Interleukin-15 Enhances Innate and Adaptive Immune Responses to Blood-Stage Malaria Infection in Mice

Abstract

Compared to C57BL/6 wild-type mice, interleukin-15^−/− (IL-15^−/−) mice showed delayed clearance of Plasmodium chabaudi AS infection, lower type 1 cytokine production, impaired dendritic cell and NK cell functions, and lower titers of malaria-specific antibodies. Thus, IL-15 supports early control and timely resolution of blood-stage malaria through promotion of Th1-dependent innate and adaptive immune responses.

In the experimental model of Plasmodium chabaudi AS infection, resistant C57BL/6 (B6) mice show moderate levels of acute primary parasitemia and resolve the infection by 4 to 5 weeks postinfection (p.i.), whereas susceptible A/J mice develop fulminating parasitemia and succumb to infection by 10 to 12 days p.i. (31, 32). Expression of resistance, including recovery from infection, requires CD4⁺ T cells, macrophages, NK cells, and gamma interferon (IFN-γ) production during the acute phase of infection, as well as Th1-dependent antibody responses during the curative phase (21, 29, 32, 36). Our laboratory has previously shown the important role of interleukin-12 (IL-12) in the induction of IFN-γ-dependent protective immunity to blood-stage malaria infection (32, 33). Other factors released during the early immune response may also influence Th1/Th2 effector choice and activation of downstream immune responses. A possible candidate is IL-15, a cytokine that promotes the expansion and activation of type 1 immune responses. In this study, we investigated the role of endogenous IL-15 in innate and adaptive immune responses to blood-stage P. chabaudi infection.

Previous studies have implicated a role for IL-15 in host resistance to intracellular pathogens, including Salmonella (13), Listeria (14), Toxoplasma (16), and Mycobacterium (39) spp. The mechanisms by which IL-15 enhances survival and host immunity to these pathogens involve the promotion of IFN-γ production, NK cell expansion and activation, and increased survival and cytolytic activity of γδ T cells or CD8⁺ T cells (13, 14, 16, 39). The role of IL-15 in the development of protective immunity to blood-stage malaria infection is not well understood. Serum IL-15 is undetectable in patients with complicated malaria involving multiple organ dysfunctions, although higher parasitemia correlates with elevated IL-10 and IL-12 levels (12). However, IL-15 increases the size of the subset, the survival, and the parasiticidal activity of γδ T cells in human peripheral blood mononuclear cells cultured with Plasmodium falciparum (9). It remains unclear whether IL-15 is involved in antimalarial immunity mediated by dendritic cells (DCs), NK cells, or Th1-dependent antibody. The results presented here show that IL-15 is required for type 1 cytokine production in vivo, NK cell responses, optimal IL-12 and IFN-γ synthesis by DCs, and malaria-specific antibody responses, all of which contribute to the early control and timely resolution of blood-stage malaria infection.

To determine the role of IL-15 in protective immunity to blood-stage malaria, the course of a primary P. chabaudi infection was monitored in wild-type (WT) and IL-15^−/− mice. Breeding pairs of IL-15^−/− mice on the B6 background were kindly provided by Jacques Peschon (Amgen, Seattle, Wash.). IL-15^−/− mice were generated by targeted disruption of the IL-15 gene in B6-derived embryonic stem cells and identified by PCR analysis (15). Age-matched littermates (IL-15^+/+) or B6 mice (Charles River Breeding Laboratories, St. Constant, Quebec, Canada) were used as WT controls. Female mice were used in all experiments and were maintained in the animal facility of the Montreal General Hospital Research Institute (Montreal, Quebec, Canada). Infections were initiated by intraperitoneal injection of 10⁶ P. chabaudi parasitized red blood cells (PRBC). All statistical analyses were performed using SAS (SAS Institute, Cary, N.C.), and a P of <0.05 was considered significant.

IL-15^−/− mice had an earlier peak in parasitemia than WT mice, developed a small recrudescent parasitemia of 4 to 6% at 24 to 32 days p.i., and failed to resolve their infections by day 32 p.i. (Fig. 1A). IL-15^−/− mice continued to demonstrate low parasitemias of 1 to 2% as late as day 48 p.i., whereas 100% of WT mice cleared the infection by day 32 p.i. (Fig. 1A, inset). Although 10% of IL-15^−/− mice died by day 13 p.i., in contrast to the WT mice, 100% of which survived, this difference was not statistically significant (Fig. 1B).

