IL-4 Plays a Crucial Role in Regulating Oxidative Damage in the Liver During Schistosomiasis¹

Over 200 million people worldwide suffer from schistosomiasis, a disease caused by infection with parasitic worms of the genus Schistosoma (1). The adult worms of Schistosoma mansoni reside in the inferior mesenteric veins where the females can lay eggs that pass through the intestinal wall to be deposited with the feces or are carried by the blood flow into the liver where they induce a vigorous granulomatous response (2). In <10% of patients this response can ultimately result in hepatic fibrosis, portal hypertension, and hepatosplenomegaly (1); this hepatosplenic form of schistosomiasis can be fatal. While the intensity of infection is an important factor influencing the development of severe liver disease, recent studies in humans and mice have clearly shown that the qualitative nature of the immune response plays a central role in determining the disease process (3, 4, 5).

In wild-type (WT)³ mice, the immune response during schistosomiasis is strongly Th2-like. Previous studies have shown that it is primarily the eggs, not the adult worms, that stimulate the Th2 response (6). The response to the eggs progresses through a Th0 stage to become Th2 dominated (7) and is characterized by high levels of IL-4, IL-5, IL-13, IL-10, and circulating IgE Ab (6, 8, 9). In the absence of IL-4, Th2 cytokine production is severely impaired, but not abolished (10, 11), and there are concurrent increases in the production of proinflammatory mediators (i.e., NO, IFN-γ, and TNF-α), morbidity, and mortality (3, 5, 10). Although granuloma formation is a Th2, CD4 T cell-dependent process (12), the absence of IL-4 does not prevent the development of these lesions (10, 11). Instead, through the use of IL-13^−/−, IL-4Rα^−/−, and Stat6^−/− mice, the contributions of both IL-4 and IL-13 have been shown to be pivotal for granuloma development (5, 9, 11). Furthermore, this Th2-induced granuloma formation has been shown to be essential for survival by preventing severe hepatocyte damage (13, 14). Finally, in human disease, the absence of a strong Th2 response and the presence of high levels of TNF-α and IFN-γ are associated with severe hepatosplenic disease (4).

The severity of the disease that develops in infected IL-4^−/− mice correlates most closely with increased production of NO. In this report, we demonstrate that in the absence of IL-4, infected mice manifest more liver damage and that this damage correlates with the overproduction of not only NO, but also O₂⁻ and ONOO⁻, and with the concurrently diminished level of catalase, an enzyme that protects against oxidative damage. IL-4^−/− mice thus have a severely impaired ability to regulate oxidative damage. Treatment of infected IL-4^−/− mice with antioxidants results in prolonged survival and decreased liver damage, further suggesting that IL-4 is playing an important role in preventing oxidative damage to the liver during S. mansoni infection.

Female IL-4^−/− C57BL/6 mice (3) were bred and used at 6–12 wk of age. Female C57BL/6 mice were purchased from Taconic Farms (Germantown, NY). For infection, mice were exposed percutaneously to approximately 35 or 70 S. mansoni cercariae (NIMR Puerto Rican strain). Egg and worm burdens were assessed as previously described (3). At autopsy, tissues from infected and uninfected mice were fixed in 10% buffered Formalin, paraffin embedded, sectioned, and stained with hematoxylin-eosin (H&E) for histological examination.

Dividing hepatocytes were assessed in Formalin-fixed, H&E-stained liver sections, and the number of hepatocytes containing mitotic bodies per 100 high-power fields (HPF) was calculated. At least 50 HPF were assessed per liver section. Reduced hepatocyte volume was indicated by assessing the number of hepatic nuclei per HPF in fixed, stained liver sections, and sinusoidal integrity was assessed by histologic examination of Formalin-fixed, H&E-stained liver sections.

Soluble egg extract (SEA) was prepared as described previously (2). LPS, PMA, bovine liver superoxide dismutase (SOD), and bovine liver catalase were purchased from Sigma (St. Louis, MO). 2,7-Dihydrodichlorofluorescein diacetate (H₂DCFDA) was purchased from Molecular Probes (Eugene, OR). FITC-, PE-, and/or CyChrome C-labeled anti-CD8, anti-CD4, anti-B220, anti-Gr-1, and anti-Mac-1 mAb were purchased from PharMingen (San Diego, CA). Biotinylated anti-F4/80 Ab was purchased from Serotec (Oxford, U.K.). Streptavidin-PE was purchased from Jackson ImmunoResearch (West Grove, PA). Plate-bound anti-CD3 mAb (PharMingen) was used at 1 μg/well.

