Dendritic Cell and NK Cell Reciprocal Cross Talk Promotes Gamma Interferon-Dependent Immunity to Blood-Stage Plasmodium chabaudi AS Infection in Mice

Abstract

Dendritic cells (DCs) are important accessory cells for promoting NK cell gamma interferon (IFN-γ) production in vitro in response to Plasmodium falciparum-infected red blood cells (iRBC). We investigated the requirements for reciprocal activation of DCs and NK cells leading to Th1-type innate and adaptive immunity to P. chabaudi AS infection. During the first week of infection, the uptake of iRBC by splenic CD11c⁺ DCs in resistant wild-type (WT) C57BL/6 mice was similar to that in interleukin 15^−/− (IL-15^−/−) and IL-12p40^−/− mice, which differ in the severity of P. chabaudi AS infection. DCs from infected IL-15^−/− mice expressed costimulatory molecules, produced IL-12, and promoted IFN-γ secretion by WT NK cells in vitro as efficiently as WT DCs. In contrast, DCs from infected IL-12p40^−/− mice exhibited alterations in maturation and cytokine production and were unable to induce NK cell IFN-γ production. Coculture of DCs and NK cells demonstrated that DC-mediated NK cell activation required IL-12 and, to a lesser extent, IL-2, as well as cell-cell contact. In turn, NK cells from infected WT mice enhanced DC maturation, IL-12 production, and priming of CD4⁺ T-cell proliferation and IFN-γ secretion. Infected WT mice depleted of NK cells, which exhibit increased parasitemia, had impaired DC maturation and DC-induced CD4⁺ Th1 cell priming. These findings indicate that DC-NK cell reciprocal cross talk is critical for control and rapid resolution of P. chabaudi AS infection and provide in vivo evidence for the importance of this interaction in IFN-γ-dependent immunity to malaria.

Studies performed with humans and with mice indicate that NK cells are a crucial and early source of gamma interferon (IFN-γ) during blood-stage malaria infection (3, 4, 12, 28). Consistent with findings with other infectious agents, tumors, and inflammatory diseases, optimal NK cell IFN-γ production during malaria requires accessory cells (10, 11). Newman et al. (29) demonstrated that IFN-γ production by human NK cells in response to Plasmodium falciparum-infected red blood cells (iRBC) occurs only following multiple contact-dependent and cytokine-mediated signals derived from myeloid dendritic cells (DCs) and monocytes. The proinflammatory cytokines type 1 IFN, interleukin 2 (IL-2), IL-12, and to a lesser extent IL-18, but not IL-15, were found to be required for NK cell activation, while transforming growth factor β suppressed NK cell activation (3, 4, 29). Reciprocal regulation of accessory cell maturation by NK cells was observed to be similar to findings of bidirectional interactions between NK cells and accessory cells in other systems (10, 11, 16, 17). On the other hand, Baratin et al. (6) reported that human monocytes, not DCs, are required to induce NK cell IFN-γ production in vitro in response to P. falciparum iRBC. That study also implicated multiple accessory cell-derived signals, including IL-12 and IL-18, as requirements for NK cell IFN-γ production in response to iRBC, and an IL-18R/MyD88-dependent pathway was demonstrated to be necessary for NK cell activation (6). Together, these studies indicate that NK cell activation following exposure to iRBC in vitro requires an accessory cell, but the identity of this cell is unclear, as is the relevance of interactions between NK cells and accessory cells to the in vivo control of blood-stage malaria.

Given the importance of the innate immune response in shaping adaptive immunity, a better understanding of the cellular mechanisms and cytokines involved in cross talk between NK cells and accessory cells during malaria is required. Control and resolution of Plasmodium chabaudi AS infection in mice critically requires IL-12-dependent IFN-γ production by NK cells and CD4⁺ T cells (37). During acute P. chabaudi AS infection in resistant C57BL/6 (B6) mice, NK cells, macrophages, and DCs rapidly accumulate and expand in the spleen, where they undergo maturation and functional activation in response to iRBC captured in the spleen for removal (24, 25, 28, 32, 46). Notably, CD11c⁺ DCs migrate from the marginal zone to the CD4⁺ T-cell-rich periarteriolar lymphoid sheaths during early infection, while macrophages expand and remain in the red pulp (25). Contrary to reports of functional impairment of DCs during lethal infections due to Plasmodium berghei, P. yoelii YM, or P. vinckei, the in vitro or in vivo exposure of bone marrow-derived or splenic DCs to P. chabaudi AS iRBC results in upregulation of major histocompatibility complex class II and costimulatory molecule expression; production of the proinflammatory cytokines IL-12, IFN-γ, tumor necrosis factor alpha (TNF-α), and IL-6; and the ability to stimulate naïve CD4⁺ T-cell proliferation and IFN-γ production (23-25, 34, 45, 46). In contrast, splenic macrophages from P. chabaudi AS-infected mice suppress T-cell proliferation and IL-2 production (1, 33). Similar findings concerning the ability of DCs to activate T cells as opposed to macrophage-mediated suppression of T-cell activation were reported in mice with a nonlethal P. yoelii infection (26, 30). Together, these findings provide compelling evidence that DCs play an important role in innate immunity during blood-stage malaria and shape an adaptive type 1 immune response necessary for control and rapid resolution of infection.

In the present study, we investigated the requirements for reciprocal activation of DCs and NK cells leading to priming of CD4⁺ Th1 cell responses in P. chabaudi AS-infected mice using in vitro and in vivo approaches. DC maturation, cytokine production, and the ability to activate NK cell IFN-γ production in vitro were compared among wild-type (WT) B6, IL-15^−/−, and IL-12p40^−/− mice that differed in the severity of infection with the parasite (23, 40). The in vivo contribution of DC-NK cell cross talk to the development of immunity to malaria was determined in NK cell-depleted WT mice, which are unable to control blood parasitemia as efficiently as intact mice (28). Our findings confirm and extend previous observations concerning DC-NK cell interactions in vitro and provide novel evidence for the in vivo relevance of DC and NK cell cross talk during the innate immune response to blood-stage malaria leading to the induction of adaptive type 1 immunity.

