Human Neutrophil-Expressed CD28 Interacts with Macrophage B7 to Induce Phosphatidylinositol 3-Kinase-Dependent IFN-γ Secretion and Restriction of Leishmania Growth¹

Introduction of foreign materials, including infectious agents, into tissue sites elicits inflammatory reactions that are initially characterized by infiltration of cells, chiefly neutrophils (1). At the site of inflammation neutrophils perform many important functions, such as control of bacterial and parasitic infections, by releasing inflammatory mediators and different cytokines (2, 3). In experimental Leishmania donovani infection, for example, the anti-leishmanial immune response begins with the release of chemokines (4) and the elimination of neutrophils increases the parasite load in hepatic tissue (5). Although these findings suggest that neutrophil migration to the site of L. donovani infection may play an important role in controlling the anti-leishmanial immune response, the stimuli required for cytokine secretion by neutrophils and their mechanism of anti-leishmanial action remain to be established.

It was reported earlier that neutrophils clear Leishmania infection by virtue of their phagocytic function (6). Because Leishmania is a protozoan parasite that resides and replicates within the macrophages, the parasite clearance by phagocytosis alone as described in previous studies limits the role of neutrophils in Leishmania infection. It was proposed recently that neutrophils may interact with macrophages through cell surface molecules like LFA-1 during infection and inflammatory reactions (7). The macrophage surface molecules, CD80 and CD86, the receptors for CD28/CD152 (CTLA-4) on T cells, were shown to play important roles in T cell response against both Leishmania major and L. donovani infections (8, 9). Because we have already demonstrated CD28 expression on human peripheral blood neutrophils and CD28-regulated neutrophil migration in response to IL-8 (10), we investigated whether neutrophil-expressed CD28 interacts with CD80/CD86 on macrophages, resulting in alteration of cellular functions.

In this work we demonstrate that L. major infection of macrophages results in increased IL-8 secretion and IL-8-directed migration of neutrophil toward the L. major-infected macrophages. The neutrophils interact with the macrophages through the CD28-CD80/CD86 pathway, resulting in phosphatidylinositol 3-kinase (PI3-kinase)³-dependent induction and secretion of IFN-γ, which in turn controls L. major growth in the macrophages. In addition, IFN-γ induction in neutrophils does not require a primary signal as in T cells or NK cells. Therefore, taken together, our results suggest that the neutrophil-expressed CD28 serves an anti-leishmanial function by interacting with the Leishmania-infected macrophages.

Following approval from the Institutional Ethics Committee, venous blood from healthy volunteers was collected in heparin and cells were isolated from the blood using Polymorphoprep (Nycomed, Oslo, Norway) as described previously (10). PBMCs were plated at a final concentration of 2 × 10⁶ cells/ml in RPMI 1640 supplemented with 10% FCS. Nonadherent lymphocytes were washed after 12 h. Monocytes were allowed to differentiate to macrophages for 72 h with washing every 24 h to remove the nonadherent cells. The cells recovered after the end of culture were >99% macrophages as judged by morphology, histochemistry, and FACS analysis (data not shown). Characterization of the neutrophils by morphology and histochemistry showed that >99.5% of the isolated cells were polymorphonuclear and that no CD3⁺ T cells were detectable (10). Cells were cultured in RPMI 1640 (Life Technologies, Grand Island, NY) with 10% FCS.

CD4⁺ T cells were isolated from the PBMC fraction by affinity chromatography using the IsoCell human T cell isolation kit (Pierce, Rockford, IL), and the purity (>99.5%) was confirmed by FACS analysis of the anti-CD3-FITC-stained cells in FACSVantage (BD Biosciences, Mountain View, CA).

Macrophages were infected with different doses of L. major promastigote (strain MHOM/SU/73/5ASKH). Extracellular parasites were washed after 4 h. The culture supernatant collected at different intervals after infection was assayed for IL-8 content by ELISA using a human IL-8 ELISA kit (Amersham Life Sciences, Little Chalfont, U.K.) following the manufacturer’s instructions.

