Role of CD80 (B7.1) and CD86 (B7.2) in the Immune Response to an Intracellular Pathogen¹

The generation of a protective immune response against invading infectious organisms requires that the immune system discriminate microbial Ags from self Ags. It has been proposed that one way the immune system accomplishes this task is by requiring the presence of a microbially induced second signal during T cell activation (1). While recognition by the TCR of peptide-MHC complexes initiates Ag-specific T cell responses, a second signal (costimulatory signal) provided by the APC is necessary for T cell clonal expansion and optimal cytokine production (2). The relevance of costimulatory signals is emphasized by the induction of T cell nonresponsiveness or anergy when T cells encounter Ags in the absence of costimulation (2, 3, 4).

Data accumulated to date suggest that the interaction between CD28 expressed on T cells and its counter-receptors CD80 (B7-1) and CD86 (B7-2) expressed on specialized APC provides the most important costimulatory signal (5, 6). Both CD80 and CD86 can provide costimulation to T cells for proliferation and IL-2 production (7, 8). However, these molecules differ in their expression on APC. Whereas dendritic cells express both costimulatory ligands, monocytes/macrophages constitutively express only CD86 (9). Resting B cells have low levels of expression of CD86 and no significant expression of CD80 (9).

It has been reported that the expression of CD80 and CD86 on APC can be regulated. Incubation with IFN-γ induces the expression of CD80 and up-regulates the expression of CD86 on monocytes (10, 11). In addition, activation of B cells results in induction of the expression of CD80 and up-regulation of the expression of CD86 (9, 12). Regulation of the expression of CD80 and CD86 on APC may be an important feature of the biology of these molecules with potential implications in self/nonself discrimination.

The outcome of infections with certain organisms has been shown to correlate with the type of cytokines produced by T cells (13). In this regard, there is ample evidence of the pivotal role that IFN-γ plays in the induction of protective immunity against intracellular micro-organisms (14). Thus, the identification of factors that influence cytokine production triggered by infectious organisms are crucial to our understanding of the mechanisms that determine whether a protective immune response is elicited. Toxoplasma gondii, an obligate intracellular protozoan that infects all nucleated cells, provides an example of a micro-organism against which cell-mediated immunity with resulting IFN-γ production plays a critical role in controlling infection (15, 16). Indeed, this parasite has become a major opportunistic pathogen in immunocompromised individuals (17). The demonstration that presumably unprimed human T cells respond in vitro to this parasite (18) made T. gondii well suited for study of the roles of CD80 and CD86 in the early events of the T cell response against an intracellular organism. Our results indicate that human monocytes discriminate between noxious and innocuous preparations of T. gondii, which translates into the induction of the expression of CD80 and the up-regulation of the expression of CD86 only when these cells encounter viable (noxious) organisms. Our results reveal that, in turn, T cell proliferation is dependent on CD80 and CD86, and that the induction/up-regulation of expression of these molecules is associated with the generation of an IFN-γ response by T cells from unexposed individuals. These data provide evidence of the importance of pathogen-monocyte interaction, especially through pathogen-mediated induction/up-regulation of costimulatory ligands, in shaping the ensuing T cell response.

The following mAbs were used for cell purifications: anti-CD2, anti-CD3, anti-CD8, anti-CD14, and anti-CD56 (all from Becton Dickinson, San Jose, CA); anti-CD16 (Medarex, Inc., Annandale, NJ); anti-CD19 (Coulter, Hialeah, FL); anti-CD45RA, anti-CD45RO, and anti-CD66b (all from Immunotech, Westbrook, MA); and anti-glycophorin A (10F7 MN, American Type Culture Collection, Rockville, MD).

Anti-CD14 (RM052, Immunotech), anti-CD80 (L307, Becton Dickinson), anti-CD86 (Fun-1, PharMingen, San Diego, CA), CTLA-4-Ig (gift from Dr. Peter Linsley) (19), and neutralizing mAb against the cytokines GM-CSF,³ IFN-γ (R&D Systems, Minneapolis, MN), and IFN-α (BioSource International, Camarillo, CA) were used in functional assays (all at 10 μg/ml). Isotype-matched mAbs and human IgG1 were obtained from PharMingen and Sigma Chemical Co. (St. Louis, MO), respectively. IFN-γ, purchased from R&D Systems, was used at 100 U/ml.

The following conjugated Abs were used for flow cytometry (purchased from Caltag, South San Francisco, CA, except when indicated): FITC-anti-CD3 (Becton Dickinson), FITC-anti-CD14, FITC-anti-CD19, FITC-anti-CD45RA (Immunotech), FITC-anti-CD45RO (Immunotech), FITC-anti-CD66b (Immunotech), phycoerythrin (PE)-anti-CD11a, PE-anti-CD16 (Becton Dickinson), PE-anti-CD54, PE-anti-CD56 (Becton Dickinson), PE-anti-CD58 (Becton Dickinson), PE-anti-CD80 (Becton Dickinson), PE-anti-CD86 (PharMingen), or PE-anti-HLA-DR.