Liver tissue was harvested directly into RNAzol (Tel-Test, Friendswood, TX) and snap frozen. Total liver mRNA was isolated, and cDNA was made using SuperScript II reverse transcriptase (Life Technologies, Gaithersburg, MD) as previously described (15). Hypoxanthine-guanine phosphoribosyl-transferase (HPRT) transcripts were amplified using competitive PCR as previously described (15) and were used to normalize cDNA levels. HPRT was amplified using 37 cycles, inducible NO synthase (iNOS) was amplified using 41 cycles, and catalase, MnSOD, and CuZnSOD were amplified using 27 cycles. Primers used for catalase, CuZnSOD, and MnSOD amplifications are as follows: catalase (forward, 5′-CCACCGGAGGCGGGAACC-3′; reverse, 5′-GCAATAGGGGTCCTCTTTCC-3′) (16), CuZnSOD (forward, 5′-GATTAACTGAAGGCCAGCATG-3′; reverse, 5′-GTCATCTTGTTTCTCATGGACC-3′) (17), and MnSOD (forward, 5′-CCCAGACCTGCCTTACGACT-3′; reverse, 5′-CGACCTTGCTCCTTATTGAA-3′) (18). Primers for HPRT and iNOS amplification were described previously (15). PCR products were run on a 2.5% agarose gel, stained using ethidium bromide, and analyzed using the Eagle Eye program (Stratagene, La Jolla, CA).

Formalin-fixed, paraffin-embedded tissue sections were deparaffinized and rehydrated, and endogenous peroxidase activity was quenched with 3% H₂O₂ in methanol. The samples were then microwaved for 10 min in citrate buffer (pH 6.0) to unmask Ab epitopes before incubation with 2 μg/ml polyclonal anti-iNOS (Transduction Laboratories, Lexington, KY), anti-nitrotyrosine (Upstate Biotechnology, Lake Placid, NY), or anti-β-galactosidase Ab (made in our laboratory). Ab binding was detected using the biotinylated goat anti-rabbit avidin-biotin peroxidase complex kit (Vector Laboratories, Burlingame, CA) and development with diaminobenzidine (Vector Laboratories) as directed by the manufacturer. Slides were counterstained with either toluene blue (Sigma) or Vector Green (Vector Laboratories).

Livers from infected mice were perfused with citrate saline, and intact granulomas were isolated by homogenization and sedimentation. Granulomas were dispersed, and single-cell suspensions were prepared as described previously (19). Granuloma cells were resuspended at 2 × 10⁶/ml in complete T cell medium containing DMEM (Life Technologies), 10% FCS (Sigma), 100 U/ml penicillin, 100 μg/ml streptomycin (Life Technologies), 10 mM HEPES (Life Technologies), l-glutamine (Life Technologies), and 5 × 10⁻⁵ M 2-ME (Sigma). Cells (4 × 10⁵/well) were cultured in 96-well flat-bottom plates (Falcon; Becton Dickinson, Franklin Lakes, NJ) at 37°C in 5% CO₂. Culture supernatants were harvested at 72 h for analysis of NO using the Greiss reaction (20). Samples of the granuloma cell preparations were cytospun onto glass slides and stained with Hema 3 (Biochemical Sciences, Swedesboro, NJ) to visually determine cell composition.

Granuloma cells were resuspended; plated at 2 × 10⁶/ml in HBSS containing 1% FCS (Sigma), 100 U/ml penicillin, 100 μg/ml streptomycin (Life Technologies), and 10 mM HEPES (Life Technologies) in 96-well flat-bottom plates (Falcon); and cultured with or without 500 ng/ml PMA for 30 min at 37°C in 5% CO₂. Nitroblue tetrazolium (NBT; Promega, Madison, WI) was added to a final concentration of 56 mM, and cells were incubated for 30 min before the addition of 10% SDS in 0.1 M HCl to solubilize the blue formazan. ODs were read at 570 nm, and O₂⁻ production was calculated as: [(OD of cells with NBT) − (OD of cells without NBT)] − [OD of wells with NBT and medium alone]. To determine O₂⁻ production by individual cell types, cells were cytospun onto slides after incubation with NBT and counterstained with Wright’s stain (Sigma).