Human monocyte-derived macrophages were infected with L. major at a 1:5 ratio for 4 h and were cocultured with neutrophils at a 1:4 ratio in the presence of neutralizing anti-IFN-γ Ab, CTLA-4-Ig (R&D Systems, Minneapolis, MN), control human Ig, or other treatments, as indicated, for 36 h. The intracellular parasite load was determined after staining the cells with Giemsa stain. Five hundred cells per culture were counted. The percentage of infection and number of parasites per 100 infected macrophages were calculated.

Neutrophil transmigration assay was performed using 24-well transwell inserts of 3-μm pore size (Falcon; BD Labware, Franklin Lakes, NJ), as described earlier (10). In brief, uninfected and L. major-infected macrophages were cultured in 24-well plates for 12 h. In some experiments, neutralizing anti-IL-8 Ab or control isotype (BD PharMingen, San Diego, CA) (10 μg/ml) were added to the lower well 30 min before placing the transwell insert. A total of 10⁵ neutrophils, pretreated with anti-IL-8R A and anti-IL-8R B or control isotype (BD PharMingen) at 4°C for 30 min, were added to each insert and were allowed to migrate for 60 min. The filter was stained with Giemsa and cells in the lower surface were counted by randomly selecting 10 fields at ×40 magnification (10).

CD28 on neutrophils was cross-linked using anti-CD28 Ab or control isotype (BD PharMingen) as described earlier (10), and the cells were cultured in the presence of indicated concentrations of wortmannin (ICN Biomedicals, Costa Mesa, CA) or cyclosporin A (Sigma-Aldrich, St. Louis, MO). The culture supernatant was assayed for IFN-γ content by ELISA using the OptEIA kit (BD PharMingen) following the manufacturer’s recommendations. CD4⁺ T cells were cultured either in the presence of only anti-CD28 (5 μg/ml) or along with immobilized anti-CD3 Ab (1 μg/ml). Supernatant was collected at different time points and ELISA was performed as described above. In parallel experiments neutrophils were also stimulated either with CD28 alone or with anti-CD3 and anti-CD28 mAbs. To compare the effect of IL-12 on IFN-γ induction by neutrophils and T cells, cells were stimulated with indicated concentrations of IL-12 and anti-CD28 mAbs. Culture supernatant was collected at different time points and ELISA was performed using the OptEIA kit (BD PharMingen) following the manufacturer’s recommendations.

Human monocyte-derived macrophages were infected with L. major at a 1:5 ratio. Six hours after infection, the cells were incubated with anti-human CD80-PE, anti-human CD86-FITC, and isotype-matched Abs (BD PharMingen) for 30 min in staining buffer (PBS containing 2% FCS and 0.1% sodium azide) at 4°C. This was followed by two washes in the staining buffer. The samples were analyzed using a FACSVantage flow cytometer.

For intracellular staining, the wells in a 96-well tissue culture plate were coated with anti-CD28 mAb (10). Neutrophils were plated to cross-link with the plate-bound anti-CD28 Ab (10). Neutrophils were recovered from the wells, fixed and permeabilized using Cytofix/Cytoperm Plus kit (BD PharMingen), and stained for IFN-γ with FITC-labeled anti-IFN-γ Ab (BD PharMingen). An isotype-matched FITC-conjugated Ab was used as isotype control. The samples were analyzed using a FACSVantage flow cytometer.

Using TRIzol (Life Technologies), total RNA was isolated from neutrophils after 4 h of incubation with medium alone or stimulation with anti-CD28 mAb or isotype-matched Ab in the presence or absence of wortmannin (10 ng/ml). For cDNA synthesis, 1 μg of total RNA from each sample was incubated with random primer, 0.1 M DTT, 500 μM dNTPs, 40 U RNase inhibitor, and 1 μl (400 U) of Moloney murine leukemia virus reverse transcriptase (Life Technologies). Samples were then incubated at 37°C for 1 h followed by a 5-min incubation at 95°C. Amplification of synthesized cDNA from each sample was conducted with DNAzyme DNA polymerase (Finnzymer, Espoo, Finland) in 50 μl under following conditions: 95°C for 2 min, 94°C 1 min, 55°C for 1 min, and 72°C for 1 min for a total of 35 cycles. Specific primers were designed to amplify the human IFN-γ coding region (sense, 5′-GCAGGTCATTCAGATGTAGC-3′; antisense, 5′-GGCCCCTGAGATAAAGCC-3′). Each sample was also amplified for either GAPDH or human β-actin to ensure equal input of cDNA.