PBMC were isolated from buffy coats of heparinized blood of healthy volunteers donors obtained from the Stanford Blood Bank (Stanford, CA). Serologic tests for detection of anti-T. gondii IgG and IgM were performed in all samples of blood (20). This allowed for division of the donors into T. gondii-seronegative (negative anti-T. gondii IgG and negative anti-T. gondii IgM) and T. gondii chronically infected individuals (positive anti-T. gondii IgG and negative anti-T. gondii IgM).

To obtain purified monocytes, PBMC were incubated with the following mAb: anti-CD2, anti-CD3, anti-CD8, anti-CD19, anti-CD56, anti-CD66b, and anti-glycophorin A. After addition of magnetic beads coated with anti-mouse IgG (Dynal, Great Neck, NY), rosetting cells were removed with a magnet (Dynal). This resulted in populations that were >96% pure for monocytes by microscopic examination of Giemsa-stained cytocentrifuge preparations. In addition, cytofluorometric analysis indicated that >92% of the cells were CD14⁺, with <0.5% CD3⁺, <0.5% CD19⁺, <0.5% CD56⁺, and <2% CD66b⁺ cells. In certain experiments monocytes were purified further by incubation with FITC-conjugated anti-CD14 mAb (Becton Dickinson) followed by FACS sorting. This procedure resulted in populations of highly purified CD14⁺ monocytes (>99% by flow cytometry). For some experiments purified monocytes were incubated for 2 days in complete medium (CM) consisting of RPMI 1640 with 10% dye test-negative human AB serum (Irvine, Santa Ana, CA) before infection with T. gondii. This resulted in populations of cells that acquired macrophage morphology (21).

Resting T cells (>99% CD3⁺) were obtained from nylon wool-nonadherent PBL that were incubated with anti-CD16 plus anti-CD56 mAb and subjected to depletion using immunomagnetic beads. To obtain purified populations of CD4⁺ CD45RA⁺ TCR-αβ⁺, anti-CD8 (OKT8, American Type Culture Collection), anti-γδ TCR (anti-TCRδ1, gift from Dr. Michael Brenner), anti-CD19 (Coulter Cytometry, Hialeah, FL), and anti-CD45RO (UCHL-1, Immunotech) were added to the combination of mAb mentioned above. Addition of magnetic beads was repeated once. The populations obtained (>98% CD4⁺ CD45RA⁺ TCR-αβ⁺) did not proliferate in response to a recall Ag (tetanus toxoid), whereas unseparated CD4⁺ TCR-αβ⁺ and CD4⁺ CD45RO⁺ TCR-αβ⁺ cells exhibited significant proliferation in response to this Ag.

T. gondii tachyzoites were obtained from both infected human foreskin fibroblasts (American Type Culture Collection) and peritoneal cavities of infected mice and exposed to UV light as previously described (22). Neither uninfected human foreskin fibroblasts nor tachyzoite-free peritoneal lavage fluids from infected mice (after passage through a 0.45-μm filter) mediated changes in expression of the surface molecules tested. In certain experiments, tachyzoites were killed by incubation in 1% paraformaldehyde in PBS (18). To obtain T. gondii lysate Ags (TLA), tachyzoites were harvested from infected human foreskin fibroblasts and lysed in H₂O by three cycles of freezing and thawing followed by reconstitution with 10× PBS. Antigenic preparations were devoid of detectable levels of endotoxin (<10 pg/ml) using a Limulus amebocyte lysate assay (Sigma).

Unless stated otherwise, PBMC were incubated with tachyzoites of T. gondii at a ratio of 1:1 for 24 h, and purified monocytes were incubated with four tachyzoites per cell for 48 h before cytofluorometric analysis. Cells were cultured in Teflon vessels in CM. The percentage of cells with intracellular parasites was determined by light microscopy (22). In some experiments, Transwell inserts (Corning Costar Corp., Cambridge, MA) were used to separate monocytes plus T. gondii tachyzoites from monocytes alone by a membrane with pores 0.4 μm in diameter.

