Deficiency in Tumor Necrosis Factor Alpha Activity Does Not Impair Early Protective Th1 Responses against Blood-Stage Malaria

Abstract

Blood-stage Plasmodium chabaudi AS infection was controlled by 4 weeks in mice with deletion of tumor necrosis factor p55 and p75 receptors (TNFR-knockout [KO]) and control wild-type (WT) mice, although female TNFR-KO mice showed slightly but significantly higher parasitemia immediately following the peak. Serum interleukin 12 (IL-12) p70 and gamma interferon (IFN-γ) levels were similar but tumor necrosis factor alpha levels were significantly higher in TNFR-KO mice than in WT controls. Splenic IL-12 receptor β1 and β2 and IFN-γ mRNA expression, as well as spleen cell production of IFN-γ and IL-4, were comparable in both mouse types, but IL-10 production was significantly higher in cells from TNFR-KO mice than in cells from WT mice. Lipopolysaccharide-induced NO secretion by splenic macrophages in vitro was significantly reduced but systemic NO₃⁻ levels were similar in infected TNFR-KO and WT mice.

Tumor necrosis factor alpha (TNF-α) has been implicated in protective as well as pathological roles in resistance of inbred mouse strains to bacterial and protozoan parasite infections, including blood-stage malaria (4, 5, 17, 21, 34). The balance between protective versus pathological actions of TNF-α depends on several factors, including the quantity, timing, and duration of TNF-α production, as well as the organ-specific site of synthesis (5, 15). The biological activities of TNF-α and TNF-β, which shares many functions with TNF-α, are mediated by two structurally related but functionally distinct receptors (TNFRs) known as TNFRp55 or TNFR1 and TNFRp75 or TNFR2, respectively (2, 3). Most of the common biological actions of TNF are attributed to signaling via TNFRp55 (23, 26). TNFRp75 is thought to function both as a TNF antagonist and as an agonist by facilitating the cell surface interaction between TNF and TNFRp55 (22).

Recently, mice with gene-targeted deletion of both the TNF p55 and p75 receptors (TNFRp55p75^−/−) have been used to study the role of TNF-α in host defense against parasitic infections. Toxoplasma gondii infection resulted in higher parasite burdens and 100% mortality within 20 to 26 days in TNFRp55p75^−/− mice compared with wild-type (WT) control mice that survived for at least 60 days (37). In contrast, parasite burdens and susceptibility to infection were comparable in TNFRp55p75^−/− and WT controls infected with Mycobacterium avium (6).

Previous work from our laboratory demonstrated that an early, Th1-associated increase in TNF-α is involved in resistance of C57BL/6 (B6) mice against blood-stage P. chabaudi AS malaria (15). Furthermore, we demonstrated that the mechanism of recombinant interleukin-12 (rIL-12)-induced protection of susceptible A/J mice against P. chabaudi AS infection is dependent on TNF-α, acting in concert with gamma interferon (IFN-γ) and nitric oxide (NO) (31). Here, mice genetically deficient in both TNFRs were used to further define the role of TNF in resistance to blood-stage malaria. Our results reveal that, similar to WT controls, doubly deficient TNFRp55p75^−/− mice produce IL-12 in vivo, mount unimpaired Th1 responses, and clear P. chabaudi AS malaria by 4 weeks postinfection.

Mice, 9 to 10 weeks old, were age and sex matched in all experiments. TNFRp55p75^−/− mice were bred in the animal facilities of the Montreal General Hospital Research Institute from breeders provided by Genentech, Inc., San Francisco, Calif. As WT controls, (B6 × 129)F₁ from Jackson Laboratories (Bar Harbor, Maine) were used. P. chabaudi AS was maintained as previously described (24). Infection was initiated by intraperitoneal injection of 10⁶ P. chabaudi AS-infected erythrocytes (PRBC), and the course of infection was monitored by previously described procedures (24).

At the indicated times, blood samples were obtained from WT or TNFRp55p75^−/− mice by cardiac puncture and allowed to clot, and sera were separated by centrifugation at 13,800 × g for 3 min. Sera were kept at 4°C and immediately analyzed for IL-12 p70, IFN-γ, and TNF-α levels by two-site sandwich enzyme-linked immunosorbent assays (ELISAs) as previously described (27, 30, 31). Serum NO₃⁻ levels were measured by the method described by Rockett et al. (25).