Purified neutrophils were stimulated with immobilized anti-CD28 Ab (10 μg/ml) for 10 min at 37°C. The reaction was stopped by adding Triton X-100 lysis buffer (pH 7.5) containing 1 mM PMSF, 1 μg/ml aprotinin, 1 μg/ml pepstatin, and 1 μg/ml leupeptin for 1 h at 4°C. The lysates were centrifuged at 10,000 × g at 4°C for 30 min. Supernatants were precleared with 5 μg of mouse IgG and 30 μl of protein G-Sepharose (Life Technologies) and were incubated with 5 μg monoclonal anti-CD28 Ab overnight at 4°C. Immune complexes were recovered by incubation with 30 μl protein G-Sepharose for 1 h at 4°C. The beads were washed five times with the lysis buffer and the immune complexes were resuspended in reducing Laemmli sample buffer (Sigma-Aldrich). The samples were heated to 95°C for 3 min and run on 10% SDS-PAGE. The separated proteins were transferred onto a nitrocellulose membrane. The membrane was incubated with blocking buffer containing 2% BSA and 0.1% Tween 20 in TBS (100 mM Tris-HCl (pH 7.5), 0.9% NaCl) for 1 h at room temperature, followed by incubation with rabbit anti-PI3-kinase polyclonal Ab at a 1/500 dilution in the same buffer containing 1% BSA and 0.1% Tween for 1 h. After washing, the blot was incubated for 1 h with HRP-conjugated anti-rabbit Ab at 1/2000 dilutions. Immunoblot signal was detected by using the ECL kit (Amersham Pharmacia Biotech, Little Chalfont, U.K.) following the manufacturer’s instructions.

Confocal microscopy of neutrophils for checking IFN-γ production was performed following the methods of Yeaman et al. (11). Highly purified neutrophils were cross-linked with anti-CD28 or isotype control Abs in the presence or absence of wortmannin (10 ng/ml) for 5 h. The culture was continued for an additional 4 h with brefeldin A. The cells were then stained with 1 μg of FITC-labeled anti-human IFN-γ Ab. After counterstaining the nuclei with propidium iodide, cells were examined with a Zeiss LSM 510 confocal microscope equipped with argon and helium lasers (Zeiss, Jena, Germany). Laser power, photomultiplier tube gains, and confocal thresholds were set using long-pass and band-pass filters. The physical parameters defined as above were kept constant throughout the acquisition and analysis of the samples.

The raw scores were analyzed using SigmaStat software program (Binary Semantics, Pune, India). The data are presented as mean ± SD of triplicates in one representative of at least three individual experiments. The fiducial limit for considering any difference of mean as significant is fixed at 0.01. The correlation coefficient for the association between parasite dose and IL-8 secretion was determined by Pearson’s r test.

Because Leishmania infection is known to induce neutrophil infiltration to the site of infection (4), we investigated whether L. major-infected macrophages secreted IL-8, a major neutrophil-activating chemokine. Whereas infection of macrophages with L. major promastigotes at ratios of 1:5, 1:10, and 1:20 resulted in the peak IL-8 secretion ∼6 h after infection, infection at a 1:1 ratio yielded the peak IL-8 concentration 24 h after infection (Fig. 1 A). With 1:1 and 1:5 macrophage:Leishmania infection ratios, the level of IL-8 in the macrophage culture supernatant was undetectable 3 h after infection. At higher macrophage:parasite ratios, the chemokine was detectable after 3 h but was still significantly less than the value observed 6 h after Leishmania infection (data not shown). Because the peak IL-8 concentrations 6 h after infection were directly proportional to the load of infection (correlation coefficient = +0.954), the IL-8 secretion by L. major-infected macrophages was parasite dose dependent. Thus, our data are consistent with a previous observation describing the appearance of IL-8 mRNA by RT-PCR 1 h after L. major infection (12).