Freshly isolated and cultured cells were stained with either FITC-conjugated anti-CD14 or FITC-conjugated anti-CD19 mAb and one of the following PE-conjugated mAb: anti-CD11a, anti-CD54, anti-CD58, anti-CD80, anti-CD86, or anti-HLA-DR. When analyzing PBMC, an electronic gate was set on FITC-stained CD14⁺ or CD19⁺ cells to identify monocytes and B cells, respectively. For surface molecules whose levels of expression followed a bimodal distribution (CD80 and CD86 on monocytes), the percentages of CD80⁺ and CD86^high cells were calculated by determining the percentage of cells that stain above the value of fluorescence obtained with isotype control mAb (CD80⁺) or the value of fluorescence of CD86^int cells (CD86^high). For surface molecules whose levels of expression followed a unimodal distribution, data are expressed as the corrected mean fluorescence intensity (MFI). The corrected MFI for PE-conjugated mAb were calculated by subtracting the MFI of the appropriate isotype control from the MFI of each specific mAb. In some experiments, the T. gondii-induced increase in the levels of expression of CD80 and CD86 is expressed as Δ MFI, which were calculated by subtracting corrected MFI for CD80 and CD86 of uninfected monocytes from corrected MFI for CD80 and CD86 on CD80⁺ and CD86^high monocytes, respectively. Sorting of monocytes into CD80⁻, CD80⁺, CD86^int, and CD86^high populations was performed 18 h after incubation with T. gondii.

T cells (5 × 10⁵/ml) were incubated with either T. gondii-infected or uninfected, γ-irradiated, autologous PBMC with or without TLA (10 μg/ml) as previously described (18). In certain experiments, monocytes were used instead of PBMC. Infected cells were added to T cells at a ratio of five T cells per one infected cell (18). Concentrations of uninfected PBMC or monocytes matched those of infected cells. CTLA-4-Ig, anti-CD80 (L307), and anti-CD86 (Fun-1) mAb or control Abs were added to either PBMC or monocytes 30 min before incubation with T cells. Unless indicated, these reagents were used at 10 μg/ml. Cells were cultured in 96-well plates for 7 days, labeled for the final 18 h with [³H]thymidine, and harvested (18). Radioactivity was measured in a beta scintillation counter (18). Results are expressed as the mean counts per minute of [³H]thymidine incorporation of triplicate wells ± SEM. Data are also presented as stimulation indexes (counts per minute of culture with T. gondii/counts per minute of culture without T. gondii).

Purified, resting peripheral blood T cells (1 × 10⁶/ml) were incubated with purified monocytes (2–5 × 10⁵/ml) that were uninfected, infected with viable UV-attenuated tachyzoites of T. gondii (8 × 10⁵/ml to 2 × 10⁶/ml), or incubated with TLA (10 μg/ml) in 96-well plates. Supernatants were collected at 24, 48, and 72 h and stored at −70°C. Concentrations of IL-2, IL-4, and IFN-γ were measured by ELISA (18) in supernatants collected at 24, 48, and 72 h, respectively. Data are presented as the mean of triplicate wells ± SEM. None of the cytokines tested was detected in supernatants obtained from wells that lacked T cells and contained only monocytes with or without T. gondii antigenic preparations.

Statistical significance was assessed by unpaired Student’s t test.

APC express accessory molecules that can enhance Ag-driven T cell responses (23, 24). The effects of T. gondii on the expression of accessory ligands such as CD11a, CD54, CD58, CD80, and CD86 and on the expression of HLA-DR molecules on monocytes and B cells were analyzed by cytofluorometry. Incubation of PBMC with T. gondii tachyzoites consistently resulted in a striking induction of expression of CD80 and up-regulation of expression of CD86 on monocytes. The levels of expression of these molecules followed a biphasic distribution, giving rise to a subpopulation of monocytes with remarkable levels of expression of CD80 (CD80⁺) as well as a subpopulation of monocytes with increased levels of expression of CD86 (CD86^high; see Fig. 1). The levels of expression of CD54 and HLA-DR on monocytes and of CD54, CD86, and HLA-DR on B cells were also increased in seven of 10 experiments in which PBMC were incubated with the parasite (Table I). However, the levels of expression of these latter molecules remained in a unimodal distribution. It is interesting to note that the expression of CD80 on B cells was never affected by T. gondii. Whether PBMC originated from individuals with or without Abs to T. gondii did not influence the magnitude of change in the levels of expression of any of the above mentioned molecules (p ≥ 0.2). T. gondii did not induce nonspecific up-regulation of expression of surface molecules on monocytes and B cells, since the levels of expression of CD11a and CD58 on these cells were not increased after incubation with the parasite. Tachyzoites used in these experiments were exposed to UV light to inhibit their intracellular multiplication and prevent destruction of infected cells (18). Changes in the levels of expression of accessory ligands were not caused by treatment with UV light, since unexposed tachyzoites exerted similar effects on the expression of these molecules (data not shown). Thus, T. gondii induced changes in the levels of expression of costimulatory ligands that were particularly striking on monocytes, since there was both induction of expression of CD80 and up-regulation of expression of CD86.