MATERIALS AND METHODS

Mice and parasite.

IL-12p40^−/−, IL-15^−/−, and IFN-γ^−/− mice on the B6 background, obtained as previously described (23, 39, 40), were bred and maintained at the Research Institute of the McGill University Health Centre (Montreal, QC, Canada). Age-matched B6 mice (Charles River Laboratories, St. Constant, QC, Canada) were used as WT controls. Female mice, aged 8 to 14 weeks, were used in all experiments. Experiments were conducted in accordance with the guidelines of the Canadian Council on Animal Care. P. chabaudi AS was maintained as previously described (38), and infections were initiated by intraperitoneal injection of 10⁶ iRBC.

Purification of CD11c⁺, DX5⁺, and CD4⁺ cells from spleen.

Spleens were harvested from naïve and infected mice at the indicated times and perfused with a sterile solution of 1% fetal calf serum (HyClone Laboratories, Logan, UT) in phosphate-buffered saline (PBS), teased apart, and pressed gently through a sterile fine wire mesh. The RBC were removed by incubation with NH₄Cl lysing buffer. To obtain DCs, splenocytes were suspended in sterile PBS containing 5 mM EDTA (Sigma-Aldrich Co., St. Louis, MO) and separated using a Nycoprep 1.077A (Axis-Shield, Oslo, Norway) density gradient. Low-density cells at the interphase were collected and further purified by positive selection using anti-CD11c microbeads (Miltenyi Biotec, Auburn, CA) according to the manufacturer's instructions. The purity of CD11c⁺ cells was confirmed by flow cytometry for each experiment and was routinely 88 to 90% CD11c⁺; 4 to 8% also expressed DX5. NK cells and CD4⁺ T cells were also enriched from spleen cells using anti-DX5 and anti-CD4 microbeads (Miltenyi Biotec), respectively. NK cells were 85 to 90% positive for DX5; 3 to 9% also expressed CD11c, and 25% expressed CD3. The purity of CD4⁺ T cells was routinely ≥90%.

Analysis of DC uptake of iRBC by flow cytometry.

DC uptake of iRBC was determined as previously described (24). Briefly, purified iRBC and normal RBC (nRBC) were obtained from P. chabaudi-infected or naïve B6 mice, respectively, as described previously (24). Purified iRBC (>96% iRBC) or nRBC were resuspended at 10⁷ per ml in RPMI 1640 (Gibco-Invitrogen, Burlington, ON, Canada) containing 5% fetal calf serum (HyClone), 10 mM HEPES (Gibco-Invitrogen), 20 μg/ml gentamicin (Sabex Inc., QC, Canada), 2 mM l-glutamine (Gibco-Invitrogen), and 0.5 μM β-mercaptoethanol (Sigma-Aldrich Co.) and stained with 2 μM carboxyfluorescein succinimidyl ester (CFSE) (Molecular Probes, Eugene, OR) for 15 min at 37°C. Splenic CD11c⁺ DCs (10⁶/well) were plated with CFSE-labeled iRBC or nRBC at a 1:20 ratio in a final volume of 200 μl of complete RPMI 1640 medium and incubated at 37°C for 4 h. Noningested iRBC or nRBC were lysed, and the DCs were washed, treated with anti-CD16/CD32 monoclonal antibody (MAb) (clone 2.4G2; BD Biosciences, Mountain View, CA) to block Fc receptors (FcR), and stained with phycoerythrin (PE)-labeled anti-CD11c MAb (clone HL3; BD Biosciences). Uptake of iRBC or nRBC was determined by gating on CD11c⁺ cells on SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) versus FL2 and analyzing CFSE staining on FL1 using a FACSCalibur equipped with CellQuest Pro software (BD Biosciences).

Splenic-DC maturation status, cytokine production, and stimulation of CD4⁺ T cells.

Purified CD11c⁺ DCs were FcR blocked and stained with PE- and fluorescein isothiocyanate (FITC)-labeled MAb (BD Biosciences) to CD11c, CD40 (clone 2/23), CD80 (clone 16-1041), or CD86 (clone GL-1). The percentages and mean fluorescence intensity (MFI) of gated CD11c⁺ cells expressing the above-mentioned markers were determined by flow cytometry. For analysis of cytokine production, CD11c⁺ DCs (10⁶/well) were plated and incubated in medium alone or with iRBC (1:10 ratio) for 48 h at 37°C. Cytokine levels were determined in supernatants by enzyme-linked immunosorbent assay (ELISA) as previously described (38, 39). To determine the ability of DCs to stimulate CD4⁺ T cells, 5 × 10⁵ DCs in 100 μl were plated with splenic CD4⁺ T cells, freshly purified from naïve B6 mice, in 96-well flat-bottom plates (Nalge Nunc International, Naperville, IL) at the indicated ratios in a final volume of 200 μl for 48 h at 37°C. Proliferation was determined by incorporation of [³H]thymidine added during the last 16 h of incubation. Supernatants of DC and T-cell cocultures were collected and analyzed for cytokine production by ELISA as described previously (38, 39).

DC-NK cell cocultures.

To determine if malaria-activated DCs promoted NK cell IFN-γ production, CD11c⁺ DCs were purified from the spleens of day 5 infected WT or cytokine-deficient mice and cocultured with NK cells purified from naïve WT mice or, in some experiments, from IFN-γ^−/− mice. For cocultures, DCs (10⁶/well) were plated together with NK cells (10⁶/well) in 96-well flat-bottom plates for 24 to 36 h at 37°C. In some experiments, NK cells were cultured in the lower well and DCs were added separately in the upper compartment of a 96-well plate fitted with transwell inserts (Nalge Nunc International). Single cultures of DCs and NK cells were always included as controls. In some experiments, DCs were pretreated with 250 ng/ml anti-IL-2 MAb (clone JES6-1A12; R&D Systems) or isotype control immunoglobulin G (IgG) for 4 h prior to coculture with NK cells.

