To address the question of how the murine host responds to a prototypic type 1 cytokine inducer while concurrently undergoing a helminth-induced type 2 cytokine response, C57BL/6 strain animals with patent schistosomiasis mansoni were orally infected with the cystogenic Toxoplasma gondii strain ME49. Schistosoma mansoni infection resulted in a significantly higher mortality rate when mice were subsequently orally infected with ME49, and these animals displayed a defective IFN-γ and NO response relative to animals infected with T. gondii alone. Plasma levels of TNF-α and aspartate transaminase in double-infected mice were greatly elevated relative to mice infected with either parasite alone. Consistent with the latter observation, these animals exhibited severe liver pathology, with regions of coagulative necrosis and hepatocyte vacuolization unapparent in mice carrying either infection alone. Interestingly, mean egg granuloma size was ∼50% of that in mice with S. mansoni infection alone. The exacerbated liver pathology in coinfected mice did not appear to be a result of uncontrolled tachyzoite replication, because both parasite-specific RT-PCR analysis and immunohistochemical staining demonstrated a low number of tachyzoites in the liver. We hypothesize that mortality in these animals results from the high level of systemic TNF-α, which mediates a severe liver pathology culminating in death of the animal.
Infection of mice with the trematode Schistosoma mansoni (Sm)4 results in a chronic infection in which parasite eggs are deposited in the liver and gut by female worms in the portal veins. Approximately 100 million people worldwide carry this infection, which is a serious and sometimes fatal disease in the third world. In mouse models of infection, chronic schistosomiasis mansoni is dominated by a type 2 cytokine response, characterized by production of IL-4 and IL-5. The cytokine IL-4 plays an important role in minimizing host morbidity and mortality, as shown by recent studies demonstrating that IL-4 knockout (KO) mice are unable to survive Sm infection. Death in these animals is attributable to overproduction of inflammatory mediators such as TNF-α and NO, most likely resulting from absence of the anti-inflammatory activity of IL-4 (1). To a large extent, host control of infection appears to be directed toward preventing a massive inflammatory response to the parasite and its products, rather than limiting presence of the parasite itself.
This situation contrasts with that occurring during infection with another even more widespread parasite, the intracellular protozoan, Toxoplasma gondii (Tg). This microbial pathogen is a major opportunistic infection in immunocompromised individuals, and congenital infection can lead to severe birth defects. Under normal conditions an asymptomatic, chronic infection is established that is associated with the presence of quiescent cysts located in tissues of the skeletal muscle and CNS. Nevertheless, the ability to survive toxoplasmosis requires an ongoing vigorous type 1 cytokine response and strong cell-mediated immunity (2). The cytokine IL-12 is crucial in initiating these responses to the parasite (3, 4, 5).
Control of toxoplasmosis is thought to involve activation of macrophages or macrophage-related cells by parasites and the inflammatory cytokines TNF-α and IFN-γ. These activated cells display potent microbicidal effector functions such as production of NO (6, 7, 8, 9). Thus, the inability to mount a strong type 1 cytokine response, either through anti-cytokine mAb administration or cytokine gene disruption, results in death associated with high parasitemia and dissemination throughout host tissues (10, 11, 12, 13, 14, 15, 16, 17). Nevertheless, in certain experimental situations such as in IL-10 KO and d-galactosamine-sensitized mice, Tg appears to induce a lethal inflammatory cytokine response (18, 19, 20). Therefore, a successful host response to the parasite requires induction of inflammatory cytokines, but the latter must be tightly regulated to prevent host pathology associated with overproduction of these mediators.
Type 1 and type 2 cytokines are well known for their cross-regulatory properties (21). For example, IL-4 and IL-10 antagonize IFN-γ-induced macrophage activation, and IFN-γ itself displays anti-proliferative activity toward Th2 cells (22). However, less well understood is how these opposing responses interfere with each other during the course of normal infections when the host simultaneously harbors type 1- and type 2-inducing pathogens. This is surprising because the worldwide prevalence of protozoan and helminth infections in humans would predict a vast number of coinfected individuals.
