|Home | About | Journals | Submit | Contact Us | Français|
Neisseria gonorrhoeae cells (gonococci [GC]), the etiological agents for gonorrhea, can cause repeated infections. During and after gonococcal infection, local and systemic antigonococcal antibody levels are low. These clinical data indicate the possibility that GC may suppress immune responses during infection. Carcinoembryonic antigen-related cellular adhesion molecule 1 (CEACAM1 or CD66a), a receptor for GC opacity (Opa) proteins, was shown to mediate inhibitory signals. In the present study, human B cells were activated by interleukin-2 to express CEACAM1 and then stimulated to secrete antibodies and simultaneously coincubated with Opa− and OpaI GC of strain MS11. Our results show that this OpaI GC has the ability to inhibit antibody production. The interaction of GC and CEACAM1 with human peripheral B cells also results in induction of cell death. The same findings were observed in DT40 B cells. This CEACAM1-promoted cell death pathway does not involve the inhibitory signals or the tyrosine phosphatases SHP-1 and SHP-2 but depends on Bruton's tyrosine kinase in DT40 cells. Our results suggest that Neisseria gonorrhoeae possesses the ability to suppress antibody production by killing CEACAM1-expressing B cells.
Neisseria gonorrhoeae cells (gonococci [GC]), etiologic agents of gonorrhea and disseminated gonococcal infection, can cause repeated infections, as seen from patients attending sexually transmitted disease clinics (22). Gonorrhea patients with a history of prior infection have low levels of local and systemic antigonococcal antibodies (22). In addition, previous infections with N. gonorrhoeae do not alter antibody levels in patients with a current infection (22). These effects are supported by recent findings showing a paucity of local and systemic antibodies against GC (44).
GC have the ability to escape the human immune response. It is thought that their tremendous capacity to change surface components, such as the Opa proteins, pili, and lipooligosaccharides (LOS) (3, 18, 28, 37, 48, 54, 55, 59), contributes to immune evasion. In addition, antigenic variation of these surface components influences virulence in human challenge models (27, 52, 53, 56, 62, 63). Significant progress has been made toward understanding how GC promotes its survival within the host through the expression of bacterial virulence genes. However, much less is known about how the host responds to these pathogens to shape the outcome of infection.
To establish infection, N. gonorrhoeae cells must interact with receptors on host cells. This process involves adherence and penetration, which may help shield GC from host-mediated killing by the innate and humoral immune responses. In strain MS11, the Opa family consists of 11 unlinked opa genes whose sequences are known (4). Some Opa proteins, such as OpaI, promote strong phagocytosis by polymorphonuclear leukocytes, while other Opa proteins elicit intermediate levels of interaction. In contrast, OpaA bacteria (GC or Escherichia coli) do not stimulate invasion and interact like Opa− organisms (2, 32). The CEA (carcinoembryonic antigen-related cellular adhesion molecule [CEACAM] or CD66) antigen family, including CEACAM1 (CD66a) and CEACAM3 (CD66d), has been identified as receptors for some GC opacity (Opa) proteins, which can promote adherence to and invasion of microorganisms (12, 13, 19, 67) in epithelial cells and neutrophils. CEACAM3 uses its immunoreceptor tyrosine-based activation motif (ITAM) to promote signaling and mediate invasion of GC in DT40 B cells (11) as well as in HeLa cells (5). ITAMs are also present in the cytoplasmic regions of B-cell receptors, T-cell receptors, and Fc receptors (9, 46, 47). The conserved tyrosines within ITAM are sites of phosphorylation events that involve the Syk kinases and are essential for propagation of signals, including those for bacterial invasion.
On the other hand, CEACAM1 activates an inhibitory pathway and carries out inhibitory signaling through the immunoreceptor tyrosine-based inhibitory motif (ITIM) in its cytoplasmic domain (1, 15, 25). ITIM-bearing molecules, such as FcγRIIB (34, 35, 40, 41), can interrupt signal transduction and ultimately inhibit antibody production (17, 31, 49). For example, measles virus, a notorious agent for suppression of antibody production (10, 36, 50), binds to human FcγRIIB, which is functionally related to CEACAM1 (15), and inhibits human B-cell antibody production (45). It is anticipated that if human B cells express CEACAM1, interaction of GC with this receptor might inhibit antibody production. Consistent with this hypothesis, Boulton and GrayOwen (8) showed that GC bind to CEACAM1 and inhibit the activation and proliferation of CD4+ T cells. The down-regulation of adaptive immune responses by GC should be advantageous to infecting gonococci.
