Planck-Insitut für Infektionsbiologie, Berlin, Germany).

Preparation and physical characterization of OMVs.OMVs were prepared from N. meningitidis and N. lactamica isolates essentially as described previously (41). Briefly, overnight liquid cultures were prepared in Franz medium, and bacteria were pelleted by centrifugation at 1,000 × g. Bacterial pellets were resuspended in OMV buffer containing 0.15 M NaCl, 0.05 M Tris-HCl, and 0.01 M EDTA (pH 7.5) and then heated to 56°C for 30 min before being sheared in a Wearing blender for 3 min. The resultant extract was centrifuged at 25,000 × g for 20 min, and the supernatant was retained. The pellet was washed, and the supernatant was again retained. These supernatants were pooled and clarified by centrifugation at 25,000 × g for 20 min and then were centrifuged at 100,000 × g for 2 h. The final OMV-containing pellet was washed twice, resuspended in phosphate-buffered saline (PBS), and stored at −80°C.

Gonococcal OMVs were prepared from recombinant gonococcal strains (25) with defined Opa phenotypes and receptor specificities (19). Bacteria were passaged overnight on solid medium (as described above). Near-stationary-phase liquid cultures were prepared by resuspension in modified brain heart infusion (Difco) containing 10 mM LiCl, 1 mM MgCl₂, 2 mM CaCl₂, 50 mM HEPES, and 1% (wt/vol) d-glucose (pH 7.2) and then incubation at 37°C in 5% CO₂ humidified air with rapid shaking. Thereafter, the cultures were incubated for a further 2 h at 40°C with rapid shaking. Bacteria were removed by centrifugation at 1,000 × g for 20 min and resuspended in PBS containing 0.05% (wt/vol) sarkosyl (Bioshop Canada, Inc., Burlington, Ontario) and 0.05% (wt/vol) sodium deoxycholate (Bioshop). Resuspended cells were incubated at 56°C for 30 min with gentle mixing and then chilled on ice. These bacterial suspensions were extracted using a Wheaton homogenizer and then sonicated on ice (five 10-s pulses). Extracts were clarified by centrifugation at 25,000 × g for 20 min, and the resulting supernatant was centrifuged at 100,000 × g for 2 h. The final pellet, which contained OMVs, was washed twice, resuspended in PBS, extruded through a 0.22-μm syringe filter, and then stored at −80°C. OMVs isolated using this protocol have been analyzed by electron microscopy and negative staining (Nl-OMVs and N. meningitidis OMVs [Nm-OMVs]) or by comparative flow cytometric analysis (N. gonorrhoeae OMVs), with no discernible difference in size distribution evident between strains. In some experiments, OMVs were labeled using fluorescein isothiocyanate (FITC; Sigma) according to the manufacturer's specifications.

Construction and expression of CEACAM1-Fc fusion proteins.An expression vector encoding the extracellular domains of CEACAM1 fused to the Fc portion of human immunoglobulin G1 (IgG1) was previously described (50). Recombinant CEACAM1-Fc protein was expressed in COS-7 cells following transient transfection using FuGENE6 reagent (Roche Molecular Biochemicals, Indianapolis, IN) according to the manufacturer's specifications. Cell culture supernatant was harvested 48 to 72 h after transfection and was clarified by centrifugation at 1,000 × g for 20 min at 4°C. The clarified supernatant was filtered using a vacuum-driven disposable filtration system (Stericup 0.22 μm; Millipore, Nepean, Ontario) and concentrated using a 10-kDa-cutoff polyethersulfone ultrafiltration concentrator (Millipore). The fusion protein was then purified by binding to protein A-Sepharose (Sigma) and the bound protein eluted using 0.2 M glycine-HCl (pH 2.5), with aliquots recovered directly into collection tubes containing 100 μl 1 M Tris (pH 9.0) to neutralize the samples. Purified eluate was dialyzed against PBS at 4°C and then concentrated to less than 1 ml with Ultrafree Biomax centrifugal filters (Millipore). The appropriate functions of the resulting CEACAM1-Fc preparations were confirmed by their specific binding to isogenic gonococcal strains expressing defined Opa protein variants (data not shown).