We have recently become interested in using murine models of Sm and Tg infection to address this important issue. In particular, we wished to determine how the murine host responds to orally administered Tg when the animals are already undergoing a type 2-dominated gut response to chronic Sm infection. Two opposite outcomes could be envisioned based upon previous studies. First, in an Sm-induced Th2 environment, Toxoplasma could fail to elicit protective Th1 effector cells, resulting in increased tachyzoite growth and host susceptibility. Second, an ongoing Sm egg-driven type 2 cytokine response in the intestinal mucosa could prevent Tg-induced inflammatory gut pathology which occurs in mice such as those of the C57BL/6 strain (23). This hypothesis predicts that Sm infection would confer resistance to the detrimental effects of Tg in the gut.
The results presented here provide evidence that underlying Sm infection does indeed minimize Tg-induced inflammatory gut pathology and prevent the appearance of serum IFN-γ and NO. Nevertheless, TNF-α levels were higher in the double-infected mice, and the animals rapidly succumbed, showing signs of catastrophic liver damage. Interestingly, the pathology of double-infected mice was not associated with greater numbers of tachyzoites in the liver. Additionally, in Sm-infected animals administered a low Tg dose, establishment of cysts in the brain was equivalent to mice receiving Tg alone. These results indicate that increased tachyzoite burden per se is not responsible for increased susceptibility of the animals, and they suggest that a secondary factor, possibly TNF-α, mediates lethality.
Materials and Methods
Female C57BL/6 strain and outbred Swiss-Webster mice (6–8 wk of age) were obtained from Taconic Farms (Germantown, NY) and The Jackson Laboratory (Bar Harbor, ME), respectively. Animals were housed under specific pathogen-free conditions in the College of Veterinary Medicine animal facility at Cornell University.
Parasites and infections
The ME49 Tg strain was maintained by i.p. inoculation of Swiss-Webster mice with brain homegenate from mice that had been infected 6–8 wk earlier. Before injection, brain suspensions were adjusted to give a concentration of 20 cysts/mouse. For experimental infections, 250 μl of cyst-containing brain homogenate from infected mice was administered by gavage to ether-anesthetized mice.
Cercariae of Sm (NMRI strain) were maintained in Biomphalaria glabrata snails (kindly supplied by Dr. F. Lewis; Biomedical Research Institute, Rockville, MD). Cercariae were recovered by shedding, then were counted and immediately used for infection. For infection, C57BL/6 animals were ether-anesthetized, their abdomens shaved, and subsequently exposed to 70 cercariae.
To measure plasma cytokines, blood was collected in EDTA-containing tubes from the tail vein or by cardiac puncture, centrifuged (12,000 × g, 10 min at 4°C), and the resulting plasma was stored at −70°C until day of assay.
IFN-γ was measured by a two-site ELISA using plate-bound mAb HB170 (anti-IFN-γ), a rabbit polyclonal anti-mouse IFN-γ, and peroxidase-conjugated donkey anti-rabbit Ig (Jackson ImmunoResearch Laboratories, West Grove, PA) as described in detail elsewhere (14). Absorbances (405 nm) were measured on a Microplate Bio Kinetics Reader (Bio-Tek Instruments, Winooski, VT) and compared with known amounts of recombinant IFN-γ standard (Genzyme, Cambridge, MA). TNF-α levels were measured using a mouse-specific TNF-α ELISA kit according to the manufacturer’s instructions (Genzyme). The sensitivities of detection in the ELISAs were 10 pg/ml (IL-12) and 15 pg/ml (TNF-α).
A modified Griess reaction was employed to measure serum NO levels as described elsewhere (20). Briefly, blood was collected, centrifuged at 12,800 × g for 5 min, and 100 μl of the resulting plasma was added to a suspension of E. coli in 1 M HEPES (Sigma, St. Louis, MO) with 3 M formate (Sigma). Bacteria were prepared in a nitrogen-rich environment to induce high levels of nitrate reductase activity, then resuspended in PBS and stored at −70°C. The bacteria-plasma suspension was incubated (1 h, 37°C), centrifuged to pellet bacteria (3 min, 12,800 × g), and the resulting supernatant was transferred to a 96-well plate with 100 μl of a 1:1 mixture of sulfanilamide (1% in 2.5% H3PO4) and napthlethylenediamine dihydrochloride (0.1% in 2.5% H3PO4). The absorbance was measured at 600 nm and compared with standard concentrations of NaNO2.