Given the remarkable changes observed in the expression of costimulatory ligands on monocytes and the importance of macrophages to the immune response against intracellular pathogens (25), our additional studies concentrated on the effects of T. gondii on monocytes/macrophages. Since T cell-mediated cognate (cell contact-mediated) signals can induce the expression of CD80 on B cells (26, 27), experiments were conducted to determine whether changes in the expression of costimulatory ligands on monocytes that were mediated by T. gondii occurred in the absence of lymphocytes. Incubation of purified monocytes with tachyzoites resulted in the appearance of subpopulations of monocytes with remarkable levels of expression of CD80 molecules (CD80⁺; Fig. 1,A) as well as subpopulations of cells with increased levels of expression of CD86 molecules (CD86^high; Fig. 1,B). Although the changes in expression of CD80 were striking, levels of expression of CD86 on CD86^high monocytes (MFI = 1579.2 ± 70.1) were, on the average, 11.7 times higher than those of CD80 on CD80⁺ monocytes (MFI = 139.5 ± 15.4; n = 7). The increase in the levels of expression of CD80 and CD86 were not caused by nonspecific effects mediated by phagocytosis of foreign particles, since incubation with latex beads did not alter the expression of any of these molecules (data not shown). Incubation of purified monocytes with increasing numbers of tachyzoites per monocyte resulted in a progressive increase in the percentages of CD80⁺ and CD86^high monocytes (Fig. 1 C). T. gondii tachyzoites also induced up-regulation of the expression of CD80 and CD86 molecules on monocyte-derived macrophages (data not shown).

Experiments were conducted to determine whether preparations of dead tachyzoites of T. gondii as well as of parasite soluble Ags could also induce changes in the expression of CD80 and CD86 on purified monocytes. Whereas preparations containing viable (UV-attenuated) tachyzoites induced the appearance of CD80⁺ and CD86^high monocytes, neither paraformaldehyde-killed tachyzoites nor TLA affected the levels of expression of these molecules (Fig. 1 D). Similar results were obtained regardless of whether monocytes originated from individuals with or without Abs to T. gondii.

Experiments were conducted to determine whether induction of the expression of CD80 and up-regulation of the expression of CD86 caused by T. gondii were due to infection of monocytes and/or to the secretion of cytokines known to affect the expression of these molecules on monocytes/macrophages. Purified monocytes incubated with the parasite were sorted by FACS into CD80⁺, CD80⁻, CD86^high, and CD86^int populations. Microscopic examination of these populations revealed that whereas at least 50% (59 ± 8%) of CD80⁺ and CD86^high monocytes had evidence of intracellular tachyzoites (mostly degenerated organisms), the percentages of CD80⁻ and CD86^int monocytes with intracellular tachyzoites were never more than 2% (1.5 ± 0.5%; n = 3).

Since IFN-α, IFN-γ, and GM-CSF have been reported to induce the expression of CD80 and/or to up-regulate the expression of CD86 on monocytes/macrophages (9, 10, 26, 27, 28, 29), we performed experiments to determine whether the effects of T. gondii on the levels of expression of these costimulatory ligands were mediated by these cytokines. Incubation of purified monocytes with saturating concentrations of neutralizing mAb against IFN-α or IFN-γ did not affect either the T. gondii-mediated induction of CD80⁺ and CD86^high monocytes (Fig. 2,A) or the increase in the levels of expression of CD80 and CD86 on CD80⁺ and CD86^high cells, respectively (Fig. 2, B and C; n = 3). Although a neutralizing mAb against GM-CSF did not significantly affect the percentages of CD80⁺ and CD86^high monocytes, this mAb induced a modest, but consistent, inhibition of parasite-mediated increase in the levels of expression of CD80 (30.6 ± 5% inhibition) and CD86 (18.8 ± 2% inhibition) on CD80⁺ and CD86^high monocytes, respectively (p ≤ 0.04; Fig. 2, A–C). At the concentration used, anti-GM-CSF mAb completely ablated the GM-CSF-mediated induction of expression of CD80 on monocytes (data not shown). Even though our results suggest that GM-CSF is involved in the T. gondii-mediated induction of expression of CD80 and the up-regulation of expression of CD86, experiments performed with Transwell inserts indicate that it is unlikely that soluble factors (such as GM-CSF) alone are sufficient to mediate these effects. Incubation of monocytes with T. gondii tachyzoites separated from monocytes alone by a membrane permeable to particles ≤0.4 μm in diameter resulted in induction of the expression CD80 and up-regulation of the expression of CD86 only on monocytes directly exposed to the parasite (data not shown).

The changes in levels of expression of these costimulatory ligands were not attributable to LPS, since preparations of T. gondii did not contain detectable levels of LPS (<10 pg/ml by amebocyte Limulus assays) and addition of a mAb against CD14 (LPS receptor) that neutralizes the effects of LPS on human monocytes (30) did not inhibit the effects of T. gondii on the levels of expression of these costimulatory ligands (data not shown). Taken together, these experiments indicate that infection of monocytes with viable tachyzoites of T. gondii results in induction of remarkable levels of expression of CD80 molecules and up-regulation of expression of CD86 molecules. In addition, our results also suggest that GM-CSF has a secondary role, probably by potentiating the effects of T. gondii infection on the expression of these molecules.