Single-cell suspensions of spleen cells and adherent splenic macrophages were prepared as previously described (28, 30). Splenic macrophages (10⁶) were cultured for 24 h in 500 μl of freshly added medium as a control, with PRBC (2 × 10⁶/ml) or Escherichia coli O127:B8 lipopolysaccharide (LPS) (1 μg/ml) (Difco, Detroit, Mich.). Cell culture supernatants were removed and assayed for nitrite levels by the Griess reaction (9, 14). Where indicated, 500-μl aliquots of unfractionated spleen cells (4 × 10⁶/ml), in medium or stimulated with concanavalin A (ConA) (5 μg/ml) or PRBC (2 × 10⁶/ml), were cultured for 48 h. Cell-free culture supernatants were assayed for cytokine levels by two-site sandwich ELISAs. For IL-4, the capturing and detecting antibodies (Abs) were BVD4-1D11 and BVD6-24G2, respectively (PharMingen, Mississauga, Ontario, Canada). For IL-10, the capturing Ab was JES 5.2A5 (American Type Culture Collection, Rockville, Md.) and the detecting Ab was SXC 1 (PharMingen).

Reverse transcription-PCR (RT-PCR) was performed as previously described (16) to detect changes in cytokine or cytokine receptor mRNA levels. To determine optimal cycling conditions, titrations of input cDNA were performed followed by PCR amplification to ensure that, for the selected number of cycles, a linear relationship exists between input cDNA and PCR product. Both positive and negative controls were included in each assay to ensure efficacy of the reaction and to rule out possible cDNA contamination of reagents. The housekeeping gene glucose-6-phosphate dehydrogenase (G6PDH) was simultaneously amplified in each assay mixture to verify that equal amounts of cDNA were added in each PCR mixture. Nucleotide sequences for primers and probes for IFN-γ (1), IL-12 receptor (IL-12R) β1 and β2 (27), and G6PDH (16) were used as previously published. After hybridization and washing, cytokine or cytokine receptor mRNA was detected by autoradiography with Kodak Biomax MR film (Rochester, N.Y.). The intensities of bands corresponding to specific cytokines were analyzed by high-resolution optical densitometry (SciScan 500; United States Biochemical) and normalized to those of G6PDH.

Results are presented as means ± standard errors of the means (SEMs). Statistical significance of differences in the means for the two groups of mice was determined by Student’s t test. Where three or more groups were compared, analysis of variance, followed by Student-Newman-Keuls method was used.

First, the effects of TNFR deficiency on survival and the course of parasitemia were examined. As shown in Fig. 1, TNFRp55p75^−/− male and female mice recovered from blood-stage P. chabaudi AS malaria, as did WT control mice, by 4 weeks postinfection. Interestingly, blood-stage P. chabaudi chabaudi AS infection resulted in 56% mortality within 20 days in female, but not male, mice with gene-targeted deficiency in IL-10, whereas 100% of heterozygous controls survived (19).

FIG. 1.

Course of P. chabaudi AS infection in male (A) and female (B) mice. WT or TNFRp55p75^−/− animals were infected, and the course of parasitemia was monitored. Data are pooled from two or three replicate experiments and are presented as means ± SEMs of 10 to 15 mice analyzed individually per time point. Statistically significant differences from the values obtained with WT mice on the same day are indicated by an asterisk (P < 0.05).

The levels of primary peak parasitemia in TNFR-deficient animals and WT controls were similar. In female mice, however, parasitemias were significantly higher in TNFR-deficient mice than in WT controls (P < 0.05) on days 9 to 14, immediately following the peak parasitemia (Fig. 1). Both TNFR-KO and WT mice were immune to reinfection (data not shown). These results suggest that TNF might be important but is not a critical requirement for resolving primary blood-stage malaria.

The development of massive splenomegaly, due in part to the dramatic amplification of splenic erythropoiesis that helps combat malaria-induced anemia, was previously correlated with resistance to P. chabaudi AS infection in resistant B6, but not susceptible A/J, mice (29, 38). Treatment of P. chabaudi AS-infected B6 mice with monoclonal antibodies (MAbs) against TNF-α alone or TNF-α and IFN-γ resulted in significant reductions in spleen weight from those of control animals (13). In addition, treatment with MAbs against TNF-α prevented the development of massive splenomegaly in mice infected with Brucella abortus (39). Therefore, it was of interest to determine whether the development of splenomegaly is affected in TNFR-deficient mice compared to WT mice infected with P. chabaudi AS.