Plasma transaminase assay
To measure the liver-associated enzyme aspartate transaminase (AST) in plasma, a protocol modified from a commercial kit (Sigma) was employed (20). Briefly, 20 μl of plasma was added to 100 μl of 0.2 M dl-aspartate and 1.8 mM α-ketoglutaric acid in PBS, tubes were mixed and incubated (37°C, 1 h), then 100 μl 2,4-dinitrophenylhydrazine (DNP) was added, and the mixture incubated a further 20 min at room temperature. The reaction was halted with the addition of 1 ml 0.4 N NaOH, and sample absorbances were measured at 490 nm after 5 min.
Total plasma IgE levels were measured using a two-site ELISA as previously described (1). Briefly, plates coated with an anti-IgE mAb (EM-95; 50 ng/well) were incubated with plasma samples and a standard curve generated with known concentrations of purified mouse IgE mAb. Bound serum IgE was detected with a FITC-conjugated anti-mouse IgE mAb (PharMingen, San Diego, CA) followed by a HRP-conjugated sheep anti-FITC mAb (Amersham, Arlington Heights, IL) and 2,2-azinodi-(3-ethylbenzthiazoline-6-sulfonic) (ABTS) substrate. Absorbances were measured at 405 nm.
Livers from mice were collected and homogenized in RNA STAT 60 (Teltest, Friendswood, TX). RNA was isolated as described (8) and reverse transcribed using 3′-specific primers in a modification of a previously published protocol (24). Briefly, 7 μg RNA was incubated at 65°C for 5 min and then chilled on ice. The following were subsequently added to the sample: 5 μl 5× RT buffer (Life Technologies, Grand Island, NY), 2.5 μl dNTP (2.5 mM; Sigma), 1 μl RNasin (Promega, Madison, WI), 0.5 μl Superscript II (Life Technologies) and 1 μl 3′ primers (100 μM; hypoxanthine phosphoribosyltransferase (HPRT) and SAG-2 primer sequences from Ref. 8). After 5 min incubation (at room temperature), the reaction mix was incubated at 42°C for 60 min followed by 52°C for 30 min on an automated thermocycler (MJ Research, Cambridge, MA). Amplification of HPRT and SAG-2 cDNA was conducted as described (8). The PCR products were resolved by electrophoresis in a 2% agarose gel, and DNA bands were visualized by staining with ethidium bromide. A 100-bp ladder was run with samples to confirm that PCR products were of the predicted length.
Livers and gastrointestinal tracts were collected from mice following CO2 asphyxiation and fixed in 10% (w/v) buffered formaldehyde. Tissues were embedded in paraffin wax, cut into 6-μm sections, and stained with hematoxylin and eosin according to standard procedures. Liver sections were screened for lipid accumulation by staining formalin-fixed frozen sections with Oil-Red-O. Tissue sectioning and staining was performed by the Cornell University Department of Biomedical Sciences Histology Laboratory.
To detect the presence of apoptotic cells in the liver, tissue sections were subjected to TUNEL employing FITC-conjugated dUTP to visualize positive-staining cells. Reagents for this assay were obtained as a kit (Boehringer Mannheim, Indianapolis, IN) and used according to instructions supplied.
Granuloma volume measurement
The volume of liver granulomas was digitally measured (Scion Image, version 1.60, Sony progressive 3CCD color video camera, ×0.45 projection lens; Nikon Optophot microscope, ×10). Only granulomas with a single, centrally located egg were chosen for analysis. The diameter of each granuloma was measured six times at 45° angles and the average calculated. Diameter was then used to calculate area. Between 15 and 25 granulomas per mouse were measured.
Significant differences in granuloma volumes, AST, NO, and cytokine levels were determined by Student’s t test or ANOVA. When required, nonparametric analysis was conducted using the Wilcoxon signed rank test. Probability values ≤0.05 were considered significant.
Percutaneous infection with Sm cercariae results in onset of egg deposition ∼35 days later, which is associated with an egg-induced type 2 cytokine response. Tg infection results in an acute infection 7–10 days later, followed by establishment of chronic disease around 1 mo after infection. Control of both stages of murine toxoplasmosis is associated with a strong type 1 cytokine response. As shown in Fig. 1, our experimental set-up was to infect mice with Sm, then 7 wk later initiate Tg infection by oral administration of cysts of the low virulence ME49 strain.