To determine the effects of malaria-activated NK cells on resting DCs, purified DX5⁺ NK cells from day 5 infected WT or IFN-γ^−/− mice were cocultured at a 1:1 ratio with CD11c⁺ DCs from naïve WT mice preactivated with iRBC for 1 h prior to coculture to prevent DC lysis by activated NK cells. DCs (1 × 10⁶) were plated with NK cells (1 × 10⁶) in 96-well flat-bottom plates for 18 to 24 h at 37°C. Costimulatory molecule expression was determined on gated CD11c⁺ DX5⁻ cells as described above, and supernatants were collected and analyzed for cytokine levels by ELISA. The ability of DCs cocultured with activated NK cells to prime naïve CD4⁺ T cells was assessed using a modification of a protocol described by Mailliard et al. (27). Briefly, DCs (2 × 10⁵) and NK cells (2 × 10⁵) were cocultured in 48-well plates at 37°C; 18 to 24 h later, splenic CD4⁺ T cells (2 × 10⁶), freshly purified from naïve mice, were added. For CFSE dilution assays, CD4⁺ T cells were stained with 5 mM CFSE (Molecular Probes) prior to addition to DC-NK cell cocultures. After 2 days, fresh medium containing recombinant IL-2 (10 ng/ml; R&D Systems) was added and the plates were further incubated for 72 h at 37°C. Cells were collected and adjusted to 5 × 10⁵/ml, and 200 μl was plated for 48 h at 37°C in 96-well flat-bottom plates coated with 5 μg/ml anti-CD3ɛ MAb (clone 145-2C11; BD Biosciences). Proliferation was determined by CFSE dilution in gated CD4⁺ T cells by flow cytometry or by incorporation of [³H]thymidine added during the last 16 h of incubation. CD4⁺ T-cell cytokine secretion was determined by intracellular cytokine staining. Supernatants of cocultures were also collected and analyzed for cytokines by ELISA.

Intracellular cytokine staining.

For intracellular IFN-γ or IL-2 expression by DCs, cells were stimulated in vitro with Golgi Stop (BD Biosciences), 10 ng/ml of phorbol myristate acetate (Sigma-Aldrich Co.), and 250 ng/ml of ionomycin (Sigma-Aldrich Co.) for 2 h. The DCs were washed, FcR blocked, and stained with FITC-conjugated anti-CD11c (BD Biosciences) MAb. The cells were fixed and permeabilized according to the manufacturer's instructions with the Cytofix/Cytoperm kit (BD Biosciences), followed by staining with allophycocyanin (APC)-conjugated anti-IFN-γ (clone XMG1.2; BD Biosciences) or APC-conjugated anti-mouse IL-2 (clone JES6-5H4; eBioscience, San Diego, CA). To exclude CD11c⁺ DX5⁺ IFN-γ⁺ cells, the cells were also stained with PE-conjugated anti-CD49b (clone DX5; BD Biosciences), and IFN-γ expression was determined in gated CD11c⁺ DX5⁻ cells. For DC and NK cell cocultures, cells were stimulated and stained with FITC-conjugated anti-CD11c and PE-conjugated anti-CD49b MAbs. Following fixing and permeabilization, the cells were stained with APC-conjugated anti-IFN-γ MAb, and IFN-γ signals in gated CD11c⁺ DX5⁻ (FL1) or CD11c⁻ DX5⁺ (FL2) populations were detected on FL4. For intracellular cytokine expression by gated CD4⁺ T cells, cells were stimulated as described above for 5 h and then stained with FITC-conjugated anti-CD4 MAb (clone GK1.5; BD Biosciences), fixed and permeabilized, and stained with APC-conjugated anti-IFN-γ, APC-conjugated anti-IL-10 (clone JES5-2A5; eBioscience), or PE-conjugated anti-IL-4 (clone 11B11; eBioscience) MAb.

In vivo NK cell depletion.

WT mice were depleted of NK cells by intravenous injection with 80 μg anti-asialo-GM1 antibody (Wako Chemicals, Richmond, VA) on days −7, −4, 0, and 2 postinfection (p.i.). This treatment resulted in >90% depletion of DX5⁺ NK cells in the spleen, as confirmed by flow cytometry (28). Control mice were treated similarly with normal rabbit IgG as an isotype control antibody or with sterile PBS. To assess DC maturation and function after NK cell depletion, splenic CD11c⁺ cells were purified as described above and analyzed for costimulatory molecule expression, cytokine production, and the ability to prime CD4⁺ T cells.

Statistical analyses.

Data are expressed as means ± standard errors of the mean (SEM). The statistical significance of differences between two groups was analyzed by Student's t test, and multiple comparisons were analyzed using analysis of variance. All statistical analyses were performed using SAS/STAT software (SAS Institute, Cary, NC).

RESULTS

IL-12, but not IL-15, is required for optimal DC maturation during blood-stage malaria.

We first compared the uptake of iRBC and costimulatory molecule expression by CD11c⁺ DCs purified from the spleens of naïve and infected WT, IL-15^−/−, and IL-12p40^−/− mice. As we previously reported (24), WT DCs from both naïve and infected mice captured iRBC at significantly higher levels than nRBC, and the uptake of iRBC was selectively enhanced to peak levels within 4 to 5 days of infection and declined thereafter (Fig. 1A and B). DCs from IL-15^−/− and IL-12p40^−/− mice captured iRBC in vitro as efficiently as WT DCs both before and after infection (Fig. 1B). On day 4 p.i., similar increases in CD40, CD80, and CD86 expression were observed in DCs from WT and IL-15^−/− mice, but IL-12p40^−/− DCs had significantly lower costimulatory molecule expression than WT DCs (Fig. 1C). These findings indicate that in the absence of IL-15 or IL-12p40, recognition and capture of iRBC by DCs were as efficient as the responses of WT mice, but DC maturation was impaired in IL-12p40^−/− mice following P. chabaudi AS infection.