Determination of Opa binding function.Interactions between OMV preparations containing defined Opa variants and CEACAM1 were quantified by an enzyme-lined immunosorbent assay (ELISA). Initially, the protein content of each OMV preparation was determined using the bicinchoninic acid assay system (Pierce Chemical Company, Rockford, IL), and samples containing equal amounts of total protein were immobilized on 96-well microtiter plates (Corning Corporation, Midland, MI). Each OMV was applied in triplicate serial doubling dilutions and then exposed to a standard concentration of the CEACAM1-Fc fusion protein. Bound CEACAM1-Fc was quantified using protein A-conjugated horseradish peroxidase and the o-phenylenediamine dihydrochloride colorimetric system (Sigma), with spectrophotometric analysis at 450 nm. A Student t test analysis was performed on the data to determine the statistical significance of differences in CEACAM1-Fc binding by different OMV preparations.

Flow cytometric analyses.Association between FITC-labeled OMVs and IL-2 prestimulated lymphocytes was assessed by flow cytometric analysis. Where indicated, T-cell activation was assessed by quantifying the expression of the well-characterized T-cell activation marker CD69 by using the allophycocyanin-conjugated monoclonal antibody FN50 (Pharmacia, Mississauga, Ontario) 16 h after exposure to the activating stimuli. Cell viability was monitored by FITC-annexin V (Becton Dickinson, Oakville, Ontario). Cells were then fixed in paraformaldehyde (3.7% [wt/vol]) prior to antibody staining. A minimum of 1 × 10⁵ gated cells from each sample were analyzed using a FACSCalibur flow cytometer with CellQuest software (Becton Dickinson, Oakville, Ontario).

Analysis of lymphocyte proliferation.The CD4⁺ T lymphocytes were routinely cultured in the presence of 1,000 units/ml of recombinant human IL-2 (BD Pharmingen) to induce CEACAM1 expression. Primary CD4⁺ T cells were further activated by coligation of the T-cell receptor (CD3) and CD28, using 1 μg/ml mouse anti-human CD3ε-specific monoclonal antibody (clone UCHT1; BD Pharmingen) and 1 μg/ml mouse anti-human CD28 monoclonal antibody (clone CD28.2; BD Pharmingen) followed by 3 μg/ml Fab₂ fragments of goat anti-mouse IgG (Jackson Immunoresearch Laboratories, West Grove, PA). Jurkat cell stimulation via the T-cell receptor was induced by exposure to the CD3ε-specific monoclonal antibody UCHT1 followed by the goat anti-mouse IgG cross-linker. Parallel, density-matched suspensions containing 3 × 10⁵ CD4⁺ T cells, as assessed by direct counting with a Levy double hemocytometer, were treated in RPMI-G with either CEACAM-specific antibody (anti-CEA; DAKO Diagnostics, Mississauga, Ontario), a nonreactive control antibody (DAKO Diagnostics), or OMVs at the indicated concentrations. Two hours after the commencement of infection, gentamicin (Bioshop, Burlington, Ontario) was added to obtain a final concentration of 100 μg/ml to prevent bacterial overgrowth. CD4⁺ T-cell culture densities were determined by direct counting using a Levy double hemocytometer at the indicated times. In each instance, proliferation was assessed using a standardized counting pattern and no fewer than six fields, each containing 10 to 100 cells. A Student t test analysis was performed on the data to determine if statistically significant differences exist between Nl-OMVs and Nm-OMVs or isotype antibody and anti-CEACAM1 antibody on T-cell proliferation for each time point. All experiments were performed on a minimum of three separate occasions.