Granuloma cells were labeled with 5 μM H₂DCFDA as previously described (21) and incubated at 37°C in 5% CO₂ for 30 min. Cells were incubated on ice with 2.5 μg/10⁶ cells Fc Block (PharMingen), stained for 20 min with FITC-, PE-, or CyChrome C-conjugated Ab against surface markers, washed twice with 1% FCS/0.08% sodium azide (NaN₃; Sigma) in PBS (Sigma), and analyzed immediately using a FACScalibur flow cytometer (Becton Dickinson) with the CellQuest program (Becton Dickinson). For cell sorting, gates were set using forward and side scatter as shown in Fig. 8, and approximately 50,000 cells/gate were collected directly into FCS using a FACScalibur flow cytometer (Becton Dickinson). Immediately after isolation, the collected cells were cytospun onto glass slides and stained with Hema 3 (Biochemical Sciences).

Livers from uninfected and infected mice were perfused with citrate saline, snap frozen, and homogenized in ice-cold lysis buffer (1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EDTA, 1 mM EGTA, 1 mM N-ethylmaleimide, 0.1 μM pepstatin, 1 mM PMSF, and 0.1 M N-tosyl-l-phenylalanine chloromethyl ketone in 150 mM NaCl/50 mM Tris-HCl, pH 7.4) and centrifuged at 14,000 rpm for 10 min at 4°C (22). Supernatants were removed, and protein concentrations were determined using the bicinchoninic acid method (Pierce, Rockford, IL). SOD and catalase activities were assayed as previously described (23, 24), and specific activity was calibrated from a standard curve generated with bovine liver SOD or catalase (Sigma).

Daily, mice were injected i.p. with 100 μl of saline alone or containing 2 mg of uric acid in suspension (Sigma) (25). Clinical scores were assigned and assessed by a veterinarian as follows: 1 = hunched posture, piloerection, shivering, periorbital edema, or lethargy; 2 = any two symptoms; 3 = any three symptoms; and 4 = any four symptoms or death.

Data were analyzed using Student’s t test or two-way ANOVA as indicated.

Previous studies showed that in the absence of IL-4, S. mansoni-infected mice die from a proinflammatory syndrome characterized by severe cachexia (3). Disease first becomes apparent shortly after the onset of egg production by adult worms and is not due to differences in infective burden, as both WT and IL-4^−/− mice have similar worm and total egg burdens following exposure to a similar infectious dose of cercariae (3) (data not shown). In contrast to the similar total egg burdens in the liver, the levels of eggs per gram of liver differ significantly (Fig. 1,a), indicating that in WT mice, but not IL-4^−/− mice, the liver responds to egg accumulation by enlarging. Liver enlargement is a normal response of WT mice to schistosome infection and can be associated with an increased number of dividing hepatocytes (Fig. 1,b). In contrast, the livers of infected IL-4^−/− mice are visibly smaller than those of infected WT animals (data not shown) and contain significantly fewer mitotic hepatocytes (Fig. 1,b). Although uninfected mice had few detectable dividing hepatocytes, hepatocyte proliferation increases after infection with S. mansoni and peaks at 45 days after infection (Fig. 1,b). The hepatocytes from infected IL-4^−/− animals also begin to proliferate after infection, and proliferation peaks at the same time as in WT mice, but the total number of dividing hepatocytes is greatly reduced (Fig. 1 b). Additionally, at this time in the infected IL-4^−/− mice, the gross pathology of the liver begins to worsen as reported previously (5) and as indicated histologically by a reduction in sinusoidal integrity and hepatocyte volume (data not shown). Together these results suggest that during infection the absence of IL-4 results in impaired hepatocyte proliferation and therefore impaired liver regeneration, which ultimately culminates in exacerbated liver pathology.

