Both Expansion of Regulatory GR1⁺ CD11b⁺ Myeloid Cells and Anergy of T Lymphocytes Participate in Hyporesponsiveness of the Lung-Associated Immune System during Acute Toxoplasmosis

Abstract

Oral infection with Toxoplasma gondii leads to transient systemic hyporesponsiveness. In this report, we characterized the presence in the lungs of GR1⁺ CD11b⁺ myeloid cells that have potent nitric oxide-dependent immunoregulatory properties. We also demonstrated the interleukin 2-reversible anergy of both pulmonary CD8⁺ and CD4⁺ activated T lymphocytes with infection.

Suppression of T-lymphocyte responses occurs in many infectious diseases with viruses (12, 28), bacteria (7, 22), or parasites (1, 24) and may account for evasion of the host immune system and persistence of the microorganisms during the life of the host. Activation of various immunoregulatory cells (3, 13, 29) had been demonstrated in these pathological events. Recently, a specific myeloid immunoregulatory cell population, named natural suppressive cells, exhibiting a GR1⁺ CD11b⁺ phenotype, was described in tumor-bearing rodents (17) and was recruited during Chagas' disease (9). This population has been actively involved in acute hyporesponsiveness of T and B cells throughout the synthesis of a large amount of nitric oxide (NO) at the place of infection. Acute infection with Toxoplasma gondii, an intracellular protozoan, induces a transient hyporesponsiveness (3 weeks), reported both in humans (19) and in mice (5, 16), of blood and spleen leukocyte response, respectively, when stimulated in vitro with specific T. gondii antigens (TAg) or with mitogenic molecules. A previous study has shown the existence of unresponsiveness in the lungs during acute toxoplasmosis for intraperitoneally infected mice (26). Characterization of the regulatory cells triggered during acute toxoplasmosis has not been completed. The present study focuses on the defective response of pulmonary T lymphocytes following oral infection of mice with T. gondii.

Infection of C57BL/6 mice by intragastric delivery of T. gondii 76K strain cysts (15 cysts/mouse, from a CBA/J mouse brain homogenate infected for 1 month) leads to a 2.3-fold increase of extravascular leukocytes in the lungs at day 11 postinfection (Table 1). Primed and unprimed pulmonary leukocytes isolated from infected and naive mice, respectively, were separated into two different subsets according to their plastic adherence properties. This result is correlated with an increase of adherent cells (4.3-fold) and to a lesser extent nonadherent cells (1.7-fold). Both alveolar macrophages and interstitial dendritic cells that compose the adherent cell fraction of pulmonary leukocytes (10, 11, 18) are potent immunosuppressive cells (2, 32). We consequently investigated the phenotypes of primed and unprimed adherent cells with triple-staining flow cytometry analysis (Fig. 1A, left panel). Most of the unprimed adherent cells (up to 70%) exhibited a typical alveolar macrophage phenotype: Mac3^hi CD11c⁺ CD11b⁻ GR1⁻ CD8α⁻. They also included almost 15% lymphoid dendritic cells (Mac3⁻ CD11c⁺ CD11b⁻ GR1⁻ CD8α⁺) and an average 4% myeloid monocyte-like cells (Mac3^low CD11c⁻ CD11b⁺ GR1⁺ CD8 α⁻). After infection, the adherent population in the lungs dramatically changed. The pulmonary Mac3^low CD11c⁻ CD11b⁺ GR1⁺ myeloid cell population was highly expanded, representing an average 70% of the adherent population (a 17.5-fold increase). Adherent primed cells also contained an undefined population expressing a CD11b⁺ CD8α⁺ GR1⁺ CD11c⁻ phenotype (10%) and a few T cells (average of 8% CD3⁺ CD11b⁻ GR1⁻ cells). Infection of mice was also associated with an increase of T lymphocytes in the nonadherent cell fraction (Fig. 1A, right panel). Despite the presence of a few nonadherent macrophages (Mac3⁺), the amount of CD3⁺ T lymphocytes was dramatically increased (4.2-fold in absolute number) over that of unprimed counterparts. The absolute numbers of CD4⁺, CD8⁺, and γδ T lymphocytes increased 9.8-, 2-, and 5.5-fold, respectively, with infection.