In the present study, we show that GC interact with CEACAM1 on human B cells and inhibit antibody production. We further find that the GC-CEACAM1 interaction in B cells leads to the induction of cell death.
Gonococcal strain MS11 was cultured and maintained as previously described (14, 61). Only pilus-negative gonococci with LOSb (lacto-N-neotetraose) phenotypes were used (14, 60). LOS background was determined by silver stain and monoclonal antibodies against LOSa and LOSb as we described previously (14). This LOSb GC is the same LOS background as the variant C strain of MS11, described in Schneider et al.'s human challenge experiment (53). However, no sialylation is involved in any experiments in this study, although LOSb GC is sialylable. Recombinant opa genes from N. gonorrhoeae MS11 were expressed in E. coli HB101 (2). Expression of Opa proteins, such as OpaI, from both MS11 and E. coli was routinely monitored by Western blotting using the Opa cross-reactive monoclonal antibody 4B12 (2, 62). Suspensions of E. coli and GC were prepared from bacteria grown for 16 to 20 h at 37°C on Luria-Bertani (LB) plates containing 50 μg/ml carbenicillin and GC plates, respectively.
Wild-type chicken DT40 B cells, their derived mutants (SHIP−/−, SHP-1−/−, SHP-2−/−, Bruton's tyrosine kinase [BTK]−/−, and Syk−/−) (6, 35, 40, 41, 64, 65), and their CEACAM1 transfectants were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 1% chicken serum, 50 μM 2-mercaptoethanol, and 2 mM l-glutamine.
The following antibodies were used in this study: COL-1, a monoclonal antibody specific for CEACAM3 and CEA (CD66e); B6.2, which reacted with CEACAM6 only (Zymed Laboratories Inc., California); and YTH71.3, which recognizes CEACAM1, CEACAM6, and CEACAM3 (CD66d), obtained from Harlan Bioproducts (Indianapolis, Indiana).
The two forms of cDNA, CEACAM1-long (L) and CEACAM1-short (S), originally cloned from a colon cell line (23), were used to transfect HeLa and DT40 B cells (hereafter HeLa-CEACAM1 and DT40-CEACAM1 cells). Unless specifically mentioned, HeLa-CEACAM1 or DT40-CEACAM1 cells are CEACAM1-long (L) transfectants.
Caspase-3 inhibitor (Ac-DEVD-CMK) was purchased from Calbiochem, San Diego, CA.
Peripheral blood mononuclear cells (PBMC) were isolated from human blood by centrifugation through Ficoll-Plaque (Amersham Pharmacia Biotech). The purified PBMC were suspended in RPMI 1640 medium (GIBCO BRL, Grand Island, NY) with 10% fetal calf serum (FCS) (HyClone, Logan, Utah) at a concentration of 106/ml. Interleukin-2 (IL-2) was added at a concentration of 250 U per ml to the PBMC suspensions, which were incubated for 3 or 4 days.
The MTT cell proliferation assay (Sigma, St. Louis, MO) is for the measurement of cell proliferation based upon the reduction of the tetrazolium salt 3,[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT). MTT is reduced to an insoluble purple formazan dye by mitochondrial enzymes associated with metabolic activity. The procedures followed the company's recommendation. Briefly, purified human B cells at a concentration of 106/ml that were treated with E. coli lipopolysaccharide (LPS) (100 ng/ml) or not treated and DT40 B cells at a concentration of 105/ml were loaded in 96-well flat-bottomed tissue culture plates in 0.1 ml of RPMI with 10% fetal bovine serum. Each day, 0.01 ml MTT solution was added to selected wells, and cultures were mixed. Samples were incubated at 37°C for cleavage of MTT insoluble dye. Then, 0.1 ml stop solution containing sodium dodecyl sulfate and acidified isopropanol was added to each well and mixed thoroughly, and wells were stored at 4°C until all samples were collected. The absorbance of each sample was measured at a wavelength of 540 nm by an enzyme-linked immunosorbent assay (ELISA) plate reader.