FIG. 1.

IL-12 is required for DC maturation but not for iRBC uptake. (A) Uptake of iRBC by splenic DCs at day 0 and day 4 p.i. The histograms represent CFSE expression by gated CD11c⁺ cells cultured alone (solid gray area) or with CFSE-labeled iRBC (lines). Splenic CD11c⁺ DCs were purified from WT (solid lines), IL-15^−/− (dashed lines), and IL-12p40^−/− (dotted lines) mice. (B) Kinetics of iRBC and nRBC uptake by splenic DCs from naïve and P. chabaudi AS-infected mice. The data were pooled from three experiments, each containing three or four mice per group. The asterisks indicate significant differences between iRBC and nRBC uptake at the specified time points: *, P < 0.05; **, P < 0.01; and ***, P < 0.001. (C) MFI of CD40, CD80, and CD86 expression by gated CD11c⁺ cells from naïve or day 4 infected mice. The data are representative of two experiments, each containing three or four mice per group. The asterisks indicate significant differences between WT and IL-12p40^−/− DCs: *, P < 0.05, and **, P < 0.01. The error bars indicate SEM.

IL-12 deficiency impairs IFN-γ but increases IL-2 secretion by DCs during blood-stage malaria.

Since the production of proinflammatory and immunoregulatory cytokines by DCs is critical for the generation of downstream innate and adaptive immune responses (5), we also compared cytokine secretion by DCs from naïve and infected WT, IL-15^−/−, and IL-12p40^−/− mice. In agreement with previous findings (23, 25, 34), DCs recovered from P. chabaudi AS-infected WT mice produced normal or even higher levels of IL-12p70, IFN-γ, IL-2, and IL-10 than DCs from naïve WT mice following in vitro stimulation with either medium or iRBC (Fig. 2). Similarly, infection did not inhibit the ability of DCs from IL-15^−/− or IL-12p40^−/− mice to secrete cytokines. As shown in Fig. 2A, after infection, WT and IL-15^−/− DCs produced equivalent levels of IL-12p70 in vitro in response to medium or iRBC, but WT DCs produced significantly higher levels of IFN-γ than IL-15^−/− DCs, a finding confirmed by intracellular IFN-γ expression (Fig. 2B). No detectable IL-2 was produced by DCs from infected WT and IL-15^−/− mice in response to medium, and only low levels of the cytokine were produced in response to iRBC. DCs from infected IL-12p40^−/− mice also produced less IFN-γ in response to medium or iRBC than infected WT DCs, as was apparent in the significantly lower levels of IFN-γ in the culture supernatants and a lower frequency of CD11c⁺ cells expressing intracellular IFN-γ ex vivo. After infection, IL-12p40^−/− DCs secreted more IL-2 than WT DCs in response to medium or iRBC, a finding confirmed by a higher frequency of CD11c⁺ IL-2⁺ cells ex vivo among DCs from infected IL-12p40^−/− mice (Fig. 2A and C). Despite differences in type 1 cytokine production by DCs from infected WT, IL-15^−/−, and IL-12p40^−/− mice, IL-10 secretion was enhanced to similar levels following iRBC stimulation of DCs from WT or cytokine-deficient mice after P. chabaudi AS infection (Fig. 2A).

FIG. 2.

IL-12 deficiency impairs IFN-γ production but increases IL-2 production by DCs during blood-stage malaria. (A) Splenic CD11c⁺ DCs from naïve or day 4 infected WT, IL-15^−/−, and IL-12p40^−/− mice were stimulated in vitro without (Medium) or with iRBC. The supernatants were analyzed for IL-12p70, IFN-γ, IL-2, and IL-10 by ELISA. The asterisks indicate significant differences between WT and cytokine-deficient DCs: *, P < 0.05; **, P < 0.01; and ***, P < 0.001. (B) Intracellular IFN-γ expression by iRBC-stimulated splenic DCs from day 4 infected WT, IL-15^−/−, and IL-12p40^−/− mice. The cells were gated on the DX5-negative population, and the percentages of IFN-γ expression by CD11c⁺ DCs are shown. (C) Intracellular IL-2 expression by splenic CD11c⁺ DCs from naïve or day 7 infected WT and IL-12p40^−/− mice. For panels B and C, appropriate isotype staining controls are shown. The data are representative of two experiments, each containing three or four mice per group. The error bars indicate SEM.

DCs from infected WT mice promote IFN-γ production by NK cells.

Next, we investigated if DCs from P. chabaudi AS-infected mice promoted IFN-γ production by NK cells by coculturing splenic DCs purified from infected WT mice with resting NK cells purified from spleens of naïve WT mice. Because both DCs and NK cells produce IFN-γ during P. chabaudi AS infection (24, 25), IFN-γ secretion contributed by each of these cell populations was assessed by determining the frequency of IFN-γ⁺ cells in gated CD11c⁺ DX5⁻ cells (DCs) or DX5⁺ CD3⁻ cells (NK cells). Prior to coculture, few CD11c⁺ DX5⁻ DCs expressed intracellular IFN-γ, and the frequency of IFN-γ⁺ DCs increased only modestly after coculture with resting NK cells compared with DCs cultured alone (Fig. 3A, pre- versus postculture). Likewise, few DX5⁺ CD3⁻ NK cells expressed intracellular IFN-γ prior to coculture (Fig. 3B). Notably, after coculture with DCs from infected WT mice, the frequency of DX5⁺ CD3⁻ IFN-γ⁺ NK cells increased markedly from 6.58% to 35.80% (Fig. 3B, pre- versus postculture). These findings indicate that DX5⁺ CD3⁻ NK cells were the major source of IFN-γ in the cocultures, a result corroborated when WT DCs from infected mice were cocultured with resting NK cells from naïve IFN-γ^−/− mice (data not shown).

