Resistance to the African Trypanosomes Is IFN-γ Dependent¹

The role of Th cell phenotype in determining the comparative level of host resistance to a variety of microbial pathogens has been well documented (reviewed in Refs. 1–5). From this body of work has emerged the idea that resistance and susceptibility are often associated with predominant Th1 or Th2 cell cytokine responses to microbial Ags. In the well-characterized Leishmania spp. model systems, for example, mouse strains that mount a strong Th1 cell response coupled with IFN-γ activation of macrophages are able to destroy the parasites and localize the infection, while mice producing IL-4, a Th2 cell cytokine, exhibit exacerbated disease and disseminated infection (reviewed in Refs. 6–10); these experimental studies have been mirrored at the clinical level as well (11, 12). Additional work in the Leishmania model system has demonstrated that in the absence of IFN-γ, resistant mice are unable to control infection, and neutralization of IL-4 allows susceptible mice to resolve infection (13). Similar findings in experimental infections with Candida albicans (14, 15), Borrelia burgdorferi (16), Listeria monocytogenes (17), Trichuris muris (18), and influenza virus (19) as well as in human infections with Mycobacterium tuberculosis (20) and M. leprae (21), support the idea that the nature of the emerging or predominant Th cell response may dramatically affect the outcome of disease.

African trypanosome infections are partially controlled by host B cell responses to variant surface glycoprotein (VSG)³ determinants expressed by trypanosome variants appearing in the bloodstream (reviewed in Refs. 22–25). However, it is clear that VSG-specific B cell responses and control of parasitemia are not functionally or genetically linked alone to overall host resistance, and that other immune elements contribute to resistance (25, 26, 27). Recent work in this laboratory has demonstrated that mice with the BL/10 or BL/6 genetic backgrounds, which are relatively resistant to Trypanosoma brucei rhodesiense, mount polarized VSG-specific Th1 cell responses following infection (28, 29, 30). This response consists of IFN-γ and IL-2 secretion in response to parasite Ags and includes a degree of tissue compartmentalization in which peritoneal CD4⁺ T cells produce the strongest cytokine responses. Because Th cell responses have been associated with resistance to other parasitic organisms, we asked whether Th1 cell-associated cytokine responses are associated with host resistance to African trypanosomes. This question assumes a controversial dimension because of studies suggesting that IFN-γ is a requisite growth factor for trypanosomes rather than a factor that contributes to host resistance (31, 32, 33, 34).

In the present study, therefore, we have examined parasite Ag-specific Th cell responses in relatively resistant B10.BR and susceptible C3H mice as well as in resistant wild-type and IFN-γ or IL-4 cytokine gene knockout (KO) mice. We found that while infected B10.BR and C57BL/6 mice produced strong Th1 cell cytokine responses upon stimulation with trypanosome VSG, C3H mice produced weak or no detectable Th1 cell cytokine responses. No infected mouse strain exhibited evidence of Th2 cell cytokine responses to parasite Ag. Using C57BL/6 wild-type and cytokine gene KO mice, we showed that IFN-γ production associated with the Th1 cell responses is a critical factor in determining the relative level of host resistance, and that the ability to produce IL-4 is neither detrimental nor helpful to the host.

Age-matched female B10.BR/SgSnJ, C3HeB/FeJ, C57BL/6J (wild-type), C57BL/6J-Ifg (IFN-γ KO) (35), C57BL/6J-Il4 (IL-4 KO) (36), and C57BL/6J-scid/SzJ (scid) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were 8–12 wk old at the time of infection. Outbred Swiss mice (retired breeders; Harlan Sprague-Dawley, Madison, WI) were used for propagation of trypanosome stabilates. All animals were housed in university-approved facilities.

Trypanosoma brucei rhodesiense clone LouTat 1 was used in all experiments. Frozen stabilates of the parasite were thawed and used to infect Swiss mice that had been immunosuppressed with cyclophosphamide (300 mg/kg body weight; Cytoxan, Mead Johnson and Co., Evansville, IN) (37). After 4–5 days of infection, blood was collected and diluted in PBS and 1% glucose (PBSG), and the trypanosomes were counted in a hemocytometer. Experimental infections were established by injecting 1 × 10⁵ trypanosomes i.p. in a total volume of 0.3 ml of PBSG.

For parasitemia studies, trypanosomes in the blood were enumerated by the method of Kolmer (38, 39). Briefly, 10 μl of tail blood was collected into a heparinized capillary tube and expelled into 90 μl of trypanosome staining buffer (2% formalin, 2% glacial acetic acid, and 2% Ziehl-Neelson’s carbol fuchsin in water). Appropriate dilutions of the suspension were made in trypanosome staining buffer, and the trypanosomes were counted using a hemocytometer.