The reagents and procedures for the B-cell enrichment were provided by One Lamda, Inc., California. Basically, PBMC were spun down and suspended in Lympho-Kwik B-cell reagent 1, containing complement and antibodies against multiple cell types in PBMC except B cells. The mixture was incubated for 1 h at 37°C and mixed occasionally by inverting the tube. Dulbecco's modified phosphate-buffered saline (DPBS) was layered on top of reagent 1, and the sample was centrifuged for 2 min at 2,000 rpm. The supernatant was discarded, and the pellet, containing B cells (60 to 75%), was resuspended in reagent 2. Samples were washed twice with DPBS and resuspended in the desired medium.
Double labeling methods were used to identify the expression level of CEACAM1 in B cells. IL-2-treated PBMC were stained with anti-CD19 antibody tagged with allophycocyanin and anti-CEACAM antibody tagged with fluorescein isothiocyanate (FITC) for 1 h in PBS containing 2% FCS. After washing once, these mixtures were analyzed by flow cytometry using a FACScan (Becton Dickinson, Mountain View, California). For several experiments, PBMC were pretreated with Lympho-Kwik solutions (as described above) to enrich the concentration of B cells before flow cytometry analysis. To obtain highly purified B cells, the B cells from PBMC were sorted with anti-CD19 antibody tagged with FITC (Vantage SE; Becton Dickinson, Franklin Lakes, New Jersey). This procedure yields approximately 98% purity of CD19+ cells. However, we noticed that the isolation of the primary B cells was subject to donor variability.
Purified human B cells were cultured in RPMI 1640 medium with 10% fetal calf serum in 96-well plates at a concentration of 106 per ml. B cells were incubated with anti-CD40 and -Mμ antibodies at a concentration of 1 μg/ml and at the same time infected with Opa− and OpaI GC at a ratio of 25 to 1 (bacteria to B cells) for 2.5 h. Then the bacteria were killed by adding gentamicin, penicillin, and streptomycin, and the samples were incubated for 3 days in RPMI medium with 10% FCS. Supernatants were collected, and antibody levels were measured by a standard ELISA method described below.
Unlabeled anti-human antibody (heavy plus light chains) (Southern Biotechnology, Inc, Birmingham, AL) was diluted to 10 μg/ml in 10 mM carbonate at pH 9.2 and was used to coat 96-well ELISA plates. After 2 h, the surfaces of plates were blocked with 1% gelatin in PBS-Tween 20. Twelve standards with serial dilutions of human immunoglobulin G (IgG) for a standard curve and samples were loaded in these wells and incubated for 1 h. After washing twice with PBS-Tween 20, the second antibody (goat anti-human, horseradish peroxidase-conjugated IgG) was used. 2,2′-Azinobis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) (BioFx Laboratories, Owings Mills, MD) was added for development, and results were recorded on a Microplate Autoreader (EL311SX; bio-tek Instruments, Inc., Winooski, Vermont).
Five hundred microliters of bacterial suspensions in RPMI (optical density at 540 nm = 0.02) was added to each well in 24-well plates and incubated for 120 min at 37°C with 5% CO2. HeLa cell monolayers were washed twice with PBS. Cells were suspended in PBS containing 0.5% saponin (Calbiochem Corp.), diluted, and plated on LB agar medium containing 50 μg/ml of carbenicillin. Invasion assays were performed as follows. DT40 or HeLa cells were suspended in RPMI with 2% FCS at a concentration of 2 × 105/ml. One-half milliliter of each of these cell suspensions was added to 24-well plates, and after addition of 50 μl of bacterial suspensions in RPMI at a concentration of 8 × 107/ml, the cells were allowed to incubate for 2.5 h at 37°C in the presence of 5% CO2. The ratio of bacteria to DT40 or HeLa cells was about 25 to 1. Gentamicin was added to each well at a final concentration of 100 μg/ml, and the cultures were incubated for 90 min. The cells again were suspended in PBS containing 0.5% saponin (Calbiochem Corp.), diluted, and plated on LB agar medium containing 50 μg/ml of carbenicillin. The level of internalization of bacteria in HeLa-CEACAM1 cells was calculated by determining the CFU recovered from lysed cells. All experiments were performed in duplicate or triplicate. Bars in figures represent standard error.