Animals infected with both Sm and Tg display increased morbidity and mortality
When C57BL/6 mice were orally infected with 200 ME49 cysts, 60% succumbed during acute infection (Fig. 2,A). Death in this group was presumably due to the lethal inflammatory gut pathology previously reported in this mouse strain (23). When C57BL/6 mice carrying a chronic Sm infection were infected with the same dose of ME49, the animals lost weight and all died within 10 days, correlating with the peak of acute toxoplasmosis (Fig. 2, A and B). At a 10-fold lower cyst dose, all Sm or Tg single-infected animals survived, whereas 34% of the double-infected animals succumbed (Fig. 2,C). Both control and Sm-infected mice displayed severe weight loss during the peak of acute Tg infection (day 10 post-Tg infection), but double-infected animals continued to lose weight at later time points, whereas mice with Tg alone regained a significant amount of their original weight (Fig. 2 D). We conclude that superimposition of Tg infection on mice with chronic schistosomiasis mansoni results in increased morbidity and mortality.
Levels of serum IgE are elevated in Sm and Sm + Tg-infected animals
To ask whether Tg infection modulated ongoing Sm-induced cytokine responses, we used serum IgE as a signature Ab isotype for measuring the relative level of the underlying type 2 cytokine response. As shown in Fig. 3, mice with Sm alone displayed elevated serum IgE at day 9 post-Tg infection. Consistent with the minimal level of IL-4 normally induced by Tg (14, 25), infection with this parasite failed to elicit an IgE response. The serum IgE levels in Sm-infected animals were not affected by subsequent oral Tg infection.
Imbalances in levels of systemic inflammatory mediators in double-infected mice
We next asked whether appearance of inflammatory mediators in the serum, which normally accompanies acute toxoplasmosis, was altered by the presence of an ongoing Sm infection. Fig. 4,A shows that serum IFN-γ levels are highly elevated in mice undergoing acute Tg infection, whereas in animals with Sm alone, serum IFN-γ remained close to background levels obtained from noninfected mice. When Sm-infected animals were subsequently challenged with Tg, production of Tg-induced IFN-γ was curtailed. This same pattern was repeated for serum NO levels (Fig. 4 B). Thus, pre-infection with Sm prevented the dramatic increase in both IFN-γ and NO, which is normally associated with acute Tg infection.
The above results are consistent with an inability to mount an appropriate type 1 response to Tg in Sm-infected animals. Nevertheless, this pattern was broken when we examined serum TNF-α levels (Fig. 4 C). While infection with Tg, but not Sm, induced serum TNF-α, the two infections in combination had a profoundly synergistic effect on levels of this inflammatory cytokine. Thus, TNF-α levels were ∼4-fold higher in Tg + Sm animals relative to the Tg alone group.
Double-infected mice display improved gut pathology
At day 9 following oral Tg infection, immediately before onset of death, intestines were removed and processed for histopathological examination. For mice infected with Sm alone (Fig. 5,A), granulomas were distributed throughout the small intestine, characterized by a heterogenous accumulation of lymphoid cells surrounding the parasite egg. Basic gut architecture, in particular the villus-to-crypt ratio, was essentially normal in these animals. This was not the case for Tg-infected mice (Fig. 5 B). These animals displayed loss of crypt structure and severe blunting of the villi, a pathology which has been reported to be associated with endogenous IFN-γ and is dependent upon the presence of CD4+ T cells (23).
When the two parasitic infections were combined (Fig. 5,C), we found evidence that the intestinal pathological changes normally associated with Tg infection were much less severe, as we had predicted might be the case. Although some blunting of the villi was evident, the basic villus and crypt structure was preserved, unlike the case of animals infected with Tg alone. These data provide evidence that patent Sm infection does indeed alleviate gut pathology associated with Tg. Nevertheless, as is clear from Fig. 2, double-infected animals display a higher rate of morbidity and mortality relative to single-infected animals, indicating that the partial alleviation of damage to the small intestine is not sufficient to prevent death. Because the liver is a major target of Sm egg-induced pathology, and because Tg has severe pathological effects in the liver after d-galactosamine sensitization (20), we turned our attention to this organ.