FIG. 1.

Course of parasitemia (A) and survival rate (B) in WT and IL-15^−/− (knockout [KO]) mice infected intraperitoneally with 10⁶ P. chabaudi PRBC. Parasitemia was monitored by counting the percentage of infected cells per 400 RBC for each mouse per time point on blood smears prepared as described previously (32). The inset (A) shows the course of parasitemia in the chronic stage of infection. Results were pooled from four independent experiments, each containing 8 to 10 mice per group, and are presented as means ± standard errors of the mean. *, WT versus KO mice, P < 0.05, as determined by repeated-measure analysis of variance. Differences in survival rates were not statistically significant as determined by the chi-square test.

The delayed parasite clearance observed in IL-15^−/− mice in this study suggested impaired production of type 1 cytokines. IL-15 promotes CD40-dependent IL-12 production by monocytes (2) and costimulates IFN-γ production by NK and T cells in synergy with IL-12 (3, 10, 11) and IL-21 (34). We have previously reported that protection against blood-stage P. chabaudi infection is induced by IL-12 and mediated critically by IFN-γ (31, 32, 35). During the first week p.i., IL-15^−/− mice had peak levels of IL-12 p70, IFN-γ, and TNF-α in serum that were significantly lower than the levels found in WT mice (Table 1). Peak serum IL-10 levels were comparable in WT and IL-15^−/− mice. Splenocytes from infected IL-15^−/− mice appeared to be as proficient as those from WT mice at producing IL-12 and IL-10 in response to PRBC in vitro. However, despite their comparable levels of IL-12 and IL-10 production, IL-15^−/− splenocytes produced significantly less IFN-γ and TNF-α in vitro than WT splenocytes.

TABLE 1.

Peak cytokine levels in serum and spleen cell supernatants following P. chabaudi infection in WT and IL-15^−/− mice^a

Cytokine	Level in:
	Serum		Spleen
	WT	KO	WT	KO
IL-12 p70 (ng/ml)	13.27 ± 0.88	8.04 ± 2.13*	4.98 ± 0.71	4.02 ± 0.74
IFN-γ (ng/ml)	8.06 ± 2.56	2.75 ± 0.49*	34.65 ± 5.16	10.15 ± 1.22*
TNF-α (pg/ml)	105.86 ± 19.32	2.41 ± 1.63**	113.97 ± 6.64	72.46 ± 15.36**
IL-10 (ng/ml)	1.274 ± 0.35	1.176 ± 0.54	1.53 ± 0.15	1.405 ± 0.32

^aWT and IL-15^−/− mice were infected intraperitoneally with 10⁶ P. chabaudi PRBC. The peak levels of each cytokine during the first week p.i. in vivo and in vitro are presented. On the day of sacrifice, sera were obtained, and single-cell suspensions of splenocytes were prepared as described previously (35, 36). Sera were stored at −20°C until assayed for cytokine levels by enzyme-linked immunosorbent assay (ELISA). Splenocytes were stimulated for 48 h with PRBC (10⁶/ml), and supernatants were assayed for cytokines by ELISA using paired capture and detection antibodies as described previously (31, 32, 35). Data are presented as means ± standard errors of the means of results from 10 to 12 mice pooled from three independent experiments. Comparisons between WT and knockout (KO) mice were determined by analysis of variance. *, P < 0.05; **, P < 0.01.

IL-15 activates NK cells, macrophages, and DCs, the last of which are influential in detecting infection and regulating innate immune responses. IL-15 has been shown to stimulate IL-12 production by DCs and monocytes in response to lipopolysaccharide or CD40 ligation (2, 27). Following P. chabaudi infection, IL-15^−/− mice had significantly lower numbers of CD11c⁺ DCs in the spleen than WT mice (at day 7 p.i., 12.03 × 10⁶ [WT mice] versus 7.82 × 10⁶ [IL-15^−/− mice]; Mann-Whitney U test, P < 0.05). To address the role of IL-15 in inducing cytokine production by DCs during blood-stage malaria, splenic CD11c⁺ cells from infected WT and IL-15^−/− mice were cultured with no stimulation (medium alone), with PRBC, with recombinant human IL-15 (rhIL-15), or with a combination of PRBC and rhIL-15. WT DCs stimulated with PRBC in vitro produced significantly less IL-12 p70 than did unstimulated cultures (Fig. 2A), a result consistent with the ability of Plasmodium-infected erythrocytes to suppress IL-12 p40 gene induction in macrophages (40). The addition of rhIL-15, in the absence of PRBC, significantly enhanced the secretion of IL-12 p70 by DCs. Importantly, the elevated levels of IL-12 p70 in IL-15-stimulated cultures were observed only with WT DCs, not with IL-15^−/− DCs.