Trypanosomes used for VSG isolation were separated from blood components by passage over Selectacel DEAE (Polysciences, Warrington, PA), and the VSG was purified by a modification of procedures previously described (28, 40). Briefly, trypanosomes were resuspended in 0.3 mM zinc acetate (1 × 10⁹ trypanosomes/ml), and incubated on ice for 5 min, followed by 10 min of centrifugation at 1000 × g. The supernatant fluid was decanted, the protease inhibitors N-α-p-tosyl-l-lysine chloromethyl ketone (0.1 mM; Sigma Chemical Co. Chemical, St. Louis, MO) or aprotinin (0.01 mg/ml; Sigma, St. Louis, MO) and leupeptin (0.04 mg/ml; Sigma) were added, and the mixture was placed on ice. The pellet from the same centrifugation step was resuspended in PB with PMSF (0.05 mM; BRL, Rockville, MD) and N-α-p-tosyl-l-lysine chloromethyl ketone (0.1 mM). The suspension was incubated at 37°C for 20 min, cooled on ice, and then centrifuged at 5,000 × g for 15 min. From this second centrifugation, the pellet was discarded; the supernatant fluid (in PB) was saved but handled separately from the earlier supernatant (in zinc acetate). Both supernatant fluids subsequently were centrifuged at 300,000 × g for 70 min at 4°C, collected, and kept on ice. The zinc acetate fraction was dialyzed in 50,000 MWC dialysis tubing against 100 vol of phosphate buffer for 5–6 h (4°C). Both fractions were pooled and concentrated using a Centriprep-30000 concentrator (Amicon, Beverly, MA). The resulting concentrate was applied to a DEAE-Sephadex column (Sigma) equilibrated with PB. Fractions containing protein (as determined by absorbance at 280 nm) were pooled. The buffer was changed to PBS, and the protein was concentrated using a Centriprep-30000. The VSG was filter-sterilized through a 0.2-μm filter and stored at −70°C. The concentration of the purified protein was determined using the BCA Protein Assay System (Pierce, Rockford, IL) compared with a BSA standard (Pierce), and the purity of VSG was determined by SDS-PAGE as we previously reported (41, 42).

Cells for in vitro culture were collected from infected mice at different times postinfection. Mice were exsanguinated by retro-orbital bleeding under anesthesia (Metofane, Mallinckrodt Veterinary, Mundelein, IL). Peritoneal cells (PC) were subsequently collected by injecting 5 ml of tissue culture medium (RPMI 1640; JRH Biosciences, Lenexa, KS) containing gentamicin (50 μg/ml; Life Technologies, Grand Island, NY), HEPES (16 mM; JRH Biosciences), and heparin (2 U/ml; Amersham, Arlington Heights, IL) without FBS i.p. and mechanically agitating the mice for 4 min at 3500 rpm on an orbital shaker. The PC were then withdrawn aseptically using an 18-gauge needle and syringe, and washed in tissue culture medium (c-RPMI) supplemented with 10% FBS and l-glutamine (JRH Biosciences). Spleens were removed and dissociated to single cell suspensions (SPC) in c-RPMI. RBC and debris removal was achieved by hypotonic lysis and subsequent passage of the cell suspension over a glass-wool column. Cells were washed in c-RPMI and counted with trypan blue and a hemocytometer. Splenocytes for adoptive transfer experiments were prepared in the same manner, except cells were washed an additional two times in PBS and resuspended in PBS immediately before i.v. injection.

Cultures of SPC (1.2 × 10⁷ cells/well) or PC (3 × 10⁶ cells/well) were incubated for 24 h at 37°C in 7% CO₂ in six-well plates (6 ml of medium/well). Cells were cultured in c-RPMI alone, with Con A (2.5 μg/ml, Sigma), or with VSG (50 μg/ml) for 24 h. At the end of the incubation period, supernatant fluids were collected, centrifuged to remove cells, and stored at −20°C until analysis. RNA was isolated from cultured cells using Ultraspec (Biotecx, Houston, TX) according to the manufacturer’s specifications.