Sm and Tg synergize to promote severe liver damage
We first observed that mice with Sm infection displayed massive hepatomegaly and multiple egg granulomas, visible macroscopically as a white speckling covering the entire organ. In livers from animals with acute toxoplasmosis, hepatic discoloration was apparent, with little or no enlargement of the organ. Livers from double-infected animals displayed severe discoloration, and a reversal of the hepatomegaly normally induced by Sm. At this level, granulomas were also no longer visible (data not shown).
Histopathological examination of the livers of Sm-infected mice revealed large multifocal eosinophilic granulomas, but in regions removed from granulomas, liver structure appeared relatively normal (Fig. 6,A). Livers from Tg-infected animals displayed occasional foci of inflammatory cells, although the essential structure of this organ was preserved (Fig. 6,B). However, in Tg + Sm-infected mice, large regions of coagulative necrosis and extensive hepatocyte vacuolization were apparent (Fig. 6,C). The extensive vacuolization was suggestive of fatty change in the liver, indicating an accumulation of lipid as a result of disrupted metabolism. We confirmed that this was the case, because cross-sections of livers from double-infected mice stained strongly with the lipid-sensitive dye, Oil-Red-O, whereas livers from Sm or Tg single-infected animals displayed very little of this staining (Fig. 7, A–C).
Further evidence for hepatocyte death is shown in Fig. 8. Thus, livers from coinfected mice possessed large TUNEL-positive regions (Fig. 8,A). This staining, which is virtually identical with that obtained when normal tissue sections were pre-incubated with DNase I before staining (Fig. 8 B), suggests that heptatocytes within this area have recently undergone programmed cell death. In situ TUNEL staining of livers from single infections of Tg and Sm failed to reveal similar apoptotic regions (data not shown).
Biochemical evidence for severe liver dysfunction is shown in Fig. 9. In mice infected with either Sm or Tg alone, the liver-associated enzyme AST is present in the serum at 2–3 fold greater levels than in noninfected animals. This result indicates that hepatic damage normally accompanies both Tg and Sm infections, and results in leakage into the serum of an enzyme normally confined to the liver. Nevertheless, in combination, the serum AST level was approximately seven times higher than background levels. Thus, Sm and Tg together exert severe pathological effects on the liver.
Finally, as shown in Fig. 6, A and C, Sm egg granuloma size was smaller in double-infected mice, as predicted based on the absence of visible gross lesions in the livers of these animals (data not shown). We measured granuloma area in several cross-sections and found that the lesions were ∼55% smaller in double-infected mice (Fig. 10 A). Despite the smaller size of the granulomas, general cellular composition appeared approximately equivalent. In both cases a broad mix of granulocytes, lymphocytes, and macrophages were present in the lesions.
Double-infected mice do not harbor increased tachyzoites relative to animals infected with Tg alone
The decreased systemic IFN-γ in double-infected animals might suggest a generalized inability to control tachyzoites, but we failed to find any evidence that this is the case. As shown in Fig. 6,C, double infection precipitates major pathological changes in the liver. Nevertheless, immunohistochemical staining does not show significant numbers of tachyzoites in either Tg or Sm + Tg groups (data not shown). Nevertheless, as shown in Fig. 11, RT-PCR amplification of transcripts for the Tg tachyzoite surface protein SAG-2 (p22) (26) revealed the presence of Toxoplasma transcripts in both Tg alone and Tg + Sm-infected animals. Importantly, this analysis indicated that there was no evidence that SAG-2 transcript levels were higher in double-infected mice, despite the fact that the latter display increased morbidity and mortality. Finally, Sm-infected and control animals were infected with a low ME49 dose, and 30 days later brains removed and cysts enumerated. As shown in Fig. 10 B, the number of cysts contained within brain tissue of Sm and control mice was not significantly different.