TABLE 1.

Increase of pulmonary cell number with infection (day 11 postinfection)

Leukocyte group	No. of cellsa
	Unprimed (106)	Primed (106)	Ratio (primed/unprimed)
Total	12.9 ± 4.0	32.2 ± 9.8b	2.8 ± 0.7
Adherent	0.6 ± 0.4	2.5 ± 1.8c	4.3 ± 1.7
Nonadherent	9.1 ± 3.7	15.3 ± 7.3d	1.7 ± 0.4

^aValues are mean ± standard deviation for seven experiments with at least four mice per group.

^bSignificantly different (P < 0.001) from the value for unprimed cells as calculated by the Mann-Whitney test.

^cSignificantly different (P < 0.004) from the value for unprimed cells as calculated by the Mann-Whitney test.

^dSignificantly different (P < 0.03) from the value for unprimed cells as calculated by the Mann-Whitney test.

FIG. 1.

(A) Phenotype of pulmonary cells from naive or infected mice (unprimed or primed cells, respectively) was determined by flow cytometry. Adherent cell fractions (left panel) obtained after 90 min of plastic adherence and removal of nonadherent fractions were triple stained with the specific anti-CD11c/anti-GR1/anti-CD11b, anti-CD11c/anti-Mac3/anti-CD11b, anti-CD8α/anti-GR1/anti-CD11b, or anti-CD3/anti-GR1/anti-CD11b antibodies (fluorescein isothiocyanate-, phycoerythrin-, or TriColor-labeled antibodies, respectively). Isotypic antibodies were used as a background control of unspecific labeling. The nonadherent (right panel) cell fractions were stained with phycoerythrin-labeled anti-Mac-3 or fluorescein isothiocyanate-labeled specific anti-CD3, anti-CD4, anti-CD8, or anti-γδ T-cell receptor antibodies (thick line) or with isotypic controls (line with shaded area). Dead cells were first excluded from analysis by viability staining using 7-aminoactinomycin D. Percentages of the population (#) represent one of three separate experiments with the same instrument settings for each group (n = 3 to 5). F.I., fluorescence intensity. (B) Proliferative responses of pulmonary nonadherent and adherent cells (5 × 10⁴) were determined by [³H]thymidine incorporation after a 3-day culture period upon stimulation by ConA or TAg. Cells were then harvested, and proliferation was measured with a scintillation counter. Results are means ± standard deviations (n = 3 to 5 mice), representative of triplicate wells per condition for each experiment, repeated there times. *, P ≤ 0.0001; **, P ≤ 0.001; ***, P ≤ 0.01 (primed versus unprimed values for the same condition [alone, ConA, or TAg]). (C) Unprimed and primed nonadherent or adherent viable cells (5 × 10⁴) from lungs were cocultured in vitro for 2 days with freshly isolated syngeneic lymphocyte-enriched fraction (5 × 10⁴) from spleen tissue of naive mice. [³H]thymidine was then added for another 18-h period. Results are representative of more than three separate experiments with means ± standard deviations (n = 3 to 5 mice). *, P ≤ 0.001 (primed versus unprimed from the same condition [alone or with ConA]).