In a transwell plate, where macrophages were cultured in the lower chamber with or without L. major infection, it was found that more neutrophils migrated to the wells containing L. major-infected macrophages than the uninfected macrophages (Fig. 1,B). The neutrophil migration was significantly inhibited by both anti-IL-8 (Fig. 1,B; p < 0.01) and anti-IL-8R (Fig. 1 C; p < 0.001) Abs, suggesting that IL-8 secreted by the L. major-infected macrophages chemoattract neutrophils toward the site of infection. However, the incomplete inhibition of neutrophil migration in both anti-IL-8 and anti-IL-8R blockade could be due to chemoattraction effected by other chemokines released by the infected macrophages and IL-8R recycling to the surface, as the migration assay was performed at 37°C. Nevertheless, 65% inhibition of neutrophil migration by IL-8 neutralization or by blocking IL-8Rs identified IL-8 as the major neutrophil chemoattractant released by the L. major-infected macrophages.

It is known that T cell-expressed CD28 interacts with CD80 or CD86, also called B7.1 and B7.2, respectively, to generate the CD28 signal in T cells (13). Likewise, neutrophil-expressed CD28 may also interact with CD80/CD86 on macrophages. To test that hypothesis, we cocultured neutrophils with L. major-infected macrophages in the presence or absence of CTLA4-Ig, a fusion protein that blocks the CD28-CD80/CD86 interaction (14). The culture containing CTLA4-Ig had higher number of parasites (Fig. 2, A and B) and less IFN-γ (Fig. 2 C), suggesting that the blockade of CD28-CD80/CD86 interaction resulted in the inhibition of the neutrophils’ anti-leishmanial function. The inhibition of the leishmanicidal effect of neutrophils by anti-IFN-γ Ab suggested that the CD28 signaling in neutrophils resulted in the secretion of IFN-γ, which mediated the anti-leishmanial function of neutrophils. In addition, the supernatant from CD28-cross-linked neutrophil culture restricted the growth of L. major in macrophages, and neutralization of IFN-γ in the supernatant reduced the anti-leishmanial effect of the supernatant significantly (data not shown).

We previously showed that the chronic Leishmania infection of murine macrophages resulted in down-regulation of the ligands for CD28 (9). Therefore, we tested CD80 and CD86 expression on L. major-infected macrophages. It was observed that 6 h after L. major infection CD86 expression was marginally increased, whereas CD80 expression remained unaltered (Fig. 2 D). Thus, the profile of CD80 and CD86 expression early after L. major infection validates our observation that the neutrophils and macrophages interact through CD28-CD80/CD86 to express the IFN-γ-dependent anti-leishmanial function.

It has been argued that neutrophils, being active phagocytes, control Leishmania infection by phagocytosis (6). In this study, phagocytosis of Leishmania was not a significant factor, because neutrophils were added to the culture after extracellular parasites were washed off following the infection of macrophages with Leishmania. We also investigated whether CD28 cross-linking, which imparts anti-leishmanial function to neutrophils by inducing cytokine secretion, increased phagocytosis. Using Plasmodium-infected human RBC and Leishmania as the models, we did not observe any difference between phagocytosis by CD28 and isotype cross-linked neutrophils (data not shown). This suggests that CD28-induced Leishmania killing by neutrophils is not via parasite phagocytosis. Instead, the fact that the parasite growth is inhibited by anti-IFN-γ Ab demonstrates the IFN-γ-dependent control of Leishmania growth within macrophages.

Because Leishmania-infected macrophages showed an increase in CD86 but not CD80 expression (Fig. 2,D), and because the regulation of T cell function was reported to be differentially modulated by CD80 and CD86 (15, 16), we examined whether neutrophil-expressed CD28 preferred CD86 to CD80 for inducing IFN-γ production. CD28 on neutrophils was cross-linked with immobilized B7.1-Ig-Fc and B7.2-Ig-Fc chimeras, the fusion proteins created by fusion of the extracellular domains of either B7.1 or B7.2 to the Fc portion of human IgG1, and assayed for IFN-γ in the supernatant after 12 h. Stimulation of neutrophils with human B7.1-Ig-Fc chimera or B7.2-Ig-Fc chimera did not show a significant difference in either IFN-γ secretion (Fig. 3,A) or IFN-γ mRNA induction (Fig. 3 B). Similarly, although both the anti-CD80 and anti-CD86 Abs alone prevented the IFN-γ-mediated restriction of parasite load in the neutrophil-macrophage coculture, they did not show a marked difference in their ability to prevent anti-leishmanial activity (data not shown). Our result thus indicates that neutrophils do not show any preference for either of the two B7 molecules.