DT40 and CEACAM1-expressing B cells do not stick to tissue culture wells very efficiently; therefore, in order to supplement the epithelial cell-based adherence assay described above, we determined cellular adherence by directly counting cells. Briefly, after interaction in 24-well plates in which each well contained a coverslip as described above, B cells were washed twice with PBS using cytospin and fixed with 2% paraformaldehyde in PBS containing Giemsa stain. The number of associated bacteria (adherent and internalized) per B cell was determined by microscopy by counting the bacteria associated with 100 cells on the coverslips.
All experiments were performed in triplicate, and data were expressed as the mean ± the standard error. Statistical significance was calculated using Student's t test. It should be noted that the infection time selected in our experiment is based on our preliminary experiments and is a balance between the numbers of bacteria killed by host cells and the time necessary for bacterial entry.
Ten micrograms of CEACAM1 cDNA cloned in pHβ expression vector (24) was cotransfected with 1 μg of pBabe-puror vector (39) into 5 × 106 DT40 or HeLa cells by electroporation at 250 V and 960 μF in 0.5 ml of PBS. Stable transfectants were selected in 0.5 μg/ml of puromycin for DT40 and 1 mg/ml of Geneticin for HeLa cells. The presence of CD66 antigens in individual clones was determined by flow cytometry using FACScan with YTH71.3 and secondary antibodies conjugated with FITC. Unless specified, CEACAM1 in DT40 and HeLa transfectants is the long form. We tested two to three different transfectants of each mutant to make sure that they behaved equally. It should be noted that antireceptor antibodies are often used to stimulate the activation of receptors, such as B-cell receptors and T-cell receptors. However, none of the anti-CEACAM antibodies we tested (COL-1, B1.1, 4-12-5, and YTH71.3) were able to stimulate calcium flux in DT40-CEACAM3, even when cross-linked with secondary antibodies.
Annexin V-FITC binding to phosphatidylserine was used as a B-cell death assay according to the manufacturer's recommendations (Pharmingen, San Diego, CA). Annexin V-FITC was used in conjunction with propidium iodide (PI) to distinguish apoptotic cells (annexin V-FITC positive and PI negative) from dead cells (annexin V-FITC positive and PI positive). Cells exhibited apoptotic cell-specific phosphatidylserine on their surface and therefore bound annexin V-FITC or took up PI, which cannot penetrate live cells. In our experiment, the cells that were either annexin V-FITC positive or PI positive were classified as dead cells (see Fig. Fig.4).4). Human B cells or DT40 cells (105) were suspended in 0.1 ml of 2% FCS-RPMI containing 1.5 × 106 bacteria (GC) and incubated at 37°C for 2.5 h with occasional shaking. The bacteria were killed by adding gentamicin, penicillin, and streptomycin, and the cells were further incubated overnight. Then, 50 μl of binding buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, and 2.5 mM CaCl2), 7.5 μl of annexin V-FITC, and 15 μl of PI were added to each vial. The cell suspensions were gently mixed and incubated for 15 min at room temperature in the dark. Finally, after addition of 550 μl of binding buffer, samples were analyzed immediately by flow cytometry.
It is reported that IL-2 stimulates the expression of CEACAM1 on human B cells (29, 58). We could detect very limited expression of CEACAM1 when purified human B cells were treated with IL-2 alone (data not shown). However, when human PBMC were treated with IL-2 for 3 days in RPMI medium, the B cells expressed CEACAM1, as shown by double labeling with anti-CEACAM1 and anti-CD19 antibodies using flow cytometry (Fig. (Fig.1B).1B). Without IL-2 treatment, the expression of CEACAM 1 on B cells was weak (Fig. (Fig.1A).1A). Therefore, IL-2 is necessary but not sufficient to stimulate the expression of CEACAM1 in human B cells. Although it is reported that there are two CEACAM1 forms (CEACAM1-L and CEACAM1-S) in human B cells (58), we were able to detect the expression of only CEACAM1-L in IL-2-treated B cells (data not shown).