Levels of IFN-γ, IL-2, and IL-4 in culture supernatant fluids were determined by sandwich ELISA. Immulon 4 plates (Dynatech, Chantilly, VA) were coated by incubating plates at 4°C overnight with capture Ab diluted in PBS. Plates were washed four times with PBS and 0.05% Tween-20 (PBST; Sigma) and blocked by incubation for 30 min at 37°C with PBS and 2% BSA. One hundred microliters of culture supernatant fluid or a dilution of supernatant fluid in medium was added to each well and incubated for 2 h at 37°C. The plate was again washed four times with PBST before addition of biotinylated detection Ab (diluted in PBS and 2% BSA). Following additional incubation (1 h at 37°C) and washing (four times with PBST) steps, avidin-conjugated alkaline phosphatase was prepared and used according to the manufacturer’s instructions (Vectastain ABC-AP, Vector Laboratories, Burlingame, CA). The substrate p-nitrophenylphosphate disodium (Sigma) was prepared at a concentration of 2 mg/ml in 0.1 M sodium carbonate/bicarbonate buffer, pH 9.6, and 0.01 M MgCl₂. The absorbance at 405 nm was measured with a microtiter plate reader (EL311, BioMetallics, Princeton, NJ), and the concentration of cytokine in each well was calculated by comparison with a standard curve of recombinant cytokine (IL-2, IL-4 (PharMingen, San Diego, CA), or IFN-γ (Genzyme, Cambridge, MA)) using Deltasoft II software (BioMetallics). Anti-IFN-γ Ab (R4-6A2 ascites fluid; 1/8000) was used as the capture Ab in the IFN-γ ELISA. Other Abs used in ELISA were purchased from PharMingen (San Diego, CA) and used at the concentrations noted: IFN-γ detection, 0.5 μg/ml; IL-2 capture, 0.5 μg/ml; IL-2 detection, 0.25 μg/ml; IL-4 capture, 0.25 μg/ml; and IL-4 detection, 0.5 μg/ml.

IL-2 and IL-4 activities were also assessed using the CTLL-2 bioassay system (43). The cells were grown in c-RPMI supplemented with 5% T-Stim (culture supplement, rat, with Con A; Collaborative Biomedical Products, Bedford, MA) and were routinely passaged or used for assays upon reaching stationary growth. Cells were washed in unsupplemented culture medium three times before the assay to remove residual growth factors provided by the culture supplement. Standard curves were prepared with rIL-2 and rIL-4 (PharMingen). Cultures of 5 × 10³ cells/well with the test culture supernatant fluids were set up in triplicate in 96-well plates. To determine which cytokine was inducing proliferation, blocking concentrations of Ab to IL-2 (S4B6 ascites), IL-4 (11B11 ascites), or both were added to replicate cultures. The plates were incubated at 37°C in 5% CO₂ in air. After 24 h of incubation, [³H]thymidine (1 μCi/well) was added, and the plates were incubated for an additional 24 h. The cultures were harvested onto nylon-backed glass-fiber filters (Packard, Meriden, CT) using an Inotech cell harvester (Inotech Biosystems International, Lansing, MI) and counted with a Matrix 9600 direct beta counter (Packard).

IFN-γ and IL-4 transcript levels in cultured cells were determined by quantitative competitive PCR (QC-PCR) as we have previously described (30). Total RNA in each cell culture sample was quantitated by measuring absorbance at 260 nm after purification. cDNA was generated in a reverse transcription reaction (AMV Reverse Transcription Kit, Promega, Madison, WI) according to the directions of the manufacturer and was then used in QC-PCR reactions using the method described by Clontech (Palo Alto, CA) and others (44, 45). Reactions containing cDNA were set up with fourfold dilutions of a known amount of competing DNA fragment (MIMIC) that contained 5′ and 3′ regions complementary to sequences of the gene-specific primers. To calculate the number of copies of a given gene transcript within an RNA sample, an internal standard for RNA quantitation and efficiency of the RT reaction, glyceraldehyde-3-phosphate dehydrogenase (G3PDH), was also determined. Primers used in quantitation reactions are as follows: G3PDH upstream, 5′-TGA AGG TCG GTG TGA ACG GAT TTG GC-3′; G3PDH downstream, 5′-CAT GTA GGC CAT GAG GTC CAC CAC-3′; IFN-γ upstream, 5′-CAT CTT GGC TTT GCA GCT CTT CCT CAT GGC-3′; IFN-γ downstream, 5′-TGG ACC TGT GGG TTG TTG ACC TCA AAC TTG GC-3′; IL-4 upstream, 5′-GAG ATC ATC GGC ATT TTG AAC-3′; IL-4 downstream, 5′-GCT CTT TAG GCT TTC CAG GAA GTC-3′; IL-2 upstream, 5′-ATG TAC AGC ATG CAG CTC GCA TC-3′; IL-2 downstream, 5′-GGC TTG TTG AGA TGA TGC TTT GAC A-3′. Following a “hot start” protocol (94°C for 3 min, cooling to 42°C), Taq polymerase (Promega) was added, and the reactions were cycled in a thermocycler (MJ Research, Watertown, MA) under the following conditions (G3PDH, IFN-γ, IL-2): initial denaturation at 94°C for 1 min, 30 cycles of denaturation at 94°C for 45 s, annealing at 60°C for 45 s, and extension at 72°C for 2 min. A final extension step of 5 min at 72°C was performed. Conditions for IL-4 were the same, except the annealing temperature was 55°C, and 33–35 cycles were run.