Opa protein function in the context of OMVs.Pathogenic (N. meningitidis) (Fig. 1A) and commensal (N. lactamica) (Fig. 1B) Neisseria spp. naturally overproduce and evaginate the outer membrane in the form of vesicles (OMVs) or “blebs” (32, 37). To test whether Opa proteins within OMVs retain their CEACAM-binding function, experimental vaccines were prepared. These typically contained vesicles of 100 ± 20 nm in diameter (data not shown), and consistent with previous work (33), the presence of Opa proteins reflected that of the outer membranes of the bacteria from which they were derived (Fig. 1C). To specifically ascribe differences in OMV binding function rather than other strain-related differences to expressed Opa variants, control OMVs that express the well-defined Opa₅₀ (which binds HSPG) (Opa_HSPG) or Opa₅₂ (which binds CEACAM1) (Opa_CCM) protein or no Opa protein (Opa⁻) were prepared from isogenic strains of N. gonorrhoeae (19).

• Open in new tab
• Download powerpoint

FIG. 1.

Expression and binding pattern of Opa proteins found on neisserial OMVs. A transmission electron micrograph illustrating intact diplococci and isolated OMVs is shown. N. meningitidis (A) and N. lactamica (B) each have closely associated naturally occurring membrane “blebs” (filled arrows) as well as others liberated from the bacteria. (C) Immunoblot probed using Opa protein-specific monoclonal antibody illustrating the presence of an Opa variant in Nm-OMVs but not in comparable Nl-OMVs as well as in OMVs derived from isogenic N. gonorrhoeae strains expressing either the CEACAM receptor-specific Opa (Opa_CCM) or the HSPG receptor-specific Opa (Opa_HSPG) protein but not in the strain expressing no Opa protein [Opa(−)]. (D) ELISA quantifying interactions between neisserial OMVs (10 μg total protein per well) and soluble CEACAM1-Fc. OMVs containing Opa variants that bind CEACAM1 are indicated with black bars. Relative binding was calculated based upon mean values for wells incubated with N. meningitidis-derived OMVs versus those for wells with no OMV, being 100 and 0%, respectively. In each instance, error bars indicate the standard deviations based upon values from three replicate wells. (E) Flow cytometry analysis of the association between FITC-labeled OMVs and IL-2-prestimulated lymphocytes. Matched cell populations were prestimulated using IL-2 and then incubated in the presence of FITC-OMVs derived from N. gonorrhoeae expressing the indicated Opa variants. Markers delineate the regions containing peak fluorescence intensities for T cells incubated with OMVs from OpaHSPG-expressing bacteria, drawn for comparative purposes.

Soluble chimeras consisting of the extracellular portion of human CEACAM1 fused to the Ig Fc domain (CEACAM1-Fc) were used to determine if Opa proteins retain their abilities to bind CEACAM1 family receptors when in the context of OMVs. Equal amounts of each OMV preparation were exposed to soluble CEACAM1-Fc fusion protein in an ELISA. The relative levels of CEACAM1 binding by N. gonorrhoeae-derived Opa_HSPG and N. lactamica-derived OMVs were not significantly different from that by OMVs derived from Opa⁻ bacteria; however, each of these remained a small fraction of that seen with OMVs derived from our control strain expressing an Opa variant that binds to CEACAM1 (Opa_CCM) or the Opa⁺N. meningitidis variant (Fig. 1D).

To confirm that this binding function allowed vesicle association with CEACAM1-expressing cells, OMVs prepared from the recombinant N. gonorrhoeae strains expressing defined Opa variants were labeled with FITC. These were incubated with IL-2-prestimulated CD4⁺ T lymphocytes, which were then analyzed by flow cytometry. The HSPG-specific Opa variant conferred a fivefold-increased association of the OMVs with the T cells compared to Opa-deficient OMVs, while the CEACAM-specific Opa variants displayed an even more pronounced (100-fold) increase in binding (Fig. 1E). When combined, these results demonstrate that Opa_CCM proteins retain their abilities to bind CEACAM1 when in the context of an OMV.