OpaI-expressing bacteria are able to interact with CEACAM1, promoting phagocytosis of bacteria by host cells (13, 19, 67). Therefore, strains of GC and E. coli (OpaI and Opa−) were tested for adherence to and invasion of purified human B cells (IL-2 treated and untreated). Figure Figure1C1C shows that OpaI GC and E. coli are efficiently bound to human CEACAM1-expressing B cells (IL treated) but not to untreated human B cells. Opa− E. coli did not bind to B cells, but the limited level of Opa− GC binding to IL-2-treated B cells might due to the fact that strain MS11 interacts with CEACAM1 regardless of Opa expression (7, 20, 42). However, both GC and E. coli (Opa− and Opa+) could not invade IL-2-treated human B cells and DT40-CEACAM1 cells, in contrast with OpaI GC invasion into DT40-CEACAM3 cells (Fig. (Fig.1D)1D) (11). These results suggest that OpaI bacteria interact with IL-2-treated B cells but do not invade them.
B cells undergo several stages of differentiation, maturation, and proliferation in lymph nodes or the spleen to become mature B cells and secrete antibodies in vivo. Mouse B cells express CEACAM1 without stimulation with IL-2 (21) and can be induced to proliferate and produce antibody by bacterial LPS in vitro (26). However, OpaI GC did not interact with mouse CEACAM1-expressing B cells, since OpaI GC bind only to human CEACAM1 (data now shown). Therefore, our study focuses on the human CEACAM1-expressing B cells. Control DT40 cells used in the proliferation assays were allowed to proliferate for 5 days before undergoing cell death (Fig. (Fig.2A).2A). Purified human B cells cannot be stimulated to proliferate by LPS (Fig. (Fig.2A).2A). However, it is reported that human B cells can be stimulated to secrete antibodies in vitro by cytokines such as B lymphocyte stimulator (38, 57, 68), CD40 ligand, or anti-Mμ and anti-CD40 antibodies. We have tested all three reagents, individually or in different combinations, for stimulation of antibody production in our system and found that the combination of anti-Mμ and CD40 antibodies is the most effective (data not shown). As described above, activation of B cells in the presence of other PBMC by IL-2 stimulates expression of CEACAM1, an inhibitory receptor. One of the functions of inhibitory receptors, such as FcγRIIB, is to suppress antibody production (45). To determine whether the interaction of CEACAM1 on B cells with Opa+ GC inhibits antibody production, PBMC were treated with IL-2 for 3 days, and CEACAM1-expressing B cells (Fig. (Fig.1B)1B) were purified by flow cytometry sorting with anti-CD19 antibodies. The purified CEACAM1-expressing B cells were incubated with anti-CD40 and anti-Mμ antibodies. We used purified and non-IL-2-treated B cells as the controls. At the same time, the samples were coincubated with Opa− and OpaI GC for 2.5 h. The bacteria in the mixtures were killed by addition of antibiotics. The cultures were incubated for another 3 days. Supernatants were then collected, and the amount of secreted antibodies was measured by ELISA. As shown in Fig. Fig.2B,2B, OpaI GC inhibited antibody production. These results indicate that OpaI bacteria can inhibit antibody production by CEACAM1-expressing B cells in vitro, whereas Opa− strains cannot. Antibody production in B cells without treatment of IL-2 could not be inhibited by either Opa− or OpaI GC. Instead, these cells were stimulated by GC to secrete a small amount of antibody (Fig. (Fig.2B2B).