Sera from normal and infected mice were tested in Ab isotype-specific ELISAs for VSG reactivity as we have previously described (30). Briefly, Immulon-4 (Dynatech) plates were coated with 4 μg/ml purified VSG overnight at 4 C. The plates then were washed, blocked, and washed again as described for the cytokine ELISAs. A dilution series of individual serum samples from different time points of infection was added to wells in triplicate and incubated for 2 h at 37 C followed by another wash. The VSG-specific Ab ELISA plates then were incubated with horseradish peroxidase (HRP)-conjugated rat anti-mouse IgM or IgG1 or with HRP-conjugated rabbit anti-mouse IgG3 (Zymed, South San Francisco, CA) or alkaline phosphatase-conjugated sheep anti-mouse IgG2a (The Binding Site, Birmingham, U.K.) for 1 h at 37 C. Plates were washed four times, then the substrate o-phenylenediamine dihydrochloride (Sigma) for the HRP conjugates or p-nitrophenylphosphate disodium for the alkaline phosphatase conjugate was added, and the assay was developed for 10 min at 25 C in the absence of light. The OD₄₉₀ (IgM, IgG1, and IgG3) or OD₄₀₅ (IgG2a) reading for each well was determined using an automated plate reader.

To determine whether there were differences in the Th cell responses of trypanosome-infected mice with different resistance phenotypes, supernatant fluids from VSG-stimulated lymphocyte cultures were assayed for the presence of IFN-γ, IL-2, and IL-4 (Fig. 1). Cytokines present in VSG-stimulated cultures were compared with those in unstimulated cultures to differentiate Ag-specific cytokine secretion from spontaneous cytokine release. As shown in Fig. 1 A, VSG-stimulated PC from relatively resistant B10.BR mice (MST, >60 days) secreted significant amounts of IFN-γ, whereas little, if any, IFN-γ was detectable in supernatant fluids derived from cells of relatively susceptible C3H mice (MST, <20 days). Secretion of IFN-γ by SPC of both mouse strains in response to parasite Ag was less than that of PC (data not shown), consistent with our earlier observations of relatively compartmentalized Th cell responses during trypanosome infection (28, 30).

The CTLL-2 bioassay data (not shown) demonstrated that biologically active IL-2, but not IL-4, was present in supernatant fluids of VSG-stimulated PC of B10.BR mice; the CTLL-2 cell proliferative responses were blocked with anti-IL-2 Ab, but not with anti-IL-4 Ab, indicating that IL-2, but not IL-4, induced proliferation of the cells. In contrast, proliferation of CTLL-2 target cells was not induced by supernatant fluids of Ag-stimulated cells from C3H mice. The presence of IL-2 in the same culture supernatant fluids was quantitated by ELISA (Fig. 1,B). VSG-stimulated PCs from B10.BR mice produced measurable IL-2, while no IL-2 was detectable in VSG-stimulated supernatant fluids of cells from C3H mice. In comparison and contrast, no IL-4 was detected by ELISA (Fig. 1 C) in culture fluids from either mouse strain. Previously we have shown that the VSG-specific Th cell response is the major component of total parasite Ag-specific Th cell responses occurring during infection (28, 29, 30, 46). Therefore, taken together, these results support our earlier findings that B10.BR mice produce a partially compartmentalized Th1 cell cytokine response and demonstrate, in addition, that C3H mice produce quantitatively weaker parasite Ag-specific Th1 cell responses without producing a qualitative shift to Th2 cell responses.