As shown in Fig. 2, spleen weights were expressed as splenic ndex (spleen mass/body mass), since male mice, WT or TNFRp55p75^−/−, had significantly higher body weights than their female counterparts (data not shown). Based on earlier work, marked increases in spleen weight can be demonstrated in P. chabaudi AS-infected mice by day 7 postinfection (29). In TNFR-deficient or WT controls, P. chabaudi AS infection resulted in significant and comparable increases in the splenic index on day 7 postinfection over that of uninfected controls (P < 0.05 and P < 0.01 for male WT and TNFRp55p75^−/− mice, respectively [Fig. 2A]; P < 0.001 for female WT and TNFRp55p75^−/− mice [Fig. 2B]).

FIG. 2.

Increases in splenic index (spleen weight/body weight) during P. chabaudi AS infection in male (A) and female (B) mice. Body and spleen weights were determined in uninfected mice and in infected WT or TNFRp55p75^−/− mice on day 7 postinfection. Data are pooled from two replicate experiments and are presented as means ± SEMs of four to eight mice analyzed individually per time point. Statistically significant differences from the values obtained with uninfected controls (∗, P < 0.05; ∗∗, P < 0.01; ∗∗∗, P < 0.001) and with infected male mice (#, P < 0.01; ##, P < 0.001) are shown.

P. chabaudi AS-infected female WT or TNFR-deficient animals had significantly higher splenic indices than their respective male counterparts (P < 0.001 and P < 0.01 for WT and TNFRp55p75^−/− mice, respectively). In contrast, there were no significant differences in splenic indices between uninfected mice, male or female, WT or TNFR-deficient. Thus, TNFR deficiency did not affect the development of splenomegaly following P. chabaudi AS infection. The differences between these results and those of earlier studies using MAb treatment against TNF-α could be related to unknown compensatory mechanisms that develop in TNFRp55p75^−/− mice.

TNF may play a role in the regulation of macrophage IL-12 production. Studies of Mycobacterium bovis BCG infection in mice deficient in TNFRp55 demonstrated a significant impairment in IL-12 synthesis in vivo and by bone marrow-derived monocytes/macrophages in vitro (7). Possibly reflecting the consequences of deficient IL-12 production on host resistance to mycobacterial infections, TNFRp55^−/−, but not WT controls, succumbed to Mycobacterium tuberculosis (8). In another study, serum IL-12 levels were found to be elevated threefold in Corynebacterium parvum-treated TNF^−/− mice over those in TNF^+/+ mice (10). Our earlier work demonstrated an early peak in systemic IL-12 p70 levels in blood-stage malaria-infected B6 mice at day 2 postinfection compared with those in uninfected controls (27). Following P. chabaudi AS infection, there were no significant differences in the early peak of serum p70 levels between TNFR-deficient and their WT controls (Table 1). However, serum p70 levels were significantly higher (P < 0.01) in infected female mice than in male mice, TNFRp55p75^−/− or WT. Based on our earlier studies, peak levels of serum TNF-α, IFN-γ, and NO₃⁻ occur by day 7 postinfection in resistant mice (14, 15, 31). Serum TNF-α levels were significantly higher (P < 0.05) in TNFR-deficient mice than in their WT counterparts at day 7, whereas IFN-γ and NO₃⁻ levels were comparable in both mouse types (Table 1). Basal levels of serum IL-12 p70, IFN-γ, TNF-α, and NO₃⁻ in control uninfected WT and TNFR-deficient mice were not significantly different (data not shown).

TABLE 1.