Proliferation of subpopulation leukocytes of the lungs was assayed by [³H]thymidine incorporation during in vitro culture in the presence of concanavaline A (ConA), TAg, or medium alone during 3 days. Neither primed nonadherent nor adherent purified cells (Fig. 1B) proliferated with ConA or TAg stimulation. Only unprimed nonadherent cells were able to respond to mitogenic stimulation. The unresponsiveness of leukocytes suggested the presence of regulatory cells in the lungs during acute infection. To determine which cell population exhibited such an immunoregulatory activity, nonadherent or adherent pulmonary cells were coincubated in vitro with the unprimed lymphocyte-enriched fraction as responding cells, recovered from spleens of syngeneic mice. Their proliferation was also evaluated by [³H]thymidine incorporation after ConA stimulation. Interestingly, only primed adherent cells inhibited the proliferative response of unprimed responding lymphocytes to mitogenic stimulation (Fig. 1C). Regulatory properties of adherent cells were not major histocompatibility complex restricted, since these cells also suppressed the proliferation of allogeneic unprimed or mouse-immunized lymphocytes (data not shown). Unprimed responding lymphocytes cocultured with either primed or unprimed nonadherent lung population cells proliferated when stimulated with ConA (Fig. 1C). Purified primed CD4⁺ or primed CD8⁺ T pulmonary cells did not inhibit the proliferation of unprimed responding lymphocytes (data not shown). Moreover, inhibition of responding lymphocyte proliferation is independent of the multiplication of the parasite, since PCR analysis showed amplification of the T. gondii-specific B1 gene both in primed adherent pulmonary cells and in primed nonadherent pulmonary cells (data not shown).

Consistent with the fact that after infection GR1⁺ CD11b⁺ myeloid cells became the predominant population of the adherent cell fraction, we investigated the chemokine receptor profile of these cells. GR1⁺ CD11b⁺ myeloid cells from infected mice expressed higher levels of mRNA for CCR1, CCR5, and to a lesser extent CCR2 than alveolar macrophages from naive animals (Fig. 2A). These results are consistent with the expression by pulmonary leukocytes of inflammatory chemokines, such as RANTES, macrophage inflammatory protein 1 alpha, and macrophage inflammatory protein 1 beta, the ligands of CCR1 and CCR5, and the CCR2 ligand, monocyte chemoattractant protein 1 (data not shown). These observations suggest that GR1⁺ CD11b⁺ cells might be recruited into inflammatory tissues during acute infection. Recent studies described the expansion of these suppressive cells in the peritoneal cavity in response to the injection of oligosaccharides from Schistosoma mansoni (4, 30). The mRNA analysis of unprimed and primed adherent cells showed a significant increase in CD80 and CD40 mRNA expression but not in CD86 mRNA expression (Fig. 2B). Expression of CD95 and CD95L mRNA (Fig. 2B) was enhanced in primed GR1⁺ CD11b⁺ cells compared with that in unprimed alveolar macrophages.

FIG. 2.

Chemokine expression (A) was assayed with 5 μg of total RNA from naive (five mice) or primed (three mice) adherent cells with RNase protection assay using the mCR5 Riboquant template set, including CCR1, CCR1b, CCR2, CCR3, CCR4, CCR5, L32, and glyceraldehyde-3-phosphate-dehydrogenase (GAPDH). Specific mRNA expression was normalized to housekeeping gene expression (L32 plus GAPDH) and is expressed in relative densitometry values. CD80, CD86, CD40, Fas, FasL, and tumor necrosis factor alpha (TNF-α) expression (B) was detected by reverse transcription-PCR analysis of 1 μg of total RNA. mRNA levels were standardized to β-actin expression as an internal control. Data are representative of three independent experiments with similar results.