Because the above experiments suggested that, upon CD28 cross-linking, neutrophils secrete IFN-γ, we investigated the signaling involved in CD28-induced IFN-γ secretion. We found that cross-linking with anti-CD28 Ab but not with the control isotype resulted in IFN-γ secretion, which peaked within 6 h of stimulation (Fig. 4 A).

It is known that T cell-expressed CD28 results in cytokine production that is sensitive to wortmannin, a PI3-kinase inhibitor (13, 17), but resistant to cyclosporin A (13). To test whether PI3-kinase was involved in CD28-induced IFN-γ secretion from neutrophils, the cells were incubated with different doses of wortmannin, followed by CD28 cross-linking. It was recorded that the CD28-induced IFN-γ secretion was inhibited by wortmannin in a dose-dependent manner (Fig. 4,A, inset), suggesting a PI3-kinase-dependent IFN-γ secretion. The observation was confirmed by confocal microscopy, which showed that the CD28 cross-linked neutrophils expressed more IFN-γ after 9 h than the isotype cross-linked neutrophils and that wortmannin treatment of the CD28 cross-linked neutrophils inhibited IFN-γ production (Fig. 4,B). We investigated whether CD28 was physically associated with PI3-kinase after cross-linking, as described in case of T cells (17). Immunoprecipitation of CD28 from the CD28 cross-linked neutrophils followed by Western blot with PI3-kinase Ab showed the presence of the reported PI3-kinase protein (18, 19, 20) only in CD28 cross-linked but not in the isotype cross-linked or unstimulated neutrophils (Fig. 4 C). The result suggested that neutrophil-expressed CD28 associated with PI3-kinase after CD28 cross-linking.

Although IFN-γ secretion was wortmannin sensitive, addition of different concentrations of cyclosporin A to the culture did not inhibit either transcription (Fig. 4,D, inset) or secretion (Fig. 4 D) of IFN-γ, indicating the cyclosporin A insensitivity of CD28-induced IFN-γ production in neutrophils as reported in case of T cells (13). The observation thus demonstrated that upon CD28 cross-linking PI3-kinase was associated with CD28 to transmit the signal inducing IFN-γ secretion.

Because CD28 enhances the TCR-induced cytokine mRNA in T cells (21), we checked whether CD28-induced IFN-γ secretion was enhanced by a concomitant stimulation with IL-12, an IFN-γ-inducing cytokine. IL-12 addition to the anti-CD28-stimulated neutrophil culture did not augment CD28-induced IFN-γ secretion from neutrophils (Fig. 5,A). Addition of anti-CD3 to neutrophils had no effect on anti-CD28-induced IFN-γ secretion (Fig. 5,B). In T cells, CD28-derived signal alone failed to induce IFN-γ (Fig. 5,C), although it enhanced anti-CD3-induced IFN-γ production (Fig. 5 D). A similar kind of synergism between the CD28 signal and LPS or IL-8, the known neutrophil activators, in IFN-γ induction was not observed in neutrophils (data not shown). Besides differential signaling requirements for IFN-γ production, the results described in this work showed two other differences in the cytokine secretion by T cells and neutrophils. Whereas the neutrophil IFN-γ secretion peaked 6–9 h after stimulation, the T cell cytokine secretion peaked ∼48 h after stimulation. Moreover, IFN-γ produced by T cells was 20 times more than that produced by neutrophils.

Because IFN-γ secretion peaked rapidly after CD28 signaling, it was possible that CD28 signaling through PI3-kinase induced IFN-γ secretion from its preformed pool. Therefore, we stained intracellular IFN-γ at various time points after CD28 cross-linking. We observed that the intracellular IFN-γ staining was not detectable above the background level until 3 h after CD28 signaling, suggesting that the cytokine was induced and secreted but not released from the preformed pool (Fig. 6 A).

To investigate whether the PI3-kinase signaling was inducing IFN-γ gene transcription, the neutrophils were stimulated with anti-CD28 Ab in the presence of different doses of actinomycin D, an RNA transcription inhibitor (22). We observed an actinomycin D dose-dependent inhibition of IFN-γ production (Fig. 6,B). In other experiments, the optimum dose of actinomycin D (50 ng/ml) was added to the anti-CD28-stimulated neutrophil cultures at different time points after initiating the culture as indicated in Fig. 6,C. Addition of the transcription inhibitor 3 h after initiation of culture had little effect on the cytokine production (Fig. 6 C). These experiments suggested that transcription was necessary for IFN-γ production after CD28 signaling.