FIG. 2.

Stimulation with IL-15 in vitro enhances IL-12 p70 (A) and IFN-γ (B) production by splenic CD11c⁺ DCs from infected WT mice. Spleens were digested with collagenase D (Roche, Laval, Quebec, Canada), and low-density cells were collected using Nycoprep (Axis-Shield, Oslo, Norway). Cells were further purified using anti-CD11c microbeads (Miltenyi Biotec, Auburn, Calif.), and the resulting cells were 80 to 88% positive for CD11c as determined by flow cytometry. CD11c⁺ cells (10⁶/well) were cultured in medium alone or in medium plus PRBC (10⁶/ml), rhIL-15 (100 ng/ml; Amgen), or a combination of PRBC and rhIL-15. After 48 h of incubation, supernatants were assayed for IL-12 p70 and IFN-γ by enzyme-linked immunosorbent assay. Data (n = 4 per group) are presented as median values ± 95% confidence intervals and are representative of two replicate experiments. ND, nondetectable. *, WT versus knockout (KO) mice, P < 0.05, as determined by the Mann-Whitney U test.

Recent studies which challenge the classical paradigm in which only lymphoid cells produce IFN-γ show that activated antigen-presenting cells secrete IFN-γ in an IL-15-dependent manner (20, 27). Based on this evidence, it was proposed that IL-15 is an early host factor that initiates a positive autocrine feedback loop for further Th1-type cytokine synthesis by DCs (20, 27). Therefore, we wanted to determine whether IL-15 modulates IFN-γ production by DCs during blood-stage malaria. DCs from infected WT mice produced significantly higher levels of IFN-γ in response to PRBC or rhIL-15 in vitro than did DCs from IL-15^−/− mice (Fig. 2B). Stimulation with both rhIL-15 and PRBC further increased IFN-γ secretion, but this secretion occurred to a significantly greater extent in WT DCs than in IL-15^−/− DCs. Taken together, these results suggest that IL-15 supports optimal synthesis of IL-12 p70 and IFN-γ by splenic CD11c⁺ DCs during the innate immune response to P. chabaudi infection. The ability of IL-15 to increase both IL-12 and IFN-γ production by DCs as shown here and in other studies (20, 27), as well as the IL-12 responsiveness of antigen-presenting cells (27), places IL-15 in a critical position to induce Th1-type innate immune responses. Recent work in our laboratory suggests that IL-15 enhances the ability of DCs from malaria-infected mice to stimulate IFN-γ production by DX5⁺ NK cells and CD4⁺ T cells (R. Ing and M. M. Stevenson, unpublished data).

IL-15 maintains NK cell survival through activation of the lymphocyte anti-apoptotic factor Bcl-2 (4). Following P. chabaudi infection, IL-15^−/− mice had significantly lower numbers of NK cells in the spleen (at day 7 p.i., 1.59 × 10⁶ [WT mice] versus 0.86 × 10⁶ [IL-15^−/− mice]; Mann-Whitney U test, P < 0.05) and higher percentages of NK cells that stained positive for Annexin V, a surface marker of apoptosis, than did WT mice (Fig. 3A). These data are consistent with previous studies showing that IL-15 is required for NK cell survival and proliferation during homeostasis and infection (4, 15). In addition, several studies indicated that IL-15 increases NK cell cytotoxicity and IFN-γ production (10, 15, 17), whereas others showed that IL-12 and IL-18 are more effective than IL-15 alone (17, 25). Since IL-15^−/− splenocytes were proficient at producing IL-12 p70 in vitro (Table 1), we questioned whether splenic NK cells from IL-15^−/− mice were capable of cytotoxicity and IFN-γ production. As shown in Fig. 3B, WT NK cells showed increasingly higher levels of cytolytic activity against NK cell-sensitive murine YAC-1 targets during the first week of P. chabaudi infection. Importantly, IL-15^−/− NK cells exhibited undetectable or significantly lower cytolytic activity at 0 to 4 days p.i., suggesting that IL-15 deficiency impairs not only the expansion but also the function of splenic NK cells during blood-stage malaria infection. Interestingly, IL-15^−/− NK cells showed increased cytolytic activity after day 4 p.i., which likely reflects the availability of other factors, such as IL-12 and IL-18, in IL-15^−/− mice to induce NK cell cytotoxicity, albeit in a suboptimal and severely delayed manner.