One plausible explanation for the inhibition of antibody production by OpaI is that CEACAM1 promotes a negative signal to inhibit proliferation and antibody production. A similar effect was observed after the interaction of CEACAM1-expressing CD4+ T cells with GC (8). However, in order to exert inhibitory effects, inhibitory receptors are usually coligated and costimulated with activated receptors, and the ITIM in their cytoplasmic domain is required. There are two forms of CEACAM1 generated by alternative splicing: the ITIM-containing long form (L) and the short form (S, without ITIM). We sought to determine the capacity of the different CEACAM1 forms to stimulate B-cell death. The first experiment was aimed at determining the capacity of stimulation of cell death by different CEACAM1 forms. cDNAs corresponding to CEACAM1-L and -S were stably transfected and expressed in DT40 B cells (Fig. (Fig.3A).3A). For CEACAM1-S, we could obtain only transfectants with low expression, indicating that high expression of CEACAM1-S may be deleterious to the cells. However, when DT40 B cells expressing either CEACAM1-L or CEACAM1-S were infected with OpaI GC, both transfectants were similarly killed. This demonstrates that CEACAM1 has the ability to induce host cell death regardless of the presence of its ITIM and suggests that the cytoplasmic domain may not be involved in mediating the function of CEACAM1 relative to B-cell death. A second set of experiments was performed utilizing the ability of CEACAM1 to mediate invasion of bacteria in HeLa cells (13, 19, 66). The two forms of cDNA were stably transfected in HeLa cells (Fig. (Fig.3B).3B). CEACAM1 (L or S) transfectants expressing the same levels were then challenged with OpaI bacteria. Internalized bacteria were recorded as bacteria recovered after gentamicin treatment. As shown in Fig. Fig.3B,3B, there are no significant differences between HeLa-CEACAM1-L and HeLa-CEACAM1-S cells in terms of adherence and internalization of bacteria, indicating again that ITIM is not involved when the receptor is induced and activated directly.
Activation via CEACAM1 led to killing of DT40 B cells (Fig. (Fig.3A).3A). We examined whether CEACAM1-mediated cell death could also be observed in CEACAM1-expressing human B cells. As described above, B cells from PBMC were treated with IL-2 to induce expression of CEACAM1. B cells (CD19-positive cells) were sorted with anti-CD19 antibody and infected with Opa− and OpaI GC and E. coli. As shown in Fig. Fig.4C,4C, OpaI introduced into E. coli stimulated greater cell death than Opa− E. coli (Fig. (Fig.4B).4B). Although both Opa− GC and OpaI GC killed CEACAM1-expressing B cells, OpaI GC (Fig. (Fig.4E)4E) killed B cells more efficiently than Opa− GC (Fig. (Fig.4D).4D). That Opa− GC can still kill CEACAM1-expressing B cells may be related to the observation that Opa− GC also interact with CEACAM1 (7, 20, 42). For example, both Opa+ and Opa− gonococci were taken up by CEACAM1-transfected HL60 cells (42). These results suggest that there are two distinct mechanisms of interaction between gonococci and CEACAM1. One is an Opa protein-dependent interaction. The other mechanism is Opa independent, involving a putative ligand on gonococci which has not yet been identified. Also, the cell death observed in CEACAM1-expressing B cells with Opa− GC could be caused by the low number of Opa+ revertants due to phase variation. Regardless, GC possess the ability to kill CEACAM1-expressing human B cells.