The best immunologic correlate of disease severity in infected IL-4^−/− mice is NO production. When produced at high levels NO has also been shown to be associated with severe morbidity in other systems (e.g., septic shock, reperfusion injury, and hemorrhagic shock) (26, 27). One major effect of elevated NO levels is the inhibition of cell proliferation, most evident in lymphocyte populations (28), and this impairment is particularly evident in schistosome-infected IL-4^−/− mice (E. A. Patton, A. C. La Flamme, and E. J. Pearce, unpublished observations). Consequently, we investigated whether the production of NO in the liver played a role in the observed impairment in hepatocyte proliferation. Focusing on iNOS, the enzyme responsible for high-level NO production during immune responses (29), we found that iNOS liver transcript levels increase during infection in WT mice as has been reported in previous work (30), but, more importantly, that the increase seen in the livers of infected IL-4^−/− animals is substantially higher at later time points (45–50 days after infection; Fig. 2) when the failure of hepatocytes to proliferate is most apparent (Fig. 1 b). The level of iNOS protein in the liver tissue was also examined using immunohistochemistry. As previously reported (30), there were more iNOS-positive cells in the IL-4^−/− mice than in the WT animals. Interestingly, sections of the liver from infected IL-4^−/− mice contained iNOS-positive hepatocytes in addition to the iNOS-positive cells within the granulomatous lesion (data not shown). Inducible NOS-positive hepatocytes were not present in the infected WT mice.

To measure NO production specifically within granulomas during infection, granuloma cells were isolated and cultured in vitro in the presence of various stimuli. The differential counts indicate that granulomas from WT mice contain more eosinophils and fewer macrophages and lymphocytes than lesions from IL-4^−/− mice (Fig. 3,a). After stimulation with anti-CD3, SEA, or LPS, granuloma cells from IL-4^−/− mice produced high levels of NO, while, in contrast, the levels of NO in the WT cultures were significantly lower (Fig. 3 b). In other experiments detectable levels of NO were found in the supernatants from WT granuloma cultures, but these levels were always significantly below those found in the IL-4^−/− granuloma cultures. NO production was highest after anti-CD3 stimulation, suggesting a role for T cells in the induction of NO in the granulomas. Taken together, these results indicate that NO production is increased in the liver during schistosome infection in the absence of IL-4 and therefore may play a role in impaired hepatocyte proliferation and more severe liver disease in infected IL-4^−/− mice.

Since IL-4 is known to suppress the production of other reactive nitrogen and oxygen species in addition to NO (32, 33, 34, 35, 36), and these intermediates may also be involved in mediating liver pathology (37, 38, 39), the generation of O₂⁻, hydrogen peroxide, and ONOO⁻ by granuloma cells was investigated. To examine O₂⁻ production, granuloma cells were stimulated with PMA, and O₂⁻ was detected by measuring the formation of blue formazan after addition of NBT. Granuloma cells from IL-4^−/− mice produced significantly more O₂⁻ than did those from WT mice (Fig. 3,c). Neither IL-4^−/− nor WT granuloma cells produced detectable levels of O₂⁻ when unstimulated (data not shown). The addition of SOD inhibited the formation of formazan, verifying that the assay specifically measures the generation of O₂⁻ (Fig. 3 c). The cell types producing O₂⁻ were visually identified by cytospinning PMA-stimulated cells after incubation with NBT. This approach indicated that macrophages, lymphocytes, and granulocytes were all capable of contributing to O₂⁻ production in the granulomas of WT and IL-4^−/− mice (data not shown).

Under conditions of increased NO and O₂⁻ production, as seen in the livers of infected IL-4^−/− mice, NO can compete with SOD for O₂⁻, resulting in the formation of ONOO⁻ (37, 40). This RNI is more reactive than either NO or O₂⁻ (40) and can cause lipid peroxidation, DNA damage, and nitrosylation of tyrosine residues (40, 41, 42, 43). The nitrosylation of tyrosine residues is a hallmark of ONOO⁻ production and can be detected in tissues using Ab specific for nitrotyrosine (37, 44). Although some nitrotyrosine-positive cells can be found in the granulomas of WT mice, a greater number of positive cells are present in the granulomas of IL-4^−/− mice (Fig. 4, b and e, short arrows). As with iNOS, some hepatocytes in the livers of IL-4^−/− mice also contain nitrosylated tyrosine residues and thus may be damaged by the overproduction of reactive species (Fig. 4,e, long arrows). These nitrotyrosine-positive hepatocytes were found primarily in mice that had severe liver pathology as determined by histological analysis of the liver and weight loss at time of euthanasia. No positive cells were detected in uninfected mice (Fig. 4, a and d) or after staining with an irrelevant Ab (Fig. 4, c and f). Therefore, ONOO⁻ is generated in the livers of schistosome-infected mice, and levels are greatly elevated in the absence of IL-4.