FIG. 3.

Malaria-activated DCs promote high IFN-γ production by resting NK cells. CD11c⁺ DCs from day 5 infected WT mice and DX5⁺ NK cells from naïve WT mice were cultured separately or together for 24 to 36 h at 37°C and analyzed for IFN-γ production. (A and B) Intracellular IFN-γ expression by gated populations of CD11c⁺ DX5⁻ DC or DX5⁺ CD3⁻ NK cells isolated from infected and naïve WT mice, respectively. (Right) IFN-γ expression by CD11c⁺ DX5⁻ DCs and DX5⁺ CD3⁻ NK cells before culture (Pre Culture) and after culture alone (lines) or together (shaded areas) (Post Culture). The data are representative of four experiments giving similar results. (C) Secreted levels of IFN-γ were determined by ELISA in single cultures of NK cells from naïve WT mice or DCs from infected WT mice (Day 5 DC) and in cocultures of WT NK cells with DCs from naïve WT mice (Day 0 DC) or day 5 p.i. (Day 5 DC). The DCs were separated from NK cells by using a transwell insert (Transwell). The data are representative of four experiments, each with three or four replicates per group. The asterisks indicate significant differences compared with DCs from day 5 infected WT mice (Day 5 DC): ***, P < 0.001. The error bars indicate SEM.

Determination of IFN-γ levels in coculture supernatants confirmed that WT DCs from infected mice stimulated higher IFN-γ production than WT DCs from naïve mice (Fig. 3C). DC-NK cell interaction was dependent on cell-cell contact, as demonstrated by the significantly lower IFN-γ production in wells in which NK cells and DCs were separated by a transwell insert. Although it has been shown that DCs also elicit TNF-α secretion by NK cells (16, 31), we observed only background levels of TNF-α in the supernatants of DC-NK cell cocultures (data not shown). Taken together, these results demonstrate that splenic DCs from malaria-infected WT mice promoted resting NK cells to produce high levels of IFN-γ.

IL-12 and IL-2, but not IL-15, are required for DC-mediated stimulation of NK cell IFN-γ production.

To determine the relative contributions of IL-15 and IL-12 to DC-NK cell interactions during blood-stage malaria, DCs from P. chabaudi AS-infected IL-15^−/− and IL-12p40^−/− mice were cocultured as described above with NK cells from naïve WT mice. As shown in Fig. 4A, DCs from infected IL-15^−/− mice stimulated a percentage of WT NK cells similar to that stimulated by WT DCs (32.7% versus 36.1%, respectively) to express intracellular IFN-γ. In contrast, DCs from infected IL-12p40^−/− mice were unable to promote IFN-γ expression by resting WT NK cells (6.3%) (Fig. 4A, right). IFN-γ levels were similar in the supernatants of cocultures containing DCs from infected WT mice and those containing DCs from infected IL-15^−/− mice, while significantly less IFN-γ was detected in the supernatants of cocultures with DCs from infected IL-12p40^−/− mice (Fig. 4B). These findings indicate that DCs from infected IL-12p40^−/− mice had impaired maturation and type 1 cytokine production and were unable to promote NK cell IFN-γ production.

FIG. 4.

IL-12 and IL-2 are key cytokines in DC-mediated stimulation of NK cells. Splenic DCs from day 5 infected WT, IL-15^−/−, or IL-12^−/− mice were cocultured for 24 to 36 h at 37°C with NK cells from naïve WT mice. (A) Intracellular IFN-γ expression by WT, IL-15^−/−, or IL-12^−/− DCs and WT NK cells after coculture. Each panel displays composite histograms of IFN-γ signals from gated CD11c⁺ DX5⁻ DC or DX5⁺ CD3⁻ NK cell populations. The lines represent expression by DCs, and the shaded histograms represent NK cells. The percentages of DC or NK cells expressing IFN-γ are shown. The data are representative of four experiments showing similar results. (B) IFN-γ levels in the coculture supernatants were quantified by ELISA. The data are representative of four experiments, each containing three or four mice per group. The asterisks indicate a significant difference between WT and IL-12p40^−/− DCs: ***, P < 0.001. (C) WT, IL-15^−/−, or IL-12^−/− DCs were cocultured with WT NK cells, and IL-2 activity was blocked using a neutralizing anti-IL-2 MAb. Control wells were treated with isotype control IgG MAb. IFN-γ levels were quantified by ELISA after 36 h of coculture. The data are representative of two experiments, each containing three or four replicates per group. The asterisks indicate significant differences between DCs treated with isotype control MAb and DCs treated with anti-IL-2 MAb: **, P < 0.01. The error bars indicate SEM.

In addition to IL-12 and IL-15, DC-derived IL-2 has been shown to be important for DC-NK cell cross talk (2, 14, 19). We therefore asked if the ability of IL-15^−/− DCs to stimulate IFN-γ production by NK cells was due to the compensatory effect of IL-2. The addition of neutralizing anti-IL-2 MAb significantly reduced IFN-γ production in cocultures of IL-15^−/− DCs and WT NK cells, as well as in cocultures of WT DCs and WT NK cells (Fig. 4C). This suggests that low levels of IL-2 production by WT and IL-15^−/− DCs from infected mice, as shown in Fig. 2A above, may partially account for the ability of malaria-activated IL-15^−/− DCs to promote IFN-γ production by resting WT NK cells. Anti-IL-2 MAb also significantly reduced NK cell IFN-γ production in cocultures of WT NK cells and infected IL-12p40^−/− DCs, suggesting that high IL-2 secretion by these DCs, as shown in Fig. 2, did not compensate for the detrimental effects of IL-12 deficiency on DC-stimulated NK cell activation. Together, these data indicate that IL-12 and to a lesser extent IL-2, but not IL-15, produced by DCs during the innate immune response to blood-stage P. chabaudi AS were required to promote optimal IFN-γ secretion by NK cells during malaria.