It has been recognized for several years that Sm and Tg induce polar opposite immune responses. Characteristically for a helminth, the dominant Th response during schistosomiasis is Th2-like, whereas for the intracellular parasite Tg, the host responds with a vigorous type 1 cytokine response (27). In both cases, the dominant response is ultimately host protective. Thus, IL-4−/− mice cannot survive Sm infection (1), and IFN-γ as well as IL-12 KO animals die when infected with Tg (10, 11, 28). This dichotomy in host response pattern suggests that mice infected with schistosomes might display high susceptibility to subsequent Tg infection as a result of Sm-induced Th2 cytokines, because the latter mediators possess down-regulatory activity on development of Th1 responses (21). A report several years ago (predating discovery of Th1 and Th2 cells) indicated double-infected mice do, indeed, display increased susceptibility (29). In addition, several more recent reports indicate that Sm infection is capable of affecting development of type 1 responses to unrelated Ag (30, 31, 32, 33).
We decided to re-examine the issue of Sm-Tg coinfection in light of recent data suggesting that C57BL/6 mice orally infected with Tg cysts die from intestinal inflammation associated with high levels of IFN-γ (23, 34). Based upon this model, Sm infection might be predicted to confer resistance to Tg, because type 2 responses associated with eggs passing across the intestinal wall would mitigate IFN-γ-mediated pathology. The latter hypothesis is supported in part by our data showing that intestinal pathology in double-infected C57BL/6 animals appears less severe than in those animals infected with either parasite alone (Fig. 5). Nevertheless, oral inoculation of Sm-infected mice with Tg resulted in severe morbidity and high mortality. Death in these animals was accompanied by catastrophic liver necrosis, suggesting that the primary site of Tg-induced pathology had shifted from the intestine to the liver.
In many situations, overt type 1 cytokine responses are associated with serious intestinal inflammation (reviewed in Ref. 35). We have hypothesized elsewhere that this places evolutionary pressure on the host to respond to chronic intestinal damage, such as occurs during Sm infection, by mounting a type 2 response consisting of production of the anti-inflammatory cytokines IL-10 and IL-4 (1). Such an environment would tend to prevent the proinflammatory response induced by Toxoplasma, thereby reducing local inflammation, but this cytokine milieu might also be expected to increase local tachyzoite proliferation and promote parasite dissemination. Nevertheless, we failed to find evidence that the Tg burden in coinfected mice was higher, as measured by cysts establishing in the brains of animals given sublethal Tg doses.
To specifically determine whether increased susceptibility after Tg infection correlated with decreased type 1 immune responses in Sm-infected mice, we examined plasma levels of IFN-γ, TNF-α, and NO. As expected, mice infected with ME49 alone displayed elevated levels of all three of these proinflammatory mediators. In contrast, plasma levels of IFN-γ and NO in double-infected animals were no greater than in mice infected with Sm alone. We do not as yet know the precise mechanism underlying defective production of IFN-γ and NO, but the most likely explanation is that Sm-induced type 2 cytokines limit development and expression of the type 1 response normally associated with Tg infection. Although plasma levels of type 1 mediators were reduced in dual-infected animals, it is unlikely that Tg-specific responses are completely absent because, as mentioned above, cyst burden in dual-infected mice were no greater than in mice infected with Tg alone.
Plasma TNF-α levels, which were high in mice infected with ME49 alone, were elevated even further in double-infected animals. The presence of high levels of TNF-α in the near absence of IFN-γ and NO was unexpected. Nevertheless, it has previously been reported that schistosome infection primes animals for massive TNF-α production when subsequently exposed to the appropriate triggering molecule (36). Thus, serum from Sm-infected mice previously injected i.v. with LPS displays potent tumoricidal activity as a result of high amounts of TNF-α. Because Tg is able to induce TNF-α during the normal course of infection, it seems likely that this pathogen functions in a manner similar to LPS in inducing TNF-α overproduction in Sm-infected animals. Moreover, while IFN-γ activated macrophages are clearly a major TNF-α source, other cell types such as mast cells, Th2 lymphocytes, and B cells can also produce this cytokine (37). Because all of these cells are present in the diseased liver of a schistosome-infected mouse (38), it is possible that Tg triggers one of these cell types to synthesize TNF-α.
Mice dying from dual infection displayed severely diseased livers, with evidence of steatosis and coagulative necrosis in areas adjacent to egg granulomas. Although our data do not directly demonstrate that this severe liver disease is responsible for death in the animals, this would appear plausible given the severity of the pathology. A seemingly likely possibility accounting for the exacerbated liver condition of double-infected mice is that decreased NO and IFN-γ levels result in high numbers of replicating tachyzoites and associated tissue destruction in this organ. However, our results do not support this hypothesis. Thus, RT-PCR analysis of Tg SAG-2 transcripts in the liver did not provide evidence for higher tachyzoite numbers in double-infected mice. In addition, immunohistopathological analysis using Tg-specific antiserum failed to reveal elevated tachyzoite numbers in the liver of double-infected mice (data not shown).