Despite potential costimulatory capacity, primed GR1⁺ CD11b⁺ adherent cells induced hyporesponsiveness of unprimed responding lymphocytes in vitro (Fig. 1C). Primed adherent cells showed an mRNA increase for gamma interferon (IFN-γ), inducible NO synthetase, and interleukin 10 (IL-10) over expression by unprimed adherent cells (not shown). These results correlated with the secretion of IFN-γ, IL-10 (enzyme-linked immunosorbent assay [ELISA], OptEIA; Pharmingen), and NO (Griess assay method) components from primed and unprimed adherent cells (10⁶ cells) after 48 h of in vitro culture (Fig. 3A). Primed adherent cells indeed secreted larger amount of IFN-γ, IL-10, and NO than unprimed counterparts. NO synthesis has a dichotomous role in toxoplasmosis (15). It participates in eradication of the parasites from infected macrophages but also plays an important role in splenic hyporesponsiveness observed during acute infection (23). We thus investigated the mechanism of immunoregulation. Primed or unprimed adherent cells were coincubated with unprimed responding lymphocytes in the presence of inhibitors, such as the inducible NO synthetase inhibitor N^G-monomethyl-l-arginine (l-NMMA) (Sigma) or Boc-d-FMK (pan-caspases inhibitor; Calbiochem), or in the presence of recombinant murine IL-2 (rmIL-2) (Sigma). l-NMMA completely abolished the inhibitory effect of adherent primed cells, since the rate of [³H]thymidine incorporation under this ConA-stimulating condition was not significantly different from that with responding lymphocytes alone (Fig. 3B). This result suggests that NO synthesis by GR1⁺ CD11b⁺ cells is responsible for the hyporesponsiveness. Neither the Boc-d-FMK inhibitor nor rmIL-2 suppressed the regulatory activity of GR1⁺ CD11b⁺ cells (data not shown). Geissmann and colleagues have recently shown that murine GR1⁺ CD11b⁺ blood monocytes migrated into the peritoneal cavity after thioglycolate injection (8) and differentiated into dendritic cells, triggering T-cell activation. We can thus speculate that GR1⁺ CD11b⁺ cells might have different properties according to the nature of the stimulation.

FIG. 3.

(A) Unprimed and primed adherent cells (10⁶) were incubated in vitro for 48 h in complete medium, and supernatants were collected. Determination of IFN-γ and IL-10 secretion into supernatants was assessed by ELISA, and NO products were evaluated by the Griess method. Data are mean values ± standard deviations for three separate experiments. *, P ≤ 0.0001; **, P ≤ 0.05 (primed versus unprimed values). (B) Responding spleen lymphocytes from naive mice were coincubated in vitro with unprimed or primed adherent pulmonary cells (n = 3 to 5 mice per group during 3 days in medium alone or supplemented with ConA and in the presence or not of NO synthesis inhibitor l-NMMA (1 mM; Sigma). Cells were tested for [³H]thymidine incorporation (for the last 18 h) to evaluate their proliferation. Data are mean values ± standard deviations for results representative of three experiments with similar results. *, P ≤ 0.002 (primed versus unprimed from the same condition [alone or with ConA]).