FIG. 3.

IL-15 deficiency impairs NK cell survival and function, as determined by NK cell apoptosis (A), enriched NK cell cytotoxicity (B), IFN-γ production (C), and intracellular IFN-γ expression (D). Splenic NK cells were purified using anti-DX5 microbeads (Miltenyi Biotec) and were >83% positive for DX5 as determined by flow cytometry. (A) NK cell apoptosis was determined by staining FcR-blocked spleen cells with antibodies to Annexin V and DX5 (BD Biosciences, Mississauga, Ontario, Canada). (B) NK cell cytotoxicity was measured by a standard ⁵¹Cr release assay performed as described previously (21). Briefly, DX5⁺ NK cells were plated with YAC-1 cells at an optimum effector cell/target cell ratio of 10:1 and presented as percentage of specific lysis calculated as described previously (21). (C) DX5⁺ NK cells (5 × 10⁵ cells/well) were cultured with rhIL-15 (100 ng/ml) for 72 h, and supernatants were assayed for IFN-γ by enzyme-linked immunosorbent assay. (D) DX5⁺ NK cells were stimulated in vitro with Golgi Stop (BD Biosciences), phorbol myristate acetate (Sigma, St. Louis, Mo.), and ionomycin (Sigma) for 2 h and then stained intracellularly for isotype control (outlined peak) or IFN-γ (shaded curve). Representative histograms of gated DX5⁺ cells are shown, with the percentages expressing intracellular IFN-γ. Data (n = 4 per group) are presented as median values ± 95% confidence intervals and are representative of two replicate experiments. *, WT versus knockout (KO) mice, P < 0.05, as determined by the Mann-Whitney U test.

Previous work in our laboratory has shown that NK cell IFN-γ production, not cytotoxicity, plays a major role in mediating protective immunity to P. chabaudi infection (21). Accordingly, we examined IL-15-stimulated IFN-γ production by splenic DX5⁺ NK cells purified from WT and IL-15^−/− mice. At 0 to 4 days p.i., IL-15^−/− NK cells produced significantly less IFN-γ than did WT cells (Fig. 3C). Although it is necessary to maintain NK cell survival in vitro, the addition of rhIL-15 to NK cell cultures may result in the underestimation of the effect of endogenous IL-15 deficiency on NK cell IFN-γ production in vivo. Therefore, we determined the percentage of DX5⁺ NK cells that expressed intracellular IFN-γ ex vivo by flow cytometry. As shown in Fig. 3D, significantly lower percentages of NK cells from infected IL-15^−/− mice expressed intracellular IFN-γ than did cells from WT mice. These results indicate that IL-15 is required during the early immune response to blood-stage malaria to support optimal NK cell IFN-γ production.

The inability of soluble rhIL-15 to fully restore IL-12 and IFN-γ production by IL-15^−/− DCs and NK cells, respectively, may reflect the fact that the potency of human IL-15 to stimulate mouse spleen cells is lower than that of the murine protein (8); murine rIL-15 was not available at the time our experiments were performed. It may also be explained by the existence of IL-15 as two isoforms with distinct biological activities. The IL-15-SSP isoform is stored intracellularly and mobilized to the plasma membrane following antigenic stimulation, while the IL-15-LSP isoform functions as a secretory cytokine (23, 38). The membrane-bound IL-15-SSP contributes the majority of the biological effects of IL-15, in part due to its ability to function as a receptor to activate cell-signaling pathways that result in cellular adhesion and proinflammatory cytokine production (22, 24, 26). Indeed, IL-15 deficiency was associated with lower intracellular IFN-γ expression by DX5⁺ NK cells during P. chabaudi infection. In agreement with our observations, Ohteki et al. (27) also reported that, unlike WT cells, exogenous IL-15 failed to completely rescue the reduced IL-12 and IFN-γ production by lipopolysaccharide-stimulated IL-15^−/− DCs and macrophages. Taken together, these data suggest that IL-15, in the membrane-bound form, may be required during development for maximum type 1 cytokine production by innate immune cells in response to intracellular infection.