To confirm that VSG-specific cytokine responses were exclusively of the Th1 type, and that the absence of IL-4 in culture supernatant fluids was not due to an inability to detect low levels of IL-4 product, levels of IFN-γ and IL-4 mRNA in cultured cells were quantitated by QC-PCR. SPC and PC from uninfected B10.BR or C3H mice cultured in medium exhibited no background levels of either IFN-γ or IL-4 mRNA, confirming what we have shown previously (30). After stimulation with VSG, SPC and PC from infected B10.BR mice showed a marked increase in IFN-γ transcripts compared with cells cultured in medium alone, while a smaller increase was observed in PC from C3H mice (data not shown); little or no increase in IFN-γ mRNA was detected in SPC of infected C3H mice. Similar analysis of IL-4 transcript levels showed no VSG-induced IL-4 mRNA in SPC or PC cultured from either mouse strain, although very low levels of IL-4 transcript were occasionally detected in freshly isolated cells from these mice, similar to our previous observations (30). Together, these findings confirm the results of the ELISA and CTLL assays, showing that while the relatively resistant B10.BR mice produce a strong Th1 cytokine response, this response is much weaker in the relatively susceptible C3H mouse strain. Furthermore, there is no evidence, by measurement of either IL-4 product or transcript, for Ag-specific Th2 cell responses in either the resistant or susceptible mouse strains. The absence of IL-4 was indicative of the absence of other Th2 cell cytokine responses, such as IL-5 (data not shown), as we have previously demonstrated, and the cytokine profiles presented here represent the peak responses observed in response to VSG following infection (28, 30, 46).

Because we detected quantitative, but not qualitative, differences in the parasite Ag-specific Th1 cytokine responses of resistant B10.BR and susceptible C3H mice, we asked whether Th cell cytokines played any role in determining host resistance. To specifically address the potential roles of IFN-γ and IL-4 in host resistance, we used IFN-γ and IL-4 KO mice with the C57BL/6 genetic background (35, 36). Wild-type C57BL/6 mice are relatively resistant to T. brucei rhodesiense LouTat 1 and have a MST of 44 days or more (47). In this study C57BL/6 mice were found to survive an average of 46 days, while the IFN-γ KO mice survived only 19 days, similar to the MST of scid mice (Fig. 2) and genetically susceptible C3H mice (27, 48), indicating that IFN-γ is critical for host resistance. Survival times of the IL-4 KO mice did not differ significantly from the wild-type mice indicating that, in contrast to IFN-γ, IL-4 has a neutral impact on host resistance.

To examine the parasite-specific immune response in the wild-type and cytokine gene KO mice, we measured the levels of cytokines in VSG-stimulated cell culture supernatant fluids. While C57BL/6 mice exhibited a polarized Th1 cell response, these mice produced quantitatively less IFN-γ than the somewhat more resistant B10.BR strain (47). IFN-γ and IL-2 were secreted by VSG-stimulated PC (Fig. 3, A and B), while Ag-stimulated SPC populations secreted relatively lower levels of IFN-γ and IL-2 (data not shown). As expected, cells from IFN-γ KO mice did not secrete IFN-γ. Some biologically active IL-2 was secreted by VSG-stimulated PCs, but levels were markedly reduced compared with those in the wild-type mice (data not shown). However, neither wild-type nor IFN-γ KO mice produced detectable IL-4, as determined by both CTLL bioassay and ELISA (Fig. 3,C). Analysis of mRNA from the same cultured cells confirmed that induction of IFN-γ transcripts occurred in VSG-stimulated cultures of PC from wild-type mice, while there was no VSG-specific increase in levels of IL-4 mRNA (Fig. 4). IL-4 KO mice had cytokine responses similar to those of the wild-type mice, although levels of IFN-γ and IL-2 secretion were higher in the IL-4 KO mice compared with those in the wild-type mice (Fig. 5). Susceptible scid mice made no detectable cytokines in response to VSG (data not shown), lending additional support to the idea that Th1 cells are mediating resistance as a result of Ag-induced IFN-γ secretion.

Because IFN-γ was shown to be important for determining the relative level of host resistance in this study, yet studies by others have suggested that IFN-γ is a growth factor for trypanosomes, we next asked whether the absence of IFN-γ in the IFN-γ KO mice had any effect on patterns of parasitemia. For this purpose, parasitemia was determined on a daily basis in blood sampled from the tail vein (Fig. 6). For the first 7 days of infection, all wild-type and IFN-γ KO mice displayed similar patterns of parasitemia; typically, the first wave of parasites reached a peak level in the blood on day 4 or 5 of infection and was completely cleared by day 7. All mice displayed a period of remission before the return of detectable numbers of trypanosomes in the blood. In wild-type mice, the second wave of parasitemia was not as high as the first wave, and three of five mice went on to clear this second wave completely by day 21, while two mice had very low, but detectable, levels of trypanosomes in the blood at that time. In contrast, the period of remission between waves was slightly shorter in the IFN-γ KO mice compared with that in the wild-type mice, and IFN-γ KO mice experienced a second wave of parasitemia that was 3–10 times higher than the first wave. This second wave of parasitemia was never cleared by the IFN-γ KO mice before they succumbed to infection. Although the pattern of parasitemia in experimental trypanosome infections is known to be somewhat heterogeneous in timing and magnitude after the first wave (49), the overall pattern that emerges from this study is one of earlier and higher relapsing parasitemia in the IFN-γ KO mice compared with that in wild-type mice.