We further evaluated the ability of pulmonary leukocytes and purified CD4⁺ and CD8⁺ T lymphocytes to secrete IFN-γ, IL-2, and IL-10 cytokines and the NO molecule after a 48-h in vitro culture period. IFN-γ synthesis from total primed leukocytes was highly enhanced compared with that from unprimed cells, especially from purified CD4⁺ T lymphocytes, which secreted the largest amount of this cytokine (Fig. 4A). These primed cells also exhibited greater production of IL-10 than unprimed cells. However, total primed leukocytes produced larger quantities of IL-10 than primed CD4⁺ T lymphocytes. Primed pulmonary leukocytes released the NO molecule, as determined by the dosage of nitrites in supernatants, whereas unprimed cells did not. Purified CD4⁺ T lymphocytes also produced nitrites. Conversely, CD8⁺ T lymphocytes generated neither IFN-γ nor IL-10 nor NO synthesis, suggesting that these cells were not implicated in the production of proinflammatory molecules (IFN-γ and NO) or regulatory cytokines (IL-10). Interestingly, IL-2 synthesis was profoundly altered. Primed leukocytes secreted a small amount of IL-2, and primed CD4⁺ T cells did not synthesize much more of this cytokine than unprimed counterparts. Surprisingly, primed CD8⁺ T cells secreted low but significant (P ≤ 0.005) level of IL-2. A defect in IL-2 synthesis has been previously observed in splenocytes (14) and was independent of NO synthesis (6). We also observed by flow cytometry that primed CD4⁺ and CD8⁺ cells had an activated phenotype expressing both CD25 and CD69 molecules with the infection (data not shown) compared with unprimed purified cells. We demonstrated that despite removal of the immunoregulatory GR1⁺ CD11b⁺ population, neither nonadherent primed lymphocytes nor microbead-purified primed CD4⁺ and CD8⁺ cells responded to ConA stimulation in an [³H]thymidine incorporation assay in vitro (Fig. 4B). The lack of proliferation of purified CD4⁺ and CD8⁺ T cells following mitogenic stimulation and the defect in IL-2 synthesis but not in that of its receptor subunit CD25 might suggest that lymphocytes were induced in anergy in vivo. To address this question, primed and unprimed CD4⁺ and CD8⁺ lymphocytes were purified and incubated in complete medium alone or stimulated with ConA in the presence or not of a large amount (30 U/ml) of exogenous rmIL-2 (Fig. 4B). We observed that unprimed CD4⁺ and CD8⁺ lymphocytes responded to IL-2 stimulation, since they significantly proliferated in medium alone compared with IL-2-untreated cells. In the presence of ConA stimulation, unprimed CD4⁺ T lymphocytes proliferated equally with or without exogenous IL-2. ConA stimulation has previously been involved in the proliferation of purified T lymphocytes through the activation of IL-2 secretion (6, 27). Unprimed CD8⁺ T lymphocytes responded efficiently only when both ConA and exogenous IL-2 were added to the culture medium, since they did proliferate maximally with these two stimulating molecules compared to results for IL-2-untreated unprimed counterparts. Our results also demonstrated that exogenous IL-2 abolished hyporesponsiveness of primed CD4⁺ and CD8⁺ T lymphocytes (Fig. 4B), suggesting that infection with T. gondii induced an IL-2-reversible anergy of these cells. Both lymphocytes indeed merely proliferated in medium alone in the presence of a large amount of recombinant murine IL-2. However, when rmIL-2 was added to ConA-supplemented medium, both CD4⁺ and CD8⁺ purified T cells from infected mice proliferated somewhat compared to IL-2-untreated but ConA-treated counterparts. Consistent with their activated phenotype, we can speculate that addition of both IL-2 and ConA induced inhibitory signals that may trigger lymphocyte activation-induced cell death, consistent with other infection models (20-21, 25). This hypothesis correlates with our unpublished observations showing an increase in cell death among primed pulmonary leukocytes incubated alone for 48 h in vitro compared to apoptosis within the unprimed cell population, expecially in the presence of ConA.

FIG. 4.

(A) Pulmonary leukocytes, CD4⁺ and CD8⁺ microbead-purified T lymphocytes (MACS; Myltenyi Biotec) from naive or D11-infected mice, were incubated (10⁶ cells per well) for 48 h in vitro in complete medium. Supernatants were then assayed for IFN-γ, IL-2, and IL-10 secretion by ELISA. NO production was titrated by the Griess experimental procedure. *, P ≤ 0.005 (primed versus unprimed values). Results are representative mean values from more than three separate experiments with three to five mice per group. (B) Unprimed and primed CD4⁺ and CD8⁺ purified T lymphocytes were tested in vitro to evaluate their proliferative response under ConA stimulation using an [³H]thymidine incorporation assay. Data are means ± standard deviations for triplicate results; *, P ≤ 0.0001; **, P ≤ 0.001; ***, P ≤ 0.01 (primed versus unprimed values for the same conditions [alone or ConA]).

Moreover, NO secretion by primed CD4⁺ T cell also might be involved in lymphocyte activation-induced cell death, as previously described (31). Taken together, our results revealed two mechanisms leading to hyporesponsiveness within the lung during acute toxoplasmosis: an IL-2-reversible anergy of T lymphocytes and the expansion of the unique GR1⁺ CD11b⁺ regulatory cells.