To determine whether the PI3-kinase signaling was inducing IFN-γ gene, RNA was prepared from the neutrophils cross-linked with CD28 in the presence or absence of wortmannin. Subsequent reverse transcription of the RNA, followed by PCR, showed that the message for IFN-γ was extremely low in the unstimulated neutrophils and that the message was increased significantly after CD28 cross-linking but not isotype cross-linking (Fig. 6,D). Addition of wortmannin to the CD28 cross-linked neutrophils reduced IFN-γ RNA levels to that found in unstimulated neutrophils (Fig. 6 D). These observations together indicate that the IFN-γ gene is induced by CD28 signaling through PI3-kinase and that the transcription is an obligatory prerequisite for the cytokine production.

Leishmania is a dimorphic protozoan parasite. Flagellated promastigotes are introduced into the host during the blood meal of the sandfly vector. Parasites then invade the cells of the monocyte-macrophage lineage of the mammalian host and are transformed into aflagellate amastigotes. It has been proposed that the host’s reaction in response to the infection initiated at this point is mediated by a monoclonal T cell repertoire of Vβ4⁺Vα8⁺ T cells (23), which produces IL-4, to set the bias toward Th2 cells in the susceptible host (24). In contrast, an earlier suggestion was that the initial T cell activation results in the production of a wide array of cytokines (25) that control parasite growth by regulating the macrophage activation or inactivation (26) or by setting a Th subset bias (24). Therefore, the time of involvement of T cells in the initiation of the anti-leishmanial immune response is not well defined, raising a possibility of involvement of non-T cells in the initiation of anti-leishmanial immune response. Among the non-T cells, neutrophils are known to infiltrate the site of infection as early as 21 h after s.c. introduction of the parasite in mice (our unpublished observation) and are present in the lesions of cutaneous leishmaniasis patients (27). A possible role of neutrophils at the site of lesion is phagocytosis of Leishmania (6). Therefore, we tested the hypothesis that neutrophils may play a crucial role in the initiation of anti-leishmanial immune response.

In this paper, we describe how the inflammatory reaction against Leishmania infection is initiated with the release of chemokines such as IL-8 by Leishmania-infected macrophages, followed by IL-8-dependent recruitment of neutrophils to the site of infection. CD28 on neutrophils interacts with CD80/CD86 on macrophages to generate CD28 signal-inducing IFN-γ secretion. In addition to restriction of Leishmania growth, IFN-γ may alter the cytokine microenvironment at the site of infection, affecting the IL-4- and IFN-γ-secreting T cell differentiation as a consequence (4, 28). Although neutrophils have already been reported to secrete IFN-γ (11), our study provides the first mechanistic description of how a signal generated by CD28 on cell surface regulates the cytokine gene transcription and secretion in neutrophils. While observations reported in this work indicate the significance of CD28 signaling in neutrophil-mediated induction of anti-leishmanial immune response, earlier observations on murine L. major infection in CD28-deficient mice suggested a limited role for CD28 in Th subset differentiation during the infection (29, 30). The apparent discrepancy between these observations can be explained by a possible absence of CD28 on murine neutrophils, as casein-elicited early peritoneal exudate cells do not stain for CD28 (data not shown). Thus, although the current hypothesis for initiation of the disease leishmaniasis by setting a Th subset bias relies heavily on T cell-secreted cytokines (31), our observations define for the first time a significant role of neutrophils in the afferent limb of the anti-leishmanial immune response, especially that which precedes the T cell response (32).