Since IFN-γ is consistently produced upon exposure of human PBMC to T. gondii (31), experiments were conducted to determine whether, in the presence of lymphocytes, IFN-γ modulated the T. gondii-induced changes in the expression of CD80 and CD86 on human monocytes. As shown in Figure 3, A and B, addition of anti-IFN-γ mAb to PBMC incubated with T. gondii did not result in a significant inhibition of the induction of CD80⁺ and CD86^high monocytes or a significant decrease in the levels of expression of these molecules on CD80⁺ and CD86^high monocytes (p ≥ 0.2; n = 4). Similar results were obtained regardless of the serologic status of the donor. At the concentration used, anti-IFN-γ ablated the IFN-γ-mediated induction of expression of CD80 and up-regulation of expression of CD86 on monocytes (Fig. 3 C). Thus, these results indicate that, even in the presence of lymphocytes, the early induction of CD80 and up-regulation of CD86 on human monocytes triggered by T. gondii are not mediated by IFN-γ.

The presence of costimulatory ligands during the early phases of T cell-APC interaction would appear to be required to affect T cell responses. Thus, experiments were conducted to determine the kinetics of the T. gondii-mediated changes in the levels of expression of CD80 and CD86 on purified monocytes. Figure 4,A demonstrates that a significant increase in the percentages of CD80⁺ and CD86^high monocytes was detected 12 h after incubation with T. gondii tachyzoites and reached a peak at 24 h. Similarly, when the MFI of CD80⁺ and CD86^high cells were analyzed, maximum levels of expression of these molecules on the CD80⁺ and CD86^high populations were observed at 24 h (Fig. 4,B). Thus, these data indicate that T. gondii triggers a rapid induction of expression of CD80 and up-regulation of expression of CD86 molecules on monocytes. It is interesting to note that CD80⁺ and CD86^high cells underwent a progressive decrease in the levels of expression of CD14 (Fig. 1, A and B). In contrast to dendritic cells derived by incubation of monocytes with GM-CSF plus IL-4, T. gondii-induced CD80⁺ and CD86^high monocytes remained CD1a⁻ (data not shown).

We have previously demonstrated that both peripheral blood resting T cells from healthy T. gondii chronically infected donors as well as those from T. gondii-seronegative donors proliferate when incubated with either parasite-infected cells or T. gondii-soluble Ags (18). We used this experimental system to assess the functional significance of CD80 and CD86 molecules on the response of resting T cells to T. gondii. T cells from T. gondii-seronegative donors were stimulated with autologous T. gondii-infected PBMC in the presence of CTLA-4-Ig, a chimeric protein with high affinity for CD80 and CD86 that blocks the interaction between these molecules and CD28 (19). Figure 5,A shows that CTLA-4-Ig induced a dose-dependent inhibition of the T. gondii-mediated T cell proliferation. To determine the individual roles of CD80 and CD86 in the proliferative response of T cells, anti-CD80 and/or anti-CD86 mAb were added to T cells stimulated with the parasite. Incubation with either anti-CD80 (p ≤ 0.03) as well as anti-CD86 mAb (p ≤ 0.008) resulted in a statistically significant inhibition of the T cell proliferative response to T. gondii-infected autologous PBMC (Fig. 5,B). Anti-CD80 mAb reduced the proliferative response by 30.1 ± 13.0%; anti-CD86 induced a 74.1 ± 5.4% inhibition of this response (n = 4). Furthermore, when anti-CD80 was used in combination with anti-CD86 mAb, the proliferative response was abrogated (96.0 ± 2.1% inhibition; p ≤ 0.01; Fig. 5,B; n = 4). In contrast, anti-CD86 mAb (p ≤ 0.02), but not anti-CD80 mAb (p ≥ 0.4), significantly inhibited T cell proliferation in response to TLA (Fig. 5 C). Anti-CD86 mAb induced a 76.8 ± 17.4% inhibition of T cell proliferation in response to TLA, and incubation with the combination of anti-CD80 and anti-CD86 did not result in any further inhibition (76.8 ± 19.4% inhibition; p ≤ 0.008) of TLA-mediated T cell proliferation (n = 4). Similar results were obtained when T cells were stimulated with monocytes, including T. gondii-infected highly purified CD14⁺ monocytes (n = 3; data not shown). These results demonstrate that both CD80 and CD86 are critical costimulatory ligands for T. gondii-mediated proliferation of T cells from seronegative individuals, and that whereas CD80 plays a role only in the T cell proliferation mediated by infected cells, CD86 plays a role in both proliferation mediated by cells infected with T. gondii and that mediated by cells pulsed with parasite-soluble Ags.