Malaria-activated NK cells enhance DC maturation, cytokine production, and the ability to activate CD4⁺ Th1 cells.

Previous studies indicated that NK cells activated by tumor antigens, bacterial products, or IL-2 induce DCs to mature and produce Th1-promoting IL-12 (15-17,31). Because splenic NK cells from WT mice have increased cytolytic activity and IFN-γ production during P. chabaudi AS infection (28), we investigated if activated NK cells from infected WT mice were capable of enhancing DC function. In preliminary studies, we obtained a lower recovery rate of CD11c⁺ cells when DCs from naïve WT mice were cocultured with NK cells from P. chabaudi AS-infected WT mice over a wide range of DC/NK cell ratios (data not shown), an observation in line with published findings that immature DCs are susceptible to NK cell lysis (16, 31). To circumvent this problem, DCs from naïve mice were incubated with iRBC for 1 h, followed by treatment with lysing buffer prior to coculture with NK cells. For the experiments shown here, cells were cultured at the optimal ratio of 1:1 for DC-NK cells.

WT DCs cocultured with WT NK cells from infected mice showed higher upregulation of CD40 and CD86 expression than control DCs cultured alone or with NK cells from naïve WT mice (Fig. 5A). DCs cocultured with WT NK cells in the presence of anti-IFN-γ MAb or isotype control MAb (data not shown) or cocultured with IFN-γ^−/− NK cells exhibited phenotypic maturation similar to that of DCs cocultured with NK cells from infected WT mice (Fig. 5A). Although the percentages of DCs expressing each maturation marker were comparable (data not shown), DCs cocultured with malaria-activated NK cells in transwells expressed significantly lower MFI levels of CD40 and CD86 (Fig. 5A). These data suggest that DC-NK cell contact, and not soluble factors, supported optimal DC maturation by NK cells.

FIG. 5.

Malaria-activated NK cells stimulate DCs to mature and secrete cytokines. Splenic NK cells from day 5 infected WT mice were cocultured for 18 to 24 h at 37°C with naïve WT DCs exposed to iRBC before coculture. DCs were analyzed for costimulatory molecule expression (A) and cytokine production (B). To determine the role of IFN-γ in these interactions, DCs were cocultured with NK cells from IFN-γ^−/− mice or with WT NK cells in the presence of anti-IFN-γ MAb or isotype control MAb (data not shown). DCs were also cocultured with NK cells in wells with a transwell insert to prevent cell-cell contact. (A) Cells were gated on the CD11c⁺ signal and analyzed for CD40, CD80, and CD86 expression (MFI) after culture with medium or coculture with NK cells from naïve mice (Day 0 NK) as controls or coculture with WT or IFN-γ^−/− NK cells from day 5 infected mice (Day 5 NK). The asterisks indicate a significant difference between DCs separated from day 5 NK cells by a transwell insert (Transwell) and DCs cocultured with day 5 NK cells (Coculture): **, P < 0.01. (B) Cytokine levels in supernatants of DCs cultured alone or with day 0 or day 5 NK cells were determined by ELISA. The data are representative of three experiments, each containing three or four replicates per group. The asterisks indicate significant differences between WT NK cells (Coculture) and IFN-γ^−/− NK cells: **, P < 0.01 and ***, P < 0.001. The error bars indicate SEM.

Malaria-activated NK cells also stimulated DCs to secrete higher levels of cytokine, especially IL-12, than those observed in DCs cultured alone (Fig. 5B). The addition of anti-IFN-γ MAb resulted in significantly decreased IL-12p40 and IL-12p70, but increased IL-10, levels in cocultures of DCs and malaria-activated WT NK cells (Fig. 5B) compared to the addition of isotype control MAb (data not shown). Similar results were obtained with WT DCs cocultured with NK cells from P. chabaudi AS-infected IFN-γ^−/− mice, demonstrating that NK cell-derived IFN-γ is required for enhanced IL-12 production by DCs during malaria. While some NK cell subsets are capable of producing IL-10 after infection, in our hands, only background levels were detected in day 5 NK cells cultured alone (data not shown). DCs cultured in wells separated from NK cells by a transwell insert showed levels of cytokine production similar to those in cocultures, suggesting that soluble factors released from NK cells, and not cell-cell contact, were responsible for cytokine production by DCs. Thus, cell-cell contact was required for optimal DC phenotypic maturation, while stimulation of high IL-12 production by malaria-activated NK cells was dependent on soluble mediators, particularly IFN-γ.

The ability of DCs stimulated by malaria-activated NK cells to prime CD4⁺ Th1 cell responses was also examined. In the absence of CD4⁺ T cells, control cultures of anti-CD3-stimulated DCs or NK cells alone or cocultures of DCs and NK cells did not proliferate or produce cytokines above background levels (data not shown). Similarly, proliferation and cytokine secretion were low in CD4⁺ T cells cultured with either DCs or NK cells alone (Fig. 6). In contrast, CD4⁺ T cells cultured with NK cell-activated DCs proliferated vigorously, as assessed by both [³H]thymidine incorporation (Fig. 6A) and CFSE staining of gated CD4⁺ T cells (Fig. 6B). Significantly higher levels of the type 1 cytokine IFN-γ were secreted in the supernatants of CD4⁺ T cells cultured with NK cell-activated DCs than in that of T cells cultured in medium or with either DCs or NK cells alone (Fig. 6C). This finding was reflected in the higher percentage of CD4⁺ T cells expressing intracellular IFN-γ in cultures with NK cell-activated DCs (Fig. 6D). Consistent with the results presented in Fig. 5, DCs activated by NK cells from infected WT mice stimulated significantly higher levels of CD4⁺ T-cell proliferation than did DCs activated by infected IFN-γ^−/− NK cells (Fig. 6A and B). DCs cultured with infected IFN-γ^−/− NK cells stimulated CD4⁺ T cells to produce significantly lower levels of IFN-γ but higher levels of IL-4, as well as IL-10, than DCs activated by NK cells from infected WT mice (Fig. 6C and D). Thus, NK cells from P. chabaudi AS-infected WT mice stimulated DCs to activate CD4⁺ Th1 cells via an IFN-γ-dependent mechanism.