In some situations when the liver is stressed, TNF-α possesses the capability of inducing programmed cell death in hepatocytes (39, 40), and it is possible that the high level of this cytokine is related to the liver damage we report here. We recently reported that mice sensitized with the hepatotoxin, d-galactosamine, are exquisitely sensitive to Tg-triggered, TNF-α-mediated liver destruction (20). Furthermore, because NO has been reported to protect against TNF-α-mediated liver pathology (41), its absence in the double-infected mice could also contribute to the severe disease observed. Indeed, nuclear condensation, which is indicative of apoptosis, is apparent within a subset of damaged hepatocytes in dual-infected mice. Furthermore, TUNEL-positive areas of liver from coinfected animals are readily demonstrable.
Related to this issue, it has recently been shown that inducible NO synthase KO (iNOS−/−) mice display increased resistance to Tg during acute infection (42). Thus, it is possible that during Tg infection, NO production, by virtue of its immunosuppressive activity, is detrimental to the host, but that in the Tg-Sm animals its absence results in failure to protect against TNF-α-mediated liver pathology. The severe cachexia afflicting Tg-infected mice also indicates a role for TNF-α in the pathology of acute toxoplasmosis, and anti-TNF-α mAb treatment of mice with lethal acute Tg infection delays time to death (20). We are currently further examining the role of this inflammatory cytokine using TNF receptor KO mice infected with both parasites.
The liver pathology in mice undergoing concurrent Tg and Sm infection is, in part, reminiscent of that occurring in schistosome-infected T cell deficient mice (43, 44, 45). In the latter animals, granulomas fail to form and the hepatocytes surrounding the eggs accumulate lipids and take on a vacuolated appearance. Liver failure and death rapidly follow. The damage to this organ is believed to be due to the action of Sm egg-derived hepatotoxins, such as the glycoprotein designated ω1 (46). In immunocompetent animals, egg hepatotoxins would be sequestered within the granuloma, and therefore liver damage is minimized. Thus, the egg granuloma is believed to be a protective host response which contributes to the continued functioning of the liver despite a large egg burden. In Sm and Tg coinfected mice, granulomas were significantly smaller as has been reported previously (47), and it is possible that liver damage in these animals was the result of inadequate sequestration of Sm egg-derived toxins. Why the granulomas should be smaller in dual-infected mice is unclear, but it is possible that local effects of a Tg-induced type 1 responses limit the extent of the largely Th2-mediated granulomatous lesions (48).
Many regions of the world are endemic for severe helminth, protozoan, bacterial, and viral infections. Nevertheless, very little is known regarding immune responses elicited by pathogens simultaneously infecting one host, and the influence of such responses on each other. Similarly, few clinical studies have examined how the immune system responds to coinfection. The studies presented here clearly demonstrate that Tg and Sm, together within one host, synergize to promote a severe pathology that could not be predicted based upon studies with either parasite alone. The grossly dysregulated cytokine production in these animals strongly suggests an immune-mediated component of the lethal pathology. Our current efforts are focused on defining the molecular and cellular basis for the pathology induced by concurrent infection with these pathogens.
We thank A. Lee (Department of Biomedical Sciences) for assistance with the granuloma volume measurement. Schistosome life stages for this work were supplied through National Institutes of Health-National Institute of Allergy and Infections Diseases Contract N01-AI-55270.
This work was supported by grants from the National Institutes of Health to E.Y.D. (AI-40540) and E.J.P. (AI-32573), and the U.S. Department of Agriculture Hatch Act and Animal Health and Disease Research Program (Cornell University, College of Veterinary Medicine). E.J.P. is a recipient of a Burroughs-Wellcome New Investigator Award. L.R.B. was supported by National Research Service Award AI-09512.
Abbreviations used in this paper: Sm, S. mansoni; Tg, T. gondii; AST, aspartate transaminase; KO, knockout; HPRT, hypoxanthine phosphoribosyltransferase.