The production of ROI/RNI by specific cell types was determined using flow cytometric analysis. H₂DCFDA is an intracellular dye that becomes fluorescent after being hydrolyzed by reactive species. This dye reacts with NO, hydrogen peroxide, and most strongly with ONOO⁻, but does not directly react with O₂⁻ (21, 45). Using parameters of size, granularity, surface marker expression, and H₂DCFDA fluorescence, the cells producing reactive species were identified. The distribution of cell types in the granulomas from WT and IL-4^−/− mice was found to differ significantly (Fig. 5). Granulomas from IL-4^−/− mice contain a higher percentage of macrophages and lymphocytes, whereas WT granulomas contain a greater percentage of eosinophils. ROI/RNI are generated principally by macrophages and lymphocytes, although granulocytes can produce ROI/RNI in both WT and IL-4^−/− mice (Fig. 5). These results indicate that the increased production of ROI/RNI in the granulomas from IL-4^−/− mice is primarily due to the increased number of macrophages and lymphocytes compared with granulomas from WT mice and is not due to increased production by a specific cell type.

When the production of reactive species by subsets of lymphocytes within the lymphocyte gate (R2) was analyzed, it was found that there was a similar, if not greater, level of ROI/RNI-positive CD4 T cells in WT vs IL-4^−/− granulomas (Fig. 6,a). Only low levels of ROI/RNI-positive CD8 T cells were found in WT and IL-4^−/− granuloma cell preparations (0.47 vs 0.29% of the total cells; WT and IL-4^−/−, respectively), and similar populations of Mac-1-positive, non-B, non-T lymphocytes were ROI/RNI-positive in granuloma cell preparations from WT and IL-4^−/− mice (0.87 vs 1.23% of total cells; WT and IL-4^−/−, respectively). Because the remaining ROI/RNI-positive population within the R2 gate was CD4⁻, CD8⁻, and Mac-1⁻ (0.9 vs 6.4% of total cells; WT and IL-4^−/−, respectively), B cells were implicated as the major ROI/RNI-positive cell type within the lymphocyte gate from IL-4^−/− granuloma preparations. Additional experiments supported this conclusion by showing that B cells from the granulomas of WT and IL-4^−/− mice were ROI/RNI positive (Fig. 6 b). The difference between the WT and IL-4^−/− mice appears to be due to the greater number of B cells found in the granulomas from IL-4^−/− animals and not to a difference in the ability of the B cells to generate ROI/RNI (B220⁺ ROI/RNI⁺ = 1.8 vs 12.2% of total cells; WT and IL-4^−/−, respectively). Taken together, these results indicate that NO, O₂⁻, and ONOO⁻ are overproduced in the livers of infected IL-4^−/− mice, and this overproduction may be due to the increased number of macrophages and B lymphocytes that infiltrate the granulomas in the absence of IL-4.

Although ROI/RNI production in the liver during infection is clearly increased, a concomitant increase in the production of antioxidant enzymes (e.g., catalase and SOD) could occur to minimize subsequent damage. To examine this possibility, we assayed the activity and production of SOD and catalase in the livers of infected and uninfected WT and IL-4^−/− mice. After an initial increase in the activity and transcription of both enzymes (data not shown), there was a decrease in both at later time points during infection (Fig. 7 and data not shown). Although the levels of SOD were not significantly different in WT compared with IL-4^−/− livers (data not shown), the levels of catalase activity (Fig. 7) and mRNA (data not shown) were significantly decreased in the IL-4^−/− mice. We predict that since SOD production is not elevated concurrently with elevations in NO and O₂⁻ in IL-4^−/− mice, the production of ONOO⁻ will be promoted. The decrease in catalase suggests that an excess of hydrogen peroxide could be available to form more ROI, such as the hydroxyl radical or hypochlorous acid. These results indicate that an increased production of protective enzymes does not occur to compensate for the overproduction of NO, O₂⁻, and ONOO⁻ in the infected IL-4^−/− mice, and therefore these species may be directly contributing to the exacerbated pathology seen in these animals.