Costimulatory signals appear to be particularly relevant in the initiation of the immune response by naive T cells (32). We have recently demonstrated that both resting CD4⁺ CD45RA⁺ T cells from T. gondii-seronegative adults and CD45RA⁺ T cells from T. gondii-seronegative newborns proliferate in response to T. gondii.⁴ Therefore, experiments were performed to study the roles of CD80 and CD86 molecules in the response to T. gondii by CD4⁺ CD45RA⁺ T cells from seronegative donors. The proliferative response of these T cells to T. gondii-infected PBMC was partially inhibited by anti-CD80 mAb (24.2 ± 1.4% inhibition; p ≤ 0.001), was almost completely inhibited by anti-CD86 mAb (83.1 ± 6.8% inhibition; p ≤ 0.0001), and was ablated by the combination of these two mAb (98.1 ± 1.8% inhibition; p ≤ 0.0001; Fig. 5 D). Thus, these results indicate that the proliferation of presumably unprimed T cells in response to T. gondii-infected cells is dependent on both CD80 and CD86 molecules.

To determine the role of costimulation through CD28 in the response of T cells from individuals previously exposed to T. gondii, experiments similar to those described above were performed using T cells from healthy, chronically infected donors. As shown in Figure 6,A, anti-CD86 mAb significantly inhibited the proliferation of T cells in response to T. gondii-infected PBMC (58.5 ± 14.0% inhibition; p ≤ 0.02; n = 4). Incubation with anti-CD80 mAb resulted in variable partial inhibition (17.7 ± 7.4% inhibition; p = 0.01–0.1) of the T cell proliferation in response to T. gondii-infected PBMC. The combination of anti-CD80 plus anti-CD86 mAb resulted in additive inhibitory effect (84.4 ± 4.1% inhibition; p ≤ 0.01), leading to a dramatic reduction in T cell proliferation. Figure 6 B shows that anti-CD86 mAb inhibited proliferation of T cells in response to TLA (67.5 ± 11.4% inhibition; p ≤ 0.001), and there was no inhibitory effect exerted by anti-CD80 mAb (2.7 ± 2.7% inhibition; p ≥ 0.1) and no additive effect observed after combining anti-CD80 plus anti-CD86 mAbs (76.1 ± 13.3% inhibition; p ≤ 0.0003). Thus, CD80 and CD86 are also important costimulatory ligands in the in vitro proliferative response to T. gondii by T cells from previously exposed individuals.

We studied whether the T. gondii-induced changes in the levels of expression of CD80 and CD86 on monocytes modulated cytokine production by T cells in response to the parasite. Whereas there was no significant production of IFN-γ when resting T cells from T. gondii-seronegative individuals were incubated with uninfected, untreated monocytes, T cells secreted significant amounts of this cytokine when incubated with T. gondii-infected monocytes (Fig. 7,A; n = 3). Interestingly, stimulation of T cells with monocytes incubated with paraformaldehyde-killed tachyzoites or TLA did not result in the production of significant amounts of IFN-γ. This was observed despite the fact that T cells proliferated not only in response to T. gondii-infected monocytes but also in response to monocytes that had phagocytosed killed parasites and to monocytes incubated with TLA (Fig. 7 A; n = 3). The lack of production of IFN-γ in response to incubation with monocytes and either killed parasites or TLA was not caused by an inherent inability of these stimuli to trigger IFN-γ production, since T cells from healthy individuals chronically infected with T. gondii produced considerable amounts of IFN-γ after stimulation with monocytes plus either killed tachyzoites (≥945 pg/ml) or TLA (≥582 pg/ml; n = 3).

These results raised the possibility that induction of expression of CD80 and/or up-regulation of expression of CD86 were involved in the production of IFN-γ by T cells from seronegative individuals. In this regard, production of IFN-γ by these T cells in response to T. gondii-infected monocytes was significantly inhibited by anti-CD86 (40.8 ± 5.2% inhibition; p ≤ 0.04). Incubation with anti-CD80 mAb resulted in variable partial inhibition (7.5 ± 7.0% inhibition; p = 0.02–0.1) of IFN-γ production; the combination of CD80 plus anti-CD86 mAb resulted in an additive inhibitory effect (80.2 ± 1.4% inhibition; p ≤ 0.002), leading to almost complete ablation of IFN-γ production (n = 3; Fig. 7,B). Similar results were obtained using highly purified CD14⁺ monocytes (data not shown). In addition, production of IFN-γ by CD4⁺ CD45RA⁺ T cells from seronegative individuals was inhibited by either anti-CD80 (69.47 ± 6.1% inhibition; p ≤ 0.001) or anti-CD86 mAb (53.0 ± 5.1% inhibition) and was ablated by the combination of mAbs (100% inhibition p ≤ 0.001; n = 2; Fig. 7 C). Taken together, these results indicate that the early production of IFN-γ by presumably unprimed T cells is dependent on the expression of CD80 and CD86 on monocytes.