Previous studies showed that CEACAM1-mediated signals involve SHP-1 and SHP-2 (1, 25) when there is coligation and activation with ITAM-containing receptors (15). However, the results described above indicate that the ITIM on CEACAM1 is not involved in death signaling if CEACAM1 is directly activated by GC. ITIM enacts its inhibitory signaling through SHIP and SHP-1 or SHP-2 (1, 25). To explore the possible signaling pathways of CEACAM1-mediated cell death, five knockout cell lines (DT40-ΔSHIP, DT40-ΔSHP-1, DT40-ΔSHP-2, DT40-ΔSyk, and DT40-ΔBTK) were transfected with CEACAM1 (Fig. (Fig.5B).5B). We chose DT40-ΔBTK because BTK participates in FcγRIIB-mediated cell death pathways in DT40 B cells, and the ITIM on the cytoplasmic domain of FcγRIIB receptor is a functional analog of CEACAM1 (15). To further examine whether phosphorylation of the ITIM on CEACAM1-L was necessary for CEACAM1-mediated cell death, we replaced the tyrosine residue (Y459) with phenylalanine (F) on ITIM to construct CEACAM-L-Y459F (Fig. (Fig.5B)5B) (15), which was then transfected into DT40 cells. Also, to determine that the specific binding of OpaI to CEACAM1, not simply binding of OpaI bacteria to the cell, is essential for cell death, we selected DT40-CEACAM3-Y196F (the ITAM knockout mutant) (11) and DT40-ΔSyk-CEACAM1 (ITAM pathway knockout mutant) as controls. CEACAM3 (CD66d) promotes both the invasion of OpaI-expressing bacteria and the death of DT40 cells through its ITAM on the cytoplasmic domain (11). These transfectants, which expressed the same levels of CEACAM1 on the surface, as measured by FACScan (Fig. (Fig.5B),5B), were challenged with OpaI GC. We reasoned that if CEACAM1-mediated cell death is compromised in a given mutant, then the missing protein is involved in this process. As shown in Fig. Fig.5A,5A, DT40-ΔSHP-1-CEACAM1, DT40-ΔSHP-2-CEACAM1, DT40-CEACAM1-L-Y459F, and DT40-ΔSyk-CEACAM1 cells were still killed by OpaI GC as effectively as DT40-CEACAM1 cells. DT40-ΔSHIP-CEACAM1 showed a slight increase in OpaI GC-stimulated cell death, but this increase did not reach statistical significance (P > 0.05). These results confirm that ITIM or ITAM signaling is not involved in CEACAM-mediated cell death. In contrast, cell death of DT40-ΔBTK-CEACAM1 was significantly decreased (P < 0.01). Therefore, BTK is involved in CEACAM1-mediated cell death signaling. These results indicate that the CEACAM1-L inhibitory signal and cell death signals use two separate pathways. Furthermore, OpaI GC stimulate the death of DT40-CEACAM1 but not DT40-CEACAM3-Y196F cells, suggesting the specific binding of CEACAM1 with GC is necessary for cell death.
Finally, to determine the nature of CEACAM1-mediated cell death, we examined whether the death of DT40-CEACAM1 cells was inhibited by the anti-caspase-3 inhibitor Ac-DEVD-CMK. As shown in Fig. Fig.3A,3A, OpaI GC-induced death of DT40-CEACAM1 cells was inhibited by this inhibitor, indicating the CEACAM1-mediated cell death pathway may be an apoptotic pathway.
After Neisseria gonorrhoeae overcomes the first line of host defense constituted by neutrophils, it encounters secondary host defenses. One host defense mechanism in combating microbial infection is the production of antibodies. In the present study, we showed that activation of human B cells by IL-2 stimulates expression of human CEACAM1, an inhibitory receptor, and the interaction of human CEACAM1 with GC inhibits antibody production. We further found that the GC-CEACAM1 interaction in human peripheral B cells results in induction of cell death.
A question arises as to whether GC interact with lymphocytes in vivo. Although GC infection is basically a local process, it is possible that GC may encounter lymphocytes for the following reasons. The existence of a local immune system in the female genital tract has been demonstrated, with a predominance of IgA-producing plasma cells and T cells in the fallopian tubes, uterine cervix, and vagina (16, 33). T cells are present in these sites in numbers approximately twofold higher than plasma cells. GC-infected segments of fallopian tubes were shown to contain 6- to 10-fold-increased amounts of plasma cells of all classes (33). These data suggest that a local immune response may provide a defense against GC infection. A delay in an immune response caused by the interaction of GC with CEACAM1 could potentially increase the chance that bacteria successfully colonize the urethral or cervical mucosa and persist for an extended period of time.