Treatment with anti-oxidants has been shown to prevent damage due to the overproduction of ROI/RNI in several disease models (25, 39, 46). Uric acid is a direct scavenger of ONOO⁻ and other peroxy radicals (47) and can prevent severe morbidity due to ONOO⁻ formation in the mouse experimental allergic encephalomyelitis model (25). To determine whether the overproduction of ONOO⁻ directly contributes to the severe morbidity and mortality observed in infected IL-4^−/− mice, treatment with uric acid was initiated. Since uric acid is quickly metabolized in mice (the half life in vivo is 2 h) (25), infected IL-4^−/− animals were each injected daily with 2 mg of uric acid, and the course of disease progression was followed. IL-4^−/− mice infected with a high dose of cercariae (70 cercariae) and treated with uric acid had significantly prolonged survival compared with IL-4^−/− mice injected with saline alone, but ultimately all mice succumbed to the infection (data not shown). However, infection with a lower dose of cercariae (35 cercariae) resulted in the rescue of a significant percentage of the infected IL-4^−/− mice (Fig. 8,a). Immediately after the beginning of treatment, both uninfected and infected IL-4^−/− mice treated with uric acid lost weight; nevertheless, these mice were active and healthy. Because of the nonspecific weight loss associated with uric acid treatment (data not shown), morbidity was assessed by visual inspection of the mice and assignment of clinical scores rather than by weighing the animals. Following an initial increase in morbidity, infected mice treated with uric acid were active and healthy compared with the infected mice treated with saline alone, which were lethargic, shivering, hunched, and had piloerection and periorbital edema (Fig. 8,b). Histological examination of the livers of uric acid- and saline-treated mice indicated that uric acid treatment resulted in increased hepatocyte proliferation (Fig. 8 c). Treatment with supranutritional doses of vitamins E and C, both well-characterized antioxidants (37), also prolonged survival and decreased weight loss in IL-4^−/− mice infected with a high dose of S. mansoni (data not shown), although they were less effective than uric acid in this regard. Therefore, a reduction in the presence of oxidative species in infected IL-4^−/− mice decreased morbidity, enhanced survival, and increased hepatocyte proliferation.

To determine whether uric acid treatment directly affected the downstream effects of ONOO⁻ production, the level of nitrotyrosine in the liver was compared between saline- and uric acid-treated mice. Mice treated with uric acid had fewer nitrotyrosine-positive cells in the liver than saline-treated mice (Fig. 8, e vs d). These results suggest that uric acid treatment was able to directly scavenge ONOO⁻ in vivo. The ability of uric acid treatment to protect infected IL-4^−/− mice from death and severe liver damage demonstrates a detrimental role for ROI/RNI in fatal schistosomiasis.

Our studies have centered on understanding the mechanism by which IL-4 can protect against severe liver damage and death in schistosome-infected mice. Using IL-4^−/− mice, we found that hepatocyte proliferation, structure, and sinusoidal integrity were compromised in the absence of IL-4, and that this damage to the liver correlated with increased production of NO and O₂⁻ by cells within the granuloma and, in the case of NO, by hepatocytes. Overproduction of these reactive species in the livers of IL-4^−/− mice also correlated to increased levels of nitrotyrosine, which is a marker of ONOO⁻ formation (40, 44). Although all cell types within the granuloma (macrophages, lymphocytes, and granulocytes) were found to produce ROI/RNI, a greater percentage of lymphocytes and macrophages than eosinophils produced ROI/RNI in general. Consequently, the primary difference between the WT and IL-4^−/− mice lay in the fact that the granulomas from infected IL-4^−/− mice contained more ROI/RNI-producing lymphocytes and macrophages than did those from WT mice. Despite the increase in ROI/RNI in the liver, enzymes responsible for protection from oxidative damage (SOD and catalase) were not up-regulated in either WT or IL-4^−/− mice. More importantly, the level of catalase in the liver was significantly reduced in the absence of IL-4, indicating that these mice are more susceptible to oxidative damage. Finally, treatments with antioxidants were effective in reducing morbidity and mortality in IL-4^−/− mice infected with S. mansoni, suggesting that the generation of ROI/RNI contributes to severe liver damage and death in IL-4^−/− mice.