FIG. 6.

NK cell-activated DCs prime CD4⁺ T cells in an IFN-γ-dependent mechanism. Splenic NK cells from day 5 infected WT or IFN-γ^−/− mice were cocultured with DCs from naïve WT mice for 24 h and then plated with freshly isolated CD4⁺ T cells from naïve mice. After 5 days, the cells were harvested, restimulated in wells coated with anti-CD3 MAb for 48 h, and analyzed for proliferation (A and B) and cytokine production (C and D). CD4⁺ T cells cultured alone (Medium) or with DCs or WT NK cells separately were included as controls. (A) DCs activated by WT NK cells induced significantly higher levels of T-cell proliferation than did DCs cultured with IFN-γ^−/− NK cells. The asterisks indicate a significant difference between WT NK cells and IFN-γ^−/− NK cells cocultured with DCs: **, P < 0.01. (B) The CFSE dilution in gated CD4⁺ T cells following coculture with medium, DCs alone, or DCs stimulated with WT or IFN-γ^−/− NK cells from infected mice was determined by flow cytometry. (C) DCs activated by WT NK cells induced higher levels of IFN-γ production by CD4⁺ T cells. The asterisks indicate significant differences in cytokine production by CD4⁺ T cells primed by DCs cocultured with malaria-activated WT versus IFN-γ^−/− NK cells: **, P < 0.01, and ***, P < 0.001. (D) Intracellular cytokine expression by gated CD4⁺ T cells as determined by flow cytometry. The percentages of cells expressing each cytokine are indicated. The data are representative of two or three experiments, each containing three or four replicates per group. The error bars indicate SEM.

Depletion of NK cells results in higher parasitemia and impairs splenic CD11c⁺ DC maturation.

To determine whether NK cells are required for optimal DC activation in vivo, WT mice were treated with anti-asialo-GM1 antibody (Ab) and infected with P. chabaudi AS. As we previously observed (28), depletion of NK cells in WT mice led to higher peak parasitemia and a higher recrudescence later during the infection than with control mice treated with isotype control rabbit IgG or with PBS-treated mice (Fig. 7A). No mortality was observed among the three groups of infected animals (data not shown). CD11c⁺ DCs purified from the spleens of NK cell-depleted infected mice expressed lower levels of costimulatory molecules (CD40 and CD86) than DCs from isotype control Ab-treated mice (Fig. 7B). This observation suggests that DC maturation was deficient in the absence of NK cells during blood-stage malaria. In addition, DCs from NK cell-depleted mice produced significantly less IL-12p40 and IL-12p70 but more IL-2 in response to iRBC in vitro than isotype control Ab-treated mice (Fig. 7C). Soluble IL-15 was not detected in the DC cultures (data not shown). Importantly, impairment of DC maturation and IL-12 production in NK cell-depleted mice resulted in significantly lower CD4⁺ T-cell proliferation (Fig. 7D) and IFN-γ production (Fig. 7E) than with DCs from isotype control Ab-treated mice. These data provide important evidence that NK cells are required in vivo for optimal DC maturation, as well as IL-12-dependent stimulation of Th1 cell development, leading to high IFN-γ production and efficient immune control of blood-stage malaria infection.

FIG. 7.

Depletion of NK cells leads to increased parasitemia and impaired DC maturation and Th1 priming. B6 mice were injected intravenously with sterile PBS, control IgG, or anti-asialo-GM1 Ab at days −7, −4, 0, and 2 p.i. (A) Courses of parasitemia in PBS-treated mice and in mice treated with control rabbit IgG or anti-asialo-GM1 Ab. Mice were infected intraperitoneally with 10⁶ P. chabaudi AS iRBC. (B) Splenic CD11c⁺ DCs from day 5 infected mice, either control or NK cell depleted, were analyzed by flow cytometry for costimulatory molecule expression. The cells were gated on the CD11c⁺ signal and analyzed for CD40, CD80, and CD86 expression (MFI). (C) IL-12p40, IL-12p70, and IL-2 production by splenic DCs from infected mice in response to iRBC in vitro. In panels D and E, CD11⁺ DCs from naïve mice (Day 0 DC) or from day 5 infected mice treated with control IgG or anti-asialo-GM1 Ab were cocultured with CD4⁺ T cells from naïve mice for 36 to 48 h, and the proliferation (D) and IFN-γ production (E) were determined. As a control, CD4⁺ T cells were cultured with iRBC alone (T cell). In panels A to E, the data are representative of two independent experiments, each with four mice per group. The asterisks indicate significant differences between DCs from control mice versus DCs from mice treated with anti-asialo-GM1Ab: *, P < 0.05; **, P < 0.01; and ***, P < 0.001. The error bars indicate SEM.

DISCUSSION

Here, we investigated the interactions between NK cells and DCs as accessory cells in the mouse model of P. chabaudi AS infection. Determination of DC maturation and function in P. chabaudi AS-infected WT B6 mice revealed that, within 4 to 5 days p.i., splenic CD11c⁺ DCs captured iRBC, expressed high levels of costimulatory molecules, and produced important proinflammatory and immunoregulatory cytokines, especially the type 1-promoting cytokine IL-12. These findings are in agreement with previous studies of DC function during early infection with this Plasmodium species, which causes a nonlethal infection in WT B6 mice (23, 25, 46). We also compared DC function in WT B6 mice to those of DCs from P. chabaudi AS-infected IL-15^−/− and IL-12p40^−/− mice, which differ in their abilities to efficiently control this infection. Our results demonstrated that recognition and capture of iRBC by DCs were intact in the absence of IL-15 or IL-12p40. DCs from infected IL-12p40^−/− mice, which are more susceptible to P. chabaudi AS infection than WT or IL-15^−/− mice, expressed significantly lower levels of CD40, CD80, and CD86 than DCs from infected WT mice. These findings support the notion that DC maturation is correlated with the severity of Plasmodium infection, as shown recently by Wykes et al. for mice infected with various nonlethal and lethal plasmodia (45, 46).