CEACAM1 and CEACAM3 surface antigens serve as receptors for some opacity (Opa) proteins of GC, promoting adherence and invasion of this microorganism (12, 13, 19, 67) in epithelial cells and neutrophils. GC may exploit a well-characterized intracellular signaling pathway, namely, the ITAM signal transduction pathway, to mediate invasion by interaction with CEACAM3 (5, 11). On the other hand, CEACAM1 was identified as an inhibitory receptor able to mediate negative signals delivered through its ITIM, which recruits SHP-1 and SHP-2 phosphatases (15). It has been shown that CEACAM1 directly interacts with SHP-1 and SHP-2 phosphatases (1, 25). In general, the ITIM is required to inhibit ITAM-promoted activating signals when ITAM- and ITIM-bearing receptors are coactivated and coligated. However, Boulton and GrayOwen showed that N. gonorrhoeae cells inhibit CD4+ T-cell proliferation, which can involve the recruitment of SHP-1 and SHP-2 (8). Different outcomes may depend on different host cells (CD4+ T cells versus B cells) used in these studies. Our results support the notion that, in order to exert their inhibitory effects, inhibitory receptors should be coligated and costimulated with activating receptors. Otherwise, the so-called inhibitory receptors may deliver an activation signal if they are induced directly. There is a functional difference between CEACAM1-S, which lacks a cytoplasmic region that includes the ITIM, and ITIM-mutated CEACAM1-L-Y459F. As shown in Fig. Fig.3A3A and described previously, we could not isolate transfectants with high expression of CEACAM1-S, but DT40 cells tolerate high levels of CEACAM1-L-Y459F, suggesting that another functional domain(s), in addition to the ITIM, is missing in CEACAM1-S.
CEACAM1 behaves like a cell death molecule. The question remains as to which signaling pathways are involved. Direct activation of CEACAM1 by GC does not involve SHP-1, SHP-2 (Fig. (Fig.5A),5A), Syk kinase, or phospholipase C activity (data not shown) but partially depends on SHIP and the BTK kinases in DT40 cells. SHIP, by hydrolyzing PIP3 [phosphatidylinositol (3, 4, 5) trisphosphate], leads to the association of BTK (6). BTK family kinases have been shown to play an important role in the regulation of various cellular processes, including apoptosis and cell motility (43, 51). Direct stimulation of CEACAM1 by GC may stimulate a cell death pathway. The activities and functions of CEACAM1 mimic another receptor, FcγRIIB, a well-documented inhibitory receptor able to stimulate the death of DT40 B cells (43) when activated directly. FcγRIIB-mediated cell death also does not involve its ITIM motif but rather its association with SHIP and BTK. Although our functional studies indicated that the same may be possible for CEACAM1, the direct association of CEACAM1 with SHIP or BTK remains to be determined. Nevertheless, these data show that the cellular responses initiated through CEACAM1 resemble the events following activation of FcγRIIB either by coligation for inhibitory activation (6, 40, 41) or direct activation for apoptotic ability demonstrated in this study. In original reports, in which binding of measles virus with FcγRIIB on B cells inhibits antibody production, the primary reason appears to be the killing of B cells by this virus (45). Very recently, the short form of CEACAM1 was demonstrated to mediate apoptosis and revert mammary carcinoma cells to a normal morphogenic phenotype (30), indicating again that CEACAM1 can function as cell death receptor and that it does not require the ITIM for that activity. Finally, it should be stated that CEACAM1-mediated cell death is dependent on which types of cells are activated, since neither forms of CEACAM1 stimulate death of HeLa cells, even when challenged with GC. CEACAM1 did not promote invasion of bacteria in DT40 cells (data not shown).
Repeated infections among gonorrhea patients are very common. GC appear to have the capability to inhibit the immune response, which may be in part due to suppression of antibody production (22, 44). The present study begins to unveil potential mechanisms of antibody suppression, i.e., GC specifically targeting the CEACAM1-expressing B cells and causing cell death, which consequently may play a role in GC-induced immunosuppression during infection. It is of interest that GC use the same molecule, CEACAM1, not only for their binding to host but also for subversion of regulatory immune pathways. Further studies should uncover mechanisms of gonococcal infection and help us understand how microbial pathogens exploit host cells.
We thank Margaret Bauer, Raoul Rosenthal, and Hal Broxmeyer for useful suggestions and editorial comments on the manuscript.
This work was supported by PHS grants R01AI 47736 and R03TW006270 to T.C.
Editor: D. L. Burns