Previous studies of infected WT mice have shown that granulomas and, moreover, macrophages isolated from them produce O₂⁻ when cultured in vitro (19, 48, 49, 50). Furthermore, production of O₂⁻ by macrophages in the granuloma was shown to parallel IFN-γ production, suggesting that IFN-γ, which has been found to be up-regulated in the livers of mice during acute infection (30, 51), may play a role in activating the macrophages to produce O₂⁻ in vivo. In addition, eosinophils purified from granulomas were reported to produce both O₂⁻ and hydroxyl radical after in vitro stimulation with PMA (19). Our data support a role for macrophages and eosinophils in the production of O₂⁻, but in addition indicate that B lymphocytes can contribute to this process as well.

This is the first report on the role of IL-4 in regulating O₂⁻ production in the livers of schistosome-infected mice. The ability of IL-4 to regulate O₂⁻ production during infection may occur through several independent mechanisms. First, IL-4 may exert a direct effect on O₂⁻ generation in the granuloma; down-regulation of O₂⁻ production in macrophages and PMN by IL-4 is well documented (33, 34, 35). Secondly, IL-4 may reduce O₂⁻ production by promoting the development of more Th2 cells that produce IL-10. Recent studies indicate that IL-10 is more effective than IL-4 or IL-13 in down-regulating O₂⁻ production by macrophages, PMN, or Kupffer cells (33, 34, 36, 52), and IL-10 production is clearly reduced in infected IL-4^−/− mice (3). Finally, IL-4 may play a role in regulating the cellular composition of the granuloma. Indeed, in agreement with previous histological studies (11), we found that the cellular composition differed significantly between WT and IL-4^−/− animals and, moreover, that the increase in reactive species appeared to be due primarily to changes in the cellular composition rather than to increased production of reactive species by any particular cell type. These results suggest that it may be the difference in cell recruitment, possibly through regulation of chemokine expression, rather than regulation of ROI production on a per cell basis that defines the role of IL-4 in controlling ROI production in the granuloma, although a direct effect of IL-4 and IL-10 on O₂⁻ production cannot be ruled out.

NO has been shown to have both anti-inflammatory as well as proinflammatory properties (26, 27, 31, 53). While NO has been shown to be important in the control of many parasitic, bacterial, and viral infections as well as certain malignancies (53), it has also been implicated in the suppression of lymphocyte proliferation and Th1 cytokine production (28), the apoptosis of macrophages (53), and the scavenging of O₂⁻ (53). NO is produced during infection with S. mansoni, and the onset of production correlates to the start of egg laying by the female worms (3, 30). In WT mice, it appears that NO serves a protective function in the liver, as treatment with aminoguanidine, an inhibitor of iNOS, leads to weight loss and severe liver damage during the acute stage of the infection (30). One possible mechanism by which NO could protect the liver involves the scavenging of the ROI that are generated in response to eggs (31, 54). Specifically, NO can compete with SOD for O₂⁻, form ONOO⁻, and thus reduce the levels of hydrogen peroxide produced after the dismutation of O₂⁻ by SOD. While ONOO⁻ is more reactive than either O₂⁻ or NO, it is less damaging than the hydroxyl radical or hypohalous acids that can be formed from excess hydrogen peroxide (37, 40). Our results indicate that in infected WT mice O₂⁻ and NO are generated in the livers of infected animals, but that ONOO⁻ production is limited, supporting the idea that NO may be regulating the generation of highly reactive oxidative species in the liver.