Ab responses directed against VSG determinants expressed on the trypanosome surface are functionally and genetically linked to the clearance of parasites from the blood (50), but are not linked alone to overall resistance to trypanosomes (26, 27). To determine whether the absence of IFN-γ was affecting host resistance and patterns of parasitemia by altering Ab isotype switch patterns, we compared the isotypes and quantities of VSG-specific Abs in wild-type and IFN-γ KO mice. Levels of VSG-specific IgM and IgG1 were similar in serum from representative 2-wk-infected wild-type and IFN-γ KO mice, while the level of IgG3 was slightly higher in wild-type mice (Fig. 7). There was, however, a complete absence of VSG-specific IgG2a in IFN-γ KO mice compared with that in wild-type mice. This is consistent with the dependency of the Ig Cμ to Cγ2a switch response on IFN-γ and confirms our recent findings on cytokine-induced Ig switching events in trypanosomiasis (30).

To determine whether the presence of VSG-specific IgG2a was associated with the relative resistance of wild-type mice, we pooled the serum from uninfected or 2-wk-infected wild-type and IFN-γ KO mice and administered 1 ml of serum i.p. to IFN-γ KO mice on day 7 of infection, just after clearance of the first peak of parasitemia. The mice were then monitored for survival. There was no difference in the MST of mice receiving serum from infected wild-type mice compared with that of mice receiving serum from normal or infected IFN-γ KO mice or serum from normal wild-type mice (Table I). These results suggest that VSG-specific Abs produced by wild-type mice are not responsible for the enhanced resistance of these mice compared with IFN-γ KO mice.

To evaluate whether IFN-γ-producing lymphocytes mediate host resistance we transferred splenocytes from uninfected wild-type or IFN-γ KO mice to IFN-γ KO mice on days −1 and +14 of infection. As shown in Fig. 2, IFN-γ KO mice receiving cells from wild-type mice survived an average of 28 days, while control mice receiving cells from IFN-γ KO mice showed no enhanced resistance compared with nonmanipulated mice. Reconstituted IFN-γ KO mice did not survive as long as wild-type mice, however, presumably due to the lower numbers of IFN-γ-producing cells in reconstituted mice compared with wild-type mice. Regardless, these results clearly show that resistance to trypanosomes in IFN-γ KO mice can be significantly enhanced by the passive administration of cells capable of producing IFN-γ in response to parasite Ag.

Infection with African trypanosomes results in a range of disease phenotypes that is dependent upon both host immunity and parasite virulence (27, 47, 51, 52, 53, 54, 55, 56). Several elements of the host immune response have been examined in experimental animal model systems for an influence on the level of host resistance. An early finding was that the ability to mount VSG-specific Ab responses and control parasitemia was not functionally or genetically linked alone to host resistance. For example, while mouse strains that produced VSG-specific Ab had a tendency to be more resistant to infection (48), some strains that were unable to make an Ab response were more resistant to trypanosome infection than mice that produced parasitemia-controlling Abs (26, 27). Further, F₁ hybrids derived from resistant and susceptible parental strains inherited the ability to produce VSG-specific Ab and to control parasitemia from the resistant parent, yet inherited the susceptibility phenotype from the susceptible parent (26, 27). Taken together, these studies demonstrated that genetically based immunological factors other than Ab responses to VSG were important for determining relative host resistance. This conclusion is supported by the present study in which infected wild-type and IFN-γ KO mice both made Ab responses to trypanosome VSG and controlled parasitemias, yet differed markedly in their susceptibility to infection.

The discovery of T cell responses to trypanosome Ags, including VSG, opened a new door in the analysis of host resistance to trypanosomes (28, 29, 30, 46). In the present study we have revealed for the first time a contributing role for VSG-specific Th1 cell IFN-γ production in determining host resistance to African trypanosomes. Using B10.BR/SgSnJ and C3HeB/FeJ mice as prototypical resistant and susceptible mouse strains (27, 47, 48, 57), respectively, we have shown that there are quantitative differences in the levels of parasite Ag-induced Th1 cell cytokines produced by these two mouse strains. There was no evidence for Ag-specific Th2 cell responses occurring in either mouse strain, however; this finding was supported by an analysis of Ag-induced cytokine mRNA. Together these results suggest that the difference in resistance to trypanosomes displayed by these mice may be due to quantitative and not qualitative differences in Th cell activation. Results of cytokine analyses and survival studies performed with wild-type, IFN-γ KO, IL-4 KO, and scid mice with the relatively resistant C57BL/6 genetic background provided direct evidence that IFN-γ production associated with Th1 cell responses is an important component of protective immunity, and that Th2 cell cytokine responses are not involved in determining host resistance or susceptibility.