In the afferent arm of an immune response, interaction between the cells of the adaptive immune system such as T cells and B cells is a well-documented fact, but such interactions between two different types of cells of the innate immune system has not been shown previously. The results described in this work show that the neutrophils and macrophages interact through CD28-B7, which are known to be the key regulatory molecules for cytokine production by T cells. We have identified three major differences between T cell and neutrophil IFN-γ secretion. First, while T cells require the first signal through the Ag-specific TCR to induce the cytokine gene transcription, the message of which is stabilized by CD28 signaling (33), neutrophils do not require any primary signal to accomplish the same function. Absence of such a requirement for the Ag-specific signal in neutrophil also justifies its Ag nonspecific function during an inflammatory response. However, it is possible (but remains to be tested) that neutrophils express a pattern recognition receptor that signals, as TCRs do in T cells, to induce more IFN-γ. Second, the kinetics of IFN-γ production in neutrophils is different from that of T cells. While in T cells IFN-γ production peaks ∼48 h after anti-CD3 plus anti-CD28 stimulation, in neutrophils IFN-γ production peaks between 6 and 12 h after stimulation. This rapid production of the cytokine correlates with the observation that the neutrophils are the first cells to migrate to the site of infection to encounter foreign materials (1), favoring the age-old paradigm that the cells of the innate immune system unleash the first assault on invaders. Third, neutrophils produce significantly less IFN-γ than T cells (p < 0.001). A recent report shows that cytokine levels as low as 50 pg/ml induce Th cell differentiation (34). This reported level of the cytokine is equal to the quantity of IFN-γ that we measured. The observation implies that, despite low IFN-γ concentration, the cytokine does not lose its functional significance, which is amply demonstrated in our anti-leishmanial assays. Besides the differences in the biological functions of these two cell types, they also vary in cell surface costimulatory molecule expression. While T cells express both CD28 and CTLA-4, which signal in a reasonably counteracting manner in regulating IL-4 and IFN-γ secretion from T cells (35), the neutrophils express only CD28 and not CTLA-4 (data not shown).

When considered from the evolutionary perspective, the marked quantitative and qualitative differences in the cytokine produced by T cells and neutrophils suggest some interesting possibilities. The Ag recognition mechanisms, perhaps the first form of Ag recognition using simple patterns on the Ag, came to existence in phagocytes much earlier than in T cells (36, 37). Therefore, as T cells evolved, the complex MHC-restricted Ag recognition system, which is much finer and detailed, covering wider range of Ags, came into operation (38). Because Igs, MHC molecules, TCR, CD28, and CTLA-4 belong to the Ig superfamily (39), it is possible that CD28 came earlier than the TCR and might have added a finer operational flexibility to the neutrophils. However, phylogenetic studies on these molecules are required to lend further support to this notion.

Because neutrophils are the first cells to encounter invading pathogen, their primary function is to restrict further invasion and mobilize a finer course of immune response. It is possible that a higher amount of IFN-γ production by neutrophils may lead to destruction of host tissue. In contrast, a much higher quantity of IFN-γ from T cells provides a wider window for finer modulation of the cytokine production under different conditions. Host tissue destruction by a high quantity of IFN-γ may be significantly inhibited by the counteracting T cell cytokines, such as IL-4 and IL-10 (40, 41), which are not secreted by neutrophils (data not shown). However, neutrophils do secrete T cell chemotactic factors, which may draw the T cells to the site of infection, and can modulate Th subset differentiation (Refs. 5 and 42 and K. Venuprasad, S. Chattopadhyay, and B. Saha, manuscript in preparation). Thus, the phylogeny of functions of these two cells is mirrored during the progress of an Ag-specific immune response in a two-tiered system. In the first tier, at the beginning of the response, the neutrophils not only encounter the Ag directly, as demonstrated using Leishmania infection here, but also lead to the second tier, where T cells dictate the final outcome of the infection (42).

In conclusion, neutrophils play a crucial role in the initiation of anti-leishmanial immune response by secreting IFN-γ and restricting parasite growth. The results described in this paper reveal some novel facts about the differential regulation of T cell and neutrophil IFN-γ production that affects the course of Leishmania infection and perhaps other intracellular infections as well; they can serve as the basis for further exploration of the role of neutrophils in Th subset differentiation.

We extend our sincere thanks to Dr. G. C. Mishra (National Center for Cell Science) for encouragement and constructive criticism and to Dr. Satyajit Rath (National Institute of Immunology, New Delhi, India) for critical reading of the manuscript and constructive criticism. We also thank Ashwini Atre for confocal microscopy and Atul Suple for FACS analysis. We thank Dr. Simon L. Croft (London School of Hygiene and Tropical Medicine, London, U.K.) for editing the manuscript.