In contrast to the beneficial effects of NO during infection of WT animals, the overproduction of NO in the absence of IL-4 contributes to severe liver damage and mortality during schistosome infection. Increased production of NO by splenocytes from infected IL-4^−/− mice has been shown previously (3, 30), and recently, it has been found to play a pivotal role in the suppression of lymphocyte responses in these animals (E. A. Patton, A. C. La Flamme, and E. J. Pearce, manuscript in preparation). In this report we expand upon these findings and demonstrate that the expression of iNOS transcripts and protein is significantly increased in the absence of IL-4 and that higher levels of NO are generated by the cells in the granulomas from IL-4^−/− mice compared with WT. Furthermore, the overproduction of both NO and O₂⁻ leads to increased ONOO⁻ formation. These results indicate that while the production of NO may be beneficial during infection in WT mice by scavenging O₂⁻, in the absence of IL-4 the generation of high levels of NO converts this protective mechanism to a damaging one.

In the liver, SOD, catalase, glutathione, and glutathione peroxidase are important enzymes that protect against oxidative damage caused by the production of ROI/RNI (37). The effect of schistosome infection on the levels of these enzymes in the liver has recently been investigated by Gharib et al. (55). This study determined that the antioxidant defenses were reduced in the livers of infected WT mice, as measured by decreased activity of catalase, SOD, glutathione peroxidase, and glutathione, and this reduction was most evident at later time points in infection (55). In agreement with these results, we found modest decreases in SOD and catalase 7 wk after infection in WT mice and greater decreases at later times (data not shown). Furthermore, the level of catalase, but not SOD, is dramatically reduced in the livers of infected IL-4^−/− mice compared with WT mice, thus making these mice more susceptible to oxidative damage mediated by hydrogen peroxide or hydroxyl radical. SOD has been shown to be important in the protection against ONOO⁻ damage in amyotrophic lateral sclerosis (56), and although only modest decreases in SOD were found in IL-4^−/− mice, this reduction combined with the increased ONOO⁻ generation may also be contributing to the increased ROI-mediated liver damage.

The mechanism responsible for the decreased production of antioxidant defenses in the absence of IL-4 is unclear. Since nearly all cells can express the receptor for IL-4 (57), this cytokine may be directly regulating the expression of antioxidant proteins by hepatocytes. Support for this view is provided by the finding that the addition of IL-4 to cultures of primary human hepatocytes increases the production of GST (58). The responsiveness of hepatocytes to IL-4 is also indicated by the finding that the cytokine inhibits lipogenesis stimulated by IL-1, IL-6, and TNF-α in murine hepatocytes (59). Because the production of all Th2 cytokines is dramatically reduced, and the production of the inflammatory cytokines (e.g., IFN-γ and TNF-α) is increased during infection in IL-4^−/− mice (3, 10), it is also probable that differences in the levels of other cytokines may be influencing the production of catalase and SOD in the liver (60, 61, 62). Together these findings suggest that IL-4 protects against severe liver damage by reducing the production of ROI/RNI as well as by helping to maintain the levels of antioxidant enzymes in the liver during infection.

To directly verify that the overproduction of ROI/RNI contributed to liver damage in infected IL-4^−/− mice, these mice were treated with various antioxidants. Uric acid has been shown to be effective in scavenging ONOO⁻ (25, 47) and has been effectively used to treat mice suffering from experimental autoimmune encephalomyelitis (25). Similarly, treatment of S. mansoni-infected IL-4^−/− mice reduced the formation of nitrotyrosine, a marker of ONOO⁻ (40, 44), indicating that uric acid was effectively scavenging ONOO⁻ in vivo. Furthermore, uric acid treatment resulted in a significant reduction in mortality and morbidity. Using the synchronously induced pulmonary granuloma model, previous studies have shown that WT mice treated with vitamin E, SOD, or catalase have a 40–60% reduction in granuloma size (48). In our study, no significant differences in granuloma sizes were observed after uric acid treatment. The amelioration of disease pathology in the infected IL-4^−/− mice by treatment with antioxidants implicates ROI/RNI in severe liver disease after infection with S. mansoni and indicates that IL-4 plays an important role not only in the generation of Th2 responses, but also in regulating ROI/RNI and maintaining anti-oxidant defenses in the liver during schistosomiasis.

We thank Drs. Sharon McGonigle, Melissa Beall, Ilma Araujo, Laura Rosa-Brunet, and Andrew MacDonald for helpful discussions.