We attempted to determine whether T. gondii-mediated changes in expression of costimulatory ligands affect T cell production of IL-2 and IL-4 in response to the parasite. Stimulation of T cells with monocytes plus PHA resulted in the secretion of measurable concentrations of IL-2 and either low or undetectable concentrations of IL-4. However, neither of these cytokines was detected in supernatants from T. gondii-stimulated T cells (data not shown).

We have established that T. gondii caused a rapid induction of the expression of CD80 and up-regulation of the expression of CD86 on human monocytes. These changes were observed after infection of monocytes with viable tachyzoites and not after phagocytosis of killed tachyzoites or incubation with T. gondii-soluble Ags. These costimulatory ligands played a crucial role in the T cell response triggered by T. gondii. Not only was T cell proliferation dependent on CD80 and CD86, but, more importantly, our data indicate that changes in the levels of expression of these costimulatory ligands on monocytes were associated with the production of IFN-γ by T cells from unexposed individuals. Our results are of particular relevance to immunity against intracellular organisms, since production of IFN-γ as a result of the interaction between infected monocytes/macrophages and T cells would lead to activation of these phagocytic cells, enabling them to act as major effectors of antimicrobial defense.

It is well recognized that cognate signals and certain cytokines can induce or up-regulate the expression of CD80 and CD86 on APC. We demonstrate that induction of the expression of CD80 and up-regulation of the expression of CD86 on human monocytes caused by T. gondii occurred in the absence of significant concentrations of lymphocytes and was not mediated by IFN-α or IFN-γ, which are cytokines capable of inducing the expression of CD80 and up-regulating the expression of CD86 on monocytes/macrophages (9, 10, 29). In these same experiments, GM-CSF was found to be involved in the process of induction of the expression of CD80 and up-regulation of the expression of CD86 triggered by T. gondii. These latter results suggest that, similar to what has been reported for dendritic cells (33), GM-CSF may have immunomodulatory activity on infected monocytes that, through promoting up-regulation of costimulatory ligands, may lead to an enhanced immunostimulatory function of these APC. In addition to the cytokines mentioned above, we have recently studied TNF-α, IL-1α, and IL-12 and demonstrated that neutralizing Abs against these cytokines failed to inhibit the T. gondii-mediated changes in the expression of CD80 and CD86 on monocytes (C. S. Subauste, unpublished observations). Despite the results that we obtained after neutralization of GM-CSF, cell-sorting experiments as well as data obtained with Transwell inserts indicated that infection of monocytes with viable tachyzoites and not soluble factors appeared to play the primary role in induction of the expression of CD80 and up-regulation of the expression of CD86. Although not all the CD80⁺ and CD86^high monocytes contained intracellular tachyzoites 18 h after incubation with T. gondii, our microscopic examination of monocytes incubated with the parasite is consistent with this conclusion. Monocytes can rapidly eliminate intracellular tachyzoites so that the percentage of infected monocytes 18 to 24 h after challenge with T. gondii will be remarkably lower than the percentage of infected monocytes 1 h after challenge (34).

T. gondii-mediated up-regulation of expression of CD86 on human monocytes has been described recently (35). The authors reported that the expression of this molecule followed a unimodal distribution, and that this up-regulation appeared to be inhibited by an anti-IFN-γ polyclonal Ab. However, in our studies, double staining with anti-CD14 and anti-CD86 mAb allowed us to demonstrate that up-regulation of CD86 occurred only in a subpopulation of monocytes. Furthermore, we demonstrated that IFN-γ did not play an important role in this up-regulation, since an increase in the levels of expression of CD86 occurred in populations of highly purified monocytes and was not affected by a neutralizing anti-IFN-γ mAb.

Except in experimental models using cells transfected with CD80 or transgenic mice expressing CD80 on pancreatic β cells (36, 37), it has been difficult to demonstrate significant costimulation by CD80 molecules (12, 38). The predominant costimulatory activity of CD86 molecules may be explained, at least in part, by the fact that after APC activation, CD86 is expressed earlier and at higher levels than CD80 (39, 40). In contrast, our results demonstrate that infection of monocytes with T. gondii results in a rapid induction of CD80 and up-regulation of CD86, indicating that both costimulatory ligands can be present in the early phases of the immune response to an intracellular pathogen. Moreover, our studies indicate that CD80 molecules provided significant costimulation under conditions in which the expression of CD80 was induced in human monocytes (infection with viable tachyzoites).