Levels of IL-12 p70, TNF-α, and NO₃⁻ in serum samples from TNFRp55p75^−/− and WT control mice following infection with P. chabaudi AS^a

Mouse	IL-12 p70 (ng/ml)	IFN-γ (ng/ml)	TNF-α (ng/ml)	NO3− (μM)
Male
WT	4.0 ± 1.0	48.2 ± 3.5	1.0 ± 0.1	22.5 ± 1.7
TNFRp55p75−/−	3.2 ± 0.7	41.6 ± 3.6	1.9 ± 0.3b	17.5 ± 3.0
Female
WT	8.8 ± 0.3c	44.6 ± 3.8	1.1 ± 0.2	22.5 ± 2.6
TNFRp55p75−/−	7.0 ± 0.6c	45.9 ± 4.9	2.2 ± 0.2b	27.5 ± 2.7

^aThe level of IL-12 p70 in serum was analyzed at day 2 and the levels of IFN-γ, TNF-α, and NO₃⁻ in serum were analyzed at day 7 following P. chabaudi AS infection. Data are presented as means ± SEMs of three to eight mice analyzed individually. 

^bStatistically significantly different from the value obtained with WT mice of the same sex (P < 0.05). 

^cStatistically significantly different from the value obtained with male mice with the same genotype (P < 0.01). 

Daily treatment with murine rIL-12 during the first 5 days of P. chabaudi AS malaria was found to rescue susceptible A/J mice from a lethal course of infection (31). Simultaneous treatment with rIL-12 and MAbs against TNF-α and IFN-γ completely abrogated IL-12-induced resistance in A/J mice to blood-stage malaria. Furthermore, a close association was reported between significant up-regulation of TNF-α mRNA levels in the spleen and the induction of early protective Th1 responses in resistant B6 mice during the first week of P. chabaudi AS infection (15). Whereas these studies suggested that both IL-12 and TNF-α play an important role in early Th1-dependent immune responses against blood-stage malaria, it was unclear whether IL-12 synthesis and Th1 responses were events downstream of TNF activity. The results of the present investigation demonstrate that TNF activity is not required for systemic IL-12 production during early blood-stage malaria.

It has been shown that TNF may play a role in the generation of Th1 responses (11). In addition, TNF-α was found to be an important cofactor for IL-12-induced production of IFN-γ by NK cells from mice with severe combined immunodeficiency (SCID) (12, 35). Therefore, we next examined TNFR-deficient and WT hosts for Th1 responses downstream of IL-12 p70 production, namely, up-regulation of splenic mRNA levels for IL-12R β1 and β2 and IFN-γ as well as IFN-γ protein production in vitro by spleen cells recovered from P. chabaudi AS-infected animals. Female WT and TNFRp55p75^−/− mice were selected for these additional experiments. Developing Th2 cells appear to lose mRNA expression for IL-12R β2 while maintaining mRNA expression for IL-12R β1 (32). Evidence suggests that both IL-12R β1 and β2 are required for high-affinity interaction between IL-12 and its receptor complex (32, 36). Our earlier work demonstrated significant increases in IL-12R β1 and β2 mRNA levels in the spleens of blood-stage malaria-infected B6 mice by day 5 postinfection over those of uninfected controls (27).

As shown in Table 2, fold increases in splenic IL-12R β1 mRNA levels in infected mice at day 5 versus uninfected controls were approximately 2 to 3 for TNFR-deficient mice compared with approximately 2 to 5 for WT controls. Splenic IL-12R β2 mRNA levels increased by 1.5- to approximately 3-fold in infected mice at day 5 versus uninfected controls for both TNFR-deficient and WT controls. Based on our earlier experiments with B6 mice, splenic IFN-γ mRNA levels were determined at day 7 when peak IFN-γ mRNA levels were expected. Fold increases in infected versus uninfected controls were approximately 3 in WT and 3 to 6 in TNFRp55p75^−/− hosts (Table 2). Correlating with splenic IFN-γ mRNA levels, there were no significant differences between TNFR-deficient and WT mice in IFN-γ production by unfractionated spleen cells stimulated with ConA or PRBC or in medium controls (Fig. 3A). IL-4 production by spleen cells was comparable, whereas spleen cells from TNFR-deficient mice produced significantly greater quantities (P < 0.05) of IL-10 in medium, ConA, or PRBC than the WT controls (Fig. 3B and C). Whether the increase in splenocyte IL-10 production is a direct or indirect response to TNFR deficiency or whether the increased serum TNF-α in these mice can act via nonclassical TNFRs is presently unknown.

TABLE 2.