The exposure of DCs to iRBC may lead to expansion of DC subsets that differ in their abilities to induce effector CD4⁺ T cells (21, 36, 44). Depending on their functional maturation or subset affiliation, DCs may also promote regulatory T-cell differentiation (22). Differences in T-cell priming may be due to underlying differences in cytokine production by DCs from Plasmodium-infected mice. During early P. chabaudi AS infection, DCs from IL-15^−/− mice produced levels of IL-12p70 equivalent to those produced by DCs from WT mice in response to stimulation in vitro with iRBC. In contrast to the observation of impaired IL-2 production by bone marrow-derived DCs from IL-15^−/− mice in response to Toll-like receptor-dependent stimuli (13), DCs from infected WT and IL-15^−/− mice stimulated with iRBC produced equivalent levels of IL-2 in vitro. Although there were no differences among the mouse strains in DC production of IL-10, DCs from infected IL-15- and IL-12p40-deficient mice produced significantly less IFN-γ than WT DCs in response to iRBC.

Of note, DCs from infected IL-12p40-deficient mice produced high levels of IL-2. To our knowledge, the ability of DCs from P. chabaudi AS-infected IL-12p40^−/− mice to produce high levels of IL-2 has not been reported previously. IL-2 is produced by DCs in response to stimulation with Toll-like receptor ligands and acts as a key molecule regulating NK cell IFN-γ production following bacterial challenge (18, 19). In addition, IL-2, together with transforming growth factor β, plays a role in the maintenance and activation of CD4⁺ CD25⁺ Foxp3⁺ regulatory T cells, which have been implicated in suppressing immunity to malaria (9, 20, 35, 42, 43). In our hands, blocking DC-derived IL-2 in cocultures of DCs from infected mice and NK cells from naïve mice significantly reduced NK cell IFN-γ production, but the impact of neutralizing DC-derived IL-2 on subsequent CD4⁺ T-cell activation is unclear.

In addition to their ability to mature phenotypically and to produce Th1-promoting cytokines in vitro during early P. chabaudi AS infection, splenic CD11c⁺ DC from infected WT mice promoted high IFN-γ production by resting NK cells from naïve WT mice via a mechanism dependent in part on cell-cell contact. Based on intracellular IFN-γ expression in gated CD11c⁺ DX5⁻ and CD11c⁻ DX5⁺ cells, the possible contribution of interferon-producing killer DCs or CD11c⁺ DX5⁺ cells, a subset of NK cells that share markers associated with DCs (7, 8, 41), as a source of IFN-γ in the DC and NK cell cocultures was eliminated. Moreover, DX5⁺ cells, unlike CD11c⁺ cells, neither ingested iRBC nor produced IL-12, but killed NK-sensitive YAC-1 cells in vitro (data not shown).

Consistent with the important role of IL-12 in promoting NK cell production of IFN-γ during infection with intracellular parasites, including malaria (28, 40), DCs from malaria-infected IL-12p40^−/− mice were unable to elicit high IFN-γ production by NK cells. IL-12 and, to a lesser extent, IL-18 have been implicated in DC stimulation of human NK cell IFN-γ secretion in vitro in response to P. falciparum iRBC (3, 4, 29). As shown by Newman et al. (29) for P. falciparum and by our findings reported here for P. chabaudi AS, IL-2 contributes, albeit to a lesser extent, to promoting IFN-γ secretion by NK cells during malaria. IFN-γ production by NK cells was impaired in cocultures of WT NK cells and DCs from infected IL-12p40^−/− mice despite the ability of these DCs to produce high levels of IL-2. This observation suggests that high IL-2 levels were unable to compensate for the lack of IL-12 production by IL-12p40^−/− DCs. During P. chabaudi AS malaria, DCs from IL-15^−/− mice promoted NK cell IFN-γ secretion to an extent similar to that of WT DCs, supporting the observation in human peripheral blood mononuclear cells that IL-15 does not contribute to NK cell IFN-γ production in response to P. falciparum iRBC (29). Together, these findings indicate that DC-derived IL-12 is the major cytokine regulating NK cell IFN-γ production during malaria.

Our findings further revealed that NK cells from infected WT mice, in turn, stimulated DCs to mature and secrete cytokines, including IL-12. NK cell-activated DCs provided a link between innate and adaptive immunity, as evidenced by their ability to prime naïve CD4⁺ T cells to proliferate and produce Th1-polarizing cytokines, indicating that these DCs were fully licensed antigen-presenting cells. The ability of malaria-activated NK cells to enhance the capacity of DCs to produce IL-12 and to promote IFN-γ-secreting CD4⁺ T cells required both cell-cell contact and NK cell-derived IFN-γ. These findings are consistent with the demonstration that in vivo depletion of NK cells in WT mice results in increased parasite growth and delayed clearance. We observed that the inability of NK cell-depleted WT mice to efficiently control malaria was correlated with defects in DC maturation and function. In particular, DCs from NK cell-depleted WT mice showed a severely compromised ability to produce IL-12 and promote IFN-γ-producing CD4⁺ T cells.

In conclusion, our results provide in vitro and in vivo evidence for the importance of DC and NK cell cross talk during early blood-stage P. chabaudi AS infection. Together, our findings define the mechanisms involved in the reciprocal activation of these cell types leading to the induction of CD4⁺ Th1-dependent protective immunity to malaria. These findings may help to identify relevant immunological targets for the development of an effective malaria vaccine and novel immunotherapy against the devastating worldwide menace of malaria.