Meningococcal OMVs inhibit T-cell proliferation.Given that Nl-OMVs differed from Nm-OMVs with respect to Opa phenotype (Fig. 1C) and CEACAM1 binding (Fig. 1D), we tested whether there was a difference in lymphocyte response to these vaccine preparations. The meningococcal OMVs caused a dose-dependent reduction in CD4⁺ T-cell proliferation in response to ligation of the T-cell receptor with the costimulatory molecule CD28, apparent as both a delay in the onset of cell culture expansion and a slower increase in cell number over time (Fig. 2A). Consistent with their lacking Opa proteins, the N. lactamica-derived OMVs had no similar effect (Fig. 2A). The reduction in CD4⁺ T-cell proliferation in the presence of Nm-OMV could result from either an increase in cell death or a reduction in the rate of proliferation. At 120 h, there was no difference in annexin V-FITC staining of cultures administered the various antibody or OMV preparations (90.9% ± 1.88% between all samples; data not shown), consistent with the Nm-OMV-inhibiting T-cell proliferation. As in our previous study (5), polyclonal CEACAM-specific antisera displayed an inhibitory effect reminiscent of the Opa-containing OMVs (Fig. 2B), consistent with CEACAM1 ligation being sufficient to inhibit expansion of the T-cell population.

• Open in new tab
• Download powerpoint

FIG. 2.

Proliferation of CD4⁺ T lymphocytes in response to CEACAM1 ligation by neisserial OMVs or antibody. Primary CD4⁺ T cells were cultured in the presence of IL-2 and cross-linked CD3ε- and CD28-specific antibodies with various protein concentrations of either Nl-OMVs (gray bars) or Nm-OMVs (black bars) (A) or CEACAM-specific antisera (black bars) or nonreactive isotype control antibodies (gray bars) or no additions (white bars) (B). The calculated increase in culture density is relative to the number of cells present at the onset of the experiment (time = 0 h). In each instance, error bars indicate the standard deviations based upon values from six replicate samples, with results being representative of three independent experiments. Asterisks indicate P values of <0.01 for comparison with Nl-OMVs (A) or isotype control antibodies (B).

CEACAM1-binding Opa proteins confer the lymphocyte inhibitory effect.Given that species-related differences in constitutive and phase-variable virulence factors other than Opa could be responsible for the differential effects of Nm- and Nl-OMVs on T-cell proliferation, we compared the effects of OMVs isolated from our well-defined isogenic strains of N. gonorrhoeae that express distinct Opa phenotypes. As with the intact bacteria (5), OMVs containing the CEACAM1-specific Opa_CCM protein caused a marked reduction in T-cell proliferation relative to cells that have been exposed to OMVs containing either no Opa protein or the Opa_HSPG protein (Fig. 3A). The CEACAM1-specific OMVs also effectively inhibit T-cell expression of the immediate-early activation marker CD69 (27, 46), indicating a block early in the normal response of the cells to activating stimuli. Exposure to gonococcal OMVs demonstrated that CEACAM1-binding Opa proteins showed a similar block in T-cell activation and CD69 expression 16 h after T-cell activation (Fig. 3B). Combined, these results establish that Opa variants within OMVs retain their abilities to bind CEACAM1 and effectively inhibit T-cell activation by a CEACAM1-dependent mechanism.

• Open in new tab
• Download powerpoint

FIG. 3.

Proliferation and CD69 expression of CD4⁺ T lymphocytes in response to CEACAM1 binding by Opa-specific neisserial OMVs. Jurkat CD4⁺ T cells were cultured in the presence of IL-2 and/or cross-linked CD3ε-specific antibodies with OMVs (50 μg/ml protein) prepared from N. gonorrhoeae-expressing defined Opa variants. OMVs that bind CEACAM1 are indicated with black bars. (A) The mean increase in relative culture density was calculated relative to the number of cells in uninfected samples, which was defined as 100%. Error bars indicate the standard deviations based upon values from six quadrants counted for each sample. (B) The proportion of cells expressing the early activation marker CD69 16 h following the onset of the experiment was determined by flow cytometry, with standard deviation calculations based upon three independent samples. Asterisks indicate P values of <0.0002 (A) or <0.005 (B) for comparison with the other OMV-treated samples in each set.