Effects of TNFR deficiency on IL-12R β1 and β2 and IFN-γ mRNA levels in the spleen following P. chabaudi AS infection^a

Strain	Fold increaseb
	IL-12R β1	IL-12R β2	IFN-γ
Expt 1
WT	4.7	1.5	3.1
TNFRp55p75−/−	3.0	1.5	6.1
Expt 2
WT	2.3	2.6	3.3
TNFRp55p75−/−	1.8	2.7	2.5

^aRT-PCR was used to detect changes in cytokine or cytokine receptor mRNA levels in whole spleen tissue (experiment 1) or unfractionated spleen cells (experiment 2) from uninfected or P. chabaudi AS-infected female WT or TNFRp55p75^−/− mice. Infected mice were analyzed on day 5 (for IL-12R β1 and β2) or day 7 (for IFN-γ). 

^bFold increase in the ratios of cytokine or cytokine receptor mRNA levels (normalized to those of housekeeping gene G6PDH) in infected versus uninfected controls. Data are pooled from two or three mice in each experiment. 

FIG. 3.

IFN-γ (A), IL-4 (B), and IL-10 (C) production by unfractionated spleen cells and NO synthesis by splenic macrophages (D) in vitro from P. chabaudi AS-infected WT mice at day 7 or TNFRp55p75^−/− mice. Unfractionated spleen cells were cultured for 48 h in the presence of medium, ConA (5 μg/ml), or PRBC (1 × 10⁶ to 2 × 10⁶/ml). Splenic macrophages were cultured for 24 h with medium, LPS (1 μg/ml), or PRBC (2 × 10⁶/ml). Data are presented as means ± SEMs of three to eight mice analyzed individually. Statistically significant differences shown from the values obtained with WT mice (∗, P < 0.05) and with TNFRp557p75^−/− mice (∗∗, P < 0.01) are shown.

TNF-α is an important activator of macrophages for NO synthesis (18). It was previously found that splenic macrophages recovered from resistant B6 mice displayed significantly greater LPS-induced release of NO in vitro than macrophages from susceptible A/J mice (14). Treatment of P. chabaudi AS-infected resistant B6 mice with MAbs against TNF-α resulted in significant decreases in serum nitrate levels from those of untreated controls (13). Furthermore, significant reductions in splenic inducible nitric oxide synthase mRNA levels were observed in P. chabaudi AS-infected resistant B6 mice treated with a combination of MAbs against TNF-α and IFN-γ (13). Hence, NO production by splenic macrophages recovered from malaria-infected TNFR-deficient or WT controls was assessed. As shown in Fig. 3D, LPS- but not PRBC-induced NO synthesis by splenic macrophages in vitro was significantly higher in infected WT mice than in TNFRp55p75^−/− mice (P < 0.01). As described above, systemic nitrate levels in vivo during infection in the two mouse strains were similar. These data suggest the existence of TNF-α-independent pathways for NO production in vivo, possibly involving IFN-γ. Interestingly, mice deficient in both TNFRs or in TNFRp55 alone, but not WT control mice, were protected against lethal endotoxin challenge with the combination of LPS and d-Gal, suggesting the possible relevance of differences between TNFR-deficient animals and WT hosts in LPS responses (22).

It is possible that in vivo TNF-α, in concert with other serum factors, exerts a direct antiplasmodial effect. However, it should be pointed out that in vitro recombinant TNF-α alone has no effect on parasite viability (33). The mice used in the present study lacked both the TNF p55 and p75 receptors; in other words, TNF-α activity was completely absent vis-a-vis the host. In mice with selective gene-targeted deficiency of either p55 or p75 TNFR, but not both, TNF-α signaling could still occur via the remaining TNF receptor. Use of these mice has been particularly useful in dissecting the differential roles of TNF-α signaling through the TNF p55 or p75 receptors in host defense against parasitic infections. For example, TNFRp75^−/− mice were significantly protected from cerebral malaria, whereas TNFRp55^−/− hosts were as susceptible as WT controls (20). In contrast, TNFRp55 deficiency resulted in increased susceptibility to infection with Listeria monocytogenes (23, 26) and M. tuberculosis (8) than with WT controls.

Taken together, our results suggest that TNF-α activity is not a critical requirement for resolving blood-stage infections with P. chabaudi AS malaria. Furthermore, neither IL-12 production nor protective Th1 responses appear to be impaired in the absence of TNF-α activity during early blood-stage malaria.