These results reveal several interesting features of experimental trypanosomiasis. First, unlike other model systems of infectious disease where qualitative differences in Th cell responses may determine disease outcome, trypanosomes appear to elicit quantitative differences in Th1 cell responses that are associated with the level of host resistance. For example, the emerging pattern in one parasitic model system, Leishmania major (58), has been that a dominant Th1 cell response leads to resistance, whereas a dominant Th2 cell response leads to susceptibility. The opposite pattern is true for the helminths Trichuris muris (59) and Nippostrongylus brasiliensis (60) as well as for organisms such as Borrelia burgdorferi (16), where Th2 cell responses lead to resolution of the infection, and Th1 cell responses are associated with susceptibility. In either case, the focus has been on qualitative differences (Th1 vs Th2) in response to a pathogen and the association of these qualitative differences with resistance or susceptibility. In contrast to this paradigm, we observed quantitative differences in Th1 cell cytokine profiles between resistant and susceptible mice, and we found no evidence for Th2 cell responses in any infected mouse strain. A similar difference in the development of Th cell responses has been reported during infection with Yersinia enterocolitica. Resistant C57BL/6 mice infected with Y. enterocolitica were strong Th1 cell responders and produced high levels of IFN-γ, while susceptible BALB/c mice produced a delayed IFN-γ response; there was no evidence for Th2 cell cytokine production (61). Thus, a mounting body of evidence, including our findings with T. brucei rhodesiense infections, challenges the general idea that differences in resistance and susceptibility are due to qualitative differences in Th cell cytokine responses. Rather, we observed that the magnitude of the Th1 cell response is important in determining host resistance, and that susceptibility is not associated with outgrowth of a Th2 cell response.

A legitimate question is whether T cell priming with VSG occurs during trypanosome infection of susceptible mouse strains, in that such strains may be incapable of recognizing parasite Ags. However, we recently demonstrated that infected susceptible IFN-γ KO mice are defective in suppressor macrophage activity, which is associated with depressed proliferative T cell responses (46), and such mice made proliferative T cell responses to VSG.⁴ Thus, this finding demonstrates the underlying existence of VSG-reactive T cells in susceptible mice.

The mechanism by which Th1, but not Th2, cell cytokine responses are preferentially generated during trypanosome infection appears to be due to two distinct effects.⁵ First, IFN-γ-independent production of IL-12 occurs within hours of infection and is necessary for the early development of VSG-specific Th1 cell responses in resistant mice. Second, in the absence of IL-12 or IFN-γ, Th2 cell responses do not develop and appear to be inhibited, directly or indirectly, by infection or by the trypanosomes themselves. Thus, the induction of IL-12 appears to play an important role in enhancing the development of Th1 cell responses, but the parasites also appear to specifically depress the outgrowth of Th2 cell responses by an unknown mechanism. Since we have observed that susceptible strains of mice produce lower levels of IL-12 during the early time points of infection compared with more resistant strains,⁵ it may be that the stimulus for Th1 cell outgrowth is weak, while the inhibitory effect in terms of Th2 cell outgrowth is strong, resulting in the T cell phenotype seen in susceptible mice.

Regardless of the mechanisms leading to polarized Th1 cell responses during infection, our studies clearly and unambiguously demonstrate that IFN-γ is important in determining host resistance. This finding contradicts one proposed role for IFN-γ in infection with African trypanosomes. Trypanosomes have been reported to produce a factor known as TLTF (31, 32, 62, 63), which stimulates naive CD8⁺ cells to release IFN-γ in an Ag nonspecific manner. It has been proposed that IFN-γ produced by CD8⁺ cells as a result of TLTF stimulation is used by the parasite as a requisite growth factor (31, 34). In support of this hypothesis, studies have demonstrated that neutralization of IFN-γ in vivo was found to enhance the survival of both susceptible C3H/HeJ mice and resistant C57BL/6 mice (64). Furthermore, it was determined that IFN-γ added to cultures of T. brucei stimulated the proliferation and survival of the parasite (34). However, our cumulative findings with T. brucei rhodesiense are inconsistent with TLTF-induced IFN-γ serving as a growth factor for trypanosomes. First, we have shown previously that the primary source of IFN-γ produced during infection in response to parasite Ags is CD4⁺ T cells expressing the TCR-αβ and recognizing antigenic peptides in an APC-dependent and MHC II-restricted manner (28). Second, we have also demonstrated in earlier studies a correlation between high IFN-γ levels in serum, low parasitemias, and host resistance; in contrast, susceptible mice had no detectable serum IFN-γ and exhibited high and uncontrolled parasitemias (57). Third, we have shown in the present study that susceptible C3HeB/FeJ mice produce less (or no) IFN-γ in response to parasite Ag, whereas resistant B10.BR mice produce high levels of IFN-γ in response to Ag. And, fourth, we have demonstrated in this study that mice genetically incapable of IFN-γ production (IFN-γ KO mice) are as susceptible to infection as mice without functional T and B cells (scid mice) and that although both wild-type and IFN-γ KO mice controlled parasitemia with an Ab response to VSG, the overall parasitemias in the IFN-γ KO mice were significantly greater. Thus, our cumulative studies have demonstrated an inverse relationship between IFN-γ production and parasite growth.