Given that antibody-mediated immunity is necessary for parasite clearance (18, 36), we questioned whether the delayed resolution of P. chabaudi infection observed in IL-15^−/− mice (Fig. 1A) was associated with impaired antibody production. Previous work in our laboratory has shown that Th1-dependent immunoglobulin G2a (IgG2a) and possibly IgG3 are preferentially required for rapid clearance of P. chabaudi infection (36). Since IL-15 synergizes with IL-12 to promote IFN-γ production and Th1-type responses, we expected to observe altered parasite-specific IgG2a and IgG3 responses in IL-15^−/− mice. As shown in Fig. 4, however, IL-15^−/− mice had significantly lower titers of total Ig, IgM, and the IgG subclasses IgG1, IgG2a, IgG2b, and IgG3. The lower levels of malaria-specific antibodies observed in IL-15^−/− mice may explain the inability of IL-15^−/− mice to resolve the infection as rapidly as WT mice. Moreover, these results suggest that IL-15 supports general antibody production, irrespective of Th1 or Th2 phenotypic affiliation, during blood-stage malaria. Studies have shown that IL-15 stimulates B-cell proliferation and Ig production (1) and modulates Ig class switching (19, 28), which supports our observation that IL-15 is important for augmenting antibody responses that are selectively induced by other immunoregulatory factors.

FIG. 4.

IL-15^−/− mice have defects in Ig production during the chronic stage of blood-stage malaria infection. P. chabaudi-specific total Ig (A), IgM (B), IgG1 (C), IgG2a (D), IgG2b (E), and IgG3 (F) were measured in sera of infected WT and knockout (KO) mice (n = 8 per group). Serum levels of P. chabaudi AS-specific antibody isotypes and subclasses were determined by enzyme-linked immunosorbent assay as described previously (36). Serum samples were serially diluted twofold, and 50 μl of each dilution was added to 96-well plates coated with parasite antigen prepared as described previously (36). Horseradish peroxidase-conjugated goat anti-mouse isotype/subclass monoclonal antibodies (Southern Biotechnology Associates, Birmingham, Ala.) were added, and reactivity was visualized using ABTS [2,2′-azinobis(3-ethylbenzthiazolinesulfonic acid)] substrate (Roche). Antibody levels are expressed as endpoint titers, the reciprocal of the highest dilution that yielded the background optical density. *, WT versus KO mice, P < 0.05, as determined by analysis of variance.

The results presented here provide the first in vivo evidence that IL-15 contributes to the early control and timely resolution of blood-stage P. chabaudi infection. We showed that IL-15 enhanced protective mechanisms involving both innate and adaptive responses. Specifically, IL-15 is required for optimal IL-12 and IFN-γ secretion by DCs, NK cell IFN-γ synthesis, and production of malaria-specific antibodies. Despite impairments in these immune functions, IL-15 deficiency did not markedly exacerbate mortality or parasitemia during the acute phase. Although the majority of IL-15^−/− mice survived the infection, these mice showed significantly higher parasitemia than WT mice during the chronic phase, particularly at 12 to 32 days p.i. following peak parasitemia, and failed to resolve the infection by 32 days p.i. Therefore, IL-15 appears to play a supporting, albeit not indispensable, role in enhancing the innate and adaptive immunity needed to control and completely clear a primary P. chabaudi infection in a timely manner.

Previous work in our laboratory has shown that vaccine-induced protection against blood-stage P. chabaudi infection requires CD4⁺ T cells and IFN-γ production (37). Early production of IFN-γ by NK cells and γδ T cells is associated with resolution of nonlethal Plasmodium yoelii and P. chabaudi infections, but is absent in lethal P. yoelii and Plasmodium berghei infections (5). Moreover, studies of liver-stage P. yoelii infection have demonstrated the critical role of early IFN-γ production by NK cells, CD8⁺ T cells, and γδ T cells in mediating protective immunity to sporozoites (6, 7). Accordingly, the most effective vaccination strategies are those that trigger early and strong IFN-γ production, which in turn induces long-lasting, IFN-γ-mediated protective immunity. Although, on its own, IL-15 is not essential for IFN-γ and antibody production, the results presented in this study and by other investigators (19, 28, 30) demonstrate the benefit of IL-15 as an adjuvant to induce maximal IFN-γ and antibody responses. The potential for IL-15 use in vaccine systems will necessitate further studies that examine how IL-15 interacts with other factors in the complex cytokine network to orchestrate protective immune responses to Plasmodium parasites.