One of the hallmarks of the immune response elicited by T. gondii is the production of IFN-γ. Indeed, this cytokine is a major mediator of protection against the parasite (16, 41). The critical role of IFN-γ in the immune response to intracellular pathogens makes identification of factors required for production of this cytokine one of the most important questions in the study of immunity to infectious organisms. We have recently demonstrated that both human αβ and γδ T cells from T. gondii-seronegative individuals secrete significant amounts of IFN-γ after in vitro stimulation with the parasite (18) (see Footnote 4). Our data indicate that IFN-γ production by human T cells from unexposed individuals is associated with a T. gondii-induced increase in the levels of expression of costimulatory ligands on infected monocytes. IFN-γ was produced only when T cells were incubated with monocytes that displayed high levels of expression of CD80 and CD86 (monocytes infected with viable tachyzoites). Moreover, anti-CD80 plus anti-CD86 mAb almost completely inhibited T cell production of IFN-γ. Of particular relevance to the events in the initiation of the immune response to intracellular pathogens are our results with CD4⁺ CD45RA⁺ cells, which indicate that presumably naive CD4⁺ T cells require CD80 and CD86 for the production of IFN-γ in response to T. gondii-infected monocytes. These data are of importance to the induction of protective immunity to the parasite, since the early production of IFN-γ may confer protection to the host not only because of direct effects of this cytokine on the growth of intracellular tachyzoites (41), but also because IFN-γ appears to play a role in promoting the generation of a Th1 cytokine pattern (42).

It is well established that proliferation and IL-2 secretion by naive T cells are dependent on the presence of costimulatory signals (32). Indeed, we have demonstrated that the proliferation of CD4⁺ CD45RA⁺ T cells in response to T. gondii is ablated by anti-CD80 plus anti CD86 mAb. However, our results indicate that these costimulatory ligands also play a central role in the parasite-triggered proliferation of T cells from chronically infected individuals. In this regard, human CD45RO⁺ memory T cells undergo optimal anti-CD3-mediated proliferation in the presence of CD80-transfected cells (43).

Our model provides a clear example of the capacity of monocytes to discriminate among different microbial preparations and illustrates that the T cell cytokine response is affected by the type of microbial preparation that elicits the immune response. In the case of T. gondii, this pathogen-monocyte-T cell interaction would result in IFN-γ production in situations (noxious stimulus; i.e., infection with viable tachyzoites) where an IFN-γ-dependent cell-mediated response would be appropriate, whereas no such a response would be triggered when encountering nonviable parasite preparations (harmless stimulus). A probable in vivo correlate to our observations can be drawn from the demonstration that infection with viable T. gondii bradyzoites, rather than immunization with TLA, is necessary for acquisition of resistance to tachyzoites of a virulent strain of the parasite (44).

There is increasing evidence of the importance of innate immunity in host defense against intracellular organisms (45, 46). It is well established that this arm of the immune system can promote the generation of a protective immune response through IFN-γ production by NK cells (47). It has also been proposed that the mechanisms used by the innate immune system to recognize pathogens would determine the type of adaptive immunity elicited (46). Our results support this proposal, since they provide evidence that nonclonal recognition of a micro-organism by monocytes can affect the nature of the T cell response to the offending pathogen. In addition, given that IFN-γ production by NK cells can be enhanced by costimulation through CD28 (48, 49), it is likely that microbially mediated induction/up-regulation of costimulatory ligands can further promote the generation of protective immunity through stimulation of secretion of IFN-γ by NK cells.

The results of our present study illustrate the importance of a microbially induced increase in the levels of expression of costimulatory ligands for the production of protective cytokines (i.e., IFN-γ). Of interest in this regard is the evidence that certain intracellular organisms, such as Leishmania donovani and Mycobacterium tuberculosis, either fail to up-regulate or actually decrease the expression of these molecules (50, 51). These effects on costimulatory ligands may represent strategies used by the pathogens to avoid recognition, induce anergy, or cause immunosuppression. Thus, intracellular organisms can have a major influence on antimicrobial immunity through regulation of costimulation.

Our demonstration that tachyzoites, but not parasite-soluble Ags, are capable of inducing these changes and that the tachyzoites have to be viable to affect the levels of expression of CD80 and CD86 on monocytes suggest that the process of induction of CD80 and up-regulation of CD86 in monocytes may be multifactorial. Signals triggered within monocytes by the presence of viable intracellular organisms, factors released by viable intracellular parasites and/or interactions between the intact pathogen and molecules on the surface of monocytes may be involved in this process. Identification of the mechanisms by which microbes induce the expression of CD80 and up-regulate the expression of CD86 are of paramount importance, since they may tie molecular events that occur during the interaction of pathogens with APC with the outcome of infections. Comparison of the molecular events triggered in monocytes by infection with T. gondii with those triggered by pathogens that fail to increase levels of expression of costimulatory ligands may provide an understanding of the mechanisms by which monocytes can direct T cell responses. Finally, our results have important implications for the efforts to establish a vaccine against intracellular pathogens, since they suggest that vectors that induce/up-regulate the expression of costimulatory ligands on APC will help generate a protective (IFN-γ-dependent) immune response.

We express our appreciation to J. Remington for generous support and for critical review of the manuscript. We thank G. Deepe, Jr., and L. Lanier for helpful suggestions, and P. Linsley and M. Brenner for providing reagents.