One possible explanation for these discrepancies may lie with the parasite itself. While the growth-enhancing activity of IFN-γ was characterized with several different species and strains of trypanosomes, including strains of T. brucei rhodesiense (33, 34), we repeatedly have been unable to confirm that trypanosomes of the LouTar serodeme of T. brucei rhodesiense are growth responsive to IFN-γ in vitro (unpublished observations), and similar results with other trypanosomes have been obtained previously (65). It may be that some trypanosomes are more susceptible to IFN-γ-induced host resistance mechanisms (see below) than strains of trypanosomes deemed to be growth stimulated by IFN-γ; thus, the result of exposure to IFN-γ in vivo may be that any potential growth stimulatory effects of IFN-γ are negated.

Overall, the present work clearly shows that IFN-γ is essential for host resistance to T. brucei rhodesiense LouTat 1 in animals capable of controlling parasitemia with Ab responses; however, it is not yet clear how IFN-γ affords protection to the host. One potential mechanism might involve the nature of parasite-specific Abs. IFN-γ promotes a Cμ to Cγ2a Ig isotype switch in VSG-specific Abs (30), and it may be that this isotype switch event is critically important for effective parasite control. However, when immune sera from wild-type or IFN-γ KO mice containing Abs to the infecting trypanosome VSG as well as to the VSG of subsequently arising variant types were passively transferred to infected IFN-γ KO recipients, there was no detectable effect on host survival. In support of this finding is an earlier body of work demonstrating that parasite clearance is effectively mediated by the VSG-specific IgM response alone and that Ab isotype switching events occur primarily after elimination of the trypanosomes from blood (30, 50, 66).

It is more likely that IFN-γ exerts its protective effect via the macrophage cell. For example, IFN-γ serves as a potent macrophage activation agent, and we have found that when trypanosomes are cultured in the presence of macrophages and IFN-γ, the parasites are rapidly destroyed; this cytotoxicity was reversible by adding arginine analogues or Ab to IFN-γ to the culture medium (J. M. Mansfield, unpublished observations). We suspect that this cytotoxicity is due to nitric oxide (NO) production by IFN-γ-activated macrophages. In this light, we have previously demonstrated that macrophages from trypanosome-infected mice exhibit up-regulated inducible NO synthase transcript levels and spontaneously release NO (46). Furthermore, others have shown that NO exerts adverse effects on trypanosomes (67, 68, 69), and products of activated macrophages such as TNF-α have also been demonstrated to exert cytotoxic effects on trypanosomes (70, 71, 72). Furthermore, we recently completed studies showing that resistant mice, including C57BL/6 mice, exhibit earlier and quantitatively greater macrophage activation responses than susceptible mice, including IFN-γ KO mice.⁶ We suggest, therefore, that quantitative or qualitative differences in IFN-γ-induced macrophage activation may dramatically affect the course of infection.

Since we have shown that IFN-γ is critical for host resistance to T. brucei rhodesiense, and activated macrophages may serve as the basis for this resistance, we formally propose that the Th1 cell cytokine response is linked to resistance through the action of IFN-γ on host macrophages. It is likely that this IFN-γ- and macrophage-dependent component of host immunity is most effective within the extravascular tissue sites, where it complements parasite-specific Ab responses that are most effective within the vascular compartment. Together, therefore, both Th and B cell responses provide the resistance mechanisms necessary to control parasite numbers in all tissues of the trypanosome-infected host.

We thank Kathleen Schleifer for her contributions to the scid mouse survival studies, we acknowledge Anthony Byers and Suping Cai for technical assistance, and we express gratitude to Jim Schrader, Lisa Bagnifsky, and Katherine Weber for general laboratory assistance and animal care.