Experimental African trypanosomiasis: IFN-γ mediates early mortality

1 Introduction

African trypanosomes are extracellular hemoprotozoa which infect humans and livestock. Trypanosoma brucei gambiense and T. b. rhodesiense cause sleeping sickness anddeath in humans. T. congolense is the most important pathogen for livestock. Mice are highly susceptible to infections with African trypanosomes. Different strains of mice, however, display different degrees of susceptibility to the disease 1, 2. BALB/c mice are highly susceptible to T. brucei and T. congolense infections, while C57BL/6 mice are relatively resistant, as measured by the levels of parasitemia, immunosuppression and survival time 2–4.

High levels of IFN-γ have been associated with enhanced susceptibility to T. brucei and T. congolense infections, mediating immunosuppression 5, 6. Paradoxically, it also has been shown, by infection of IFN-γ-deficient C57BL/6 mice, that IFN-γ was required for resistance to T. brucei 7, 8. Recently, IL-10-deficient relatively resistant C57BL/6 mice, infected with T. brucei have been shown to have significantly shorter survival than infected wild-type mice, indicating that IL-10 might be crucial for enhanced survival of mice infected with T. brucei 8. Using highly susceptible BALB/c and relatively resistant C57BL/6 mice as a model,we have investigated the role of cytokines in immunosuppression and pathogenesis of T. congolense infections in BALB/c and C57BL/6 mice. Higher levels of IL-4, IL-10 and IFN-γ were detected in the plasma and supernatant fluids of spleen cell cultures of highly susceptible BALB/c mice infected with T. congolense as compared to those found in infected relatively resistant C57BL/6 mice 2, 6, 9. Furthermore, treatment of T. congolense-infected BALB/c mice with anti-IFN-γ mAb dramatically increased the survival time 6. In vitro, anti-IL-10 antibodies markedly decreased the immunosuppressive effect of spleen cells of T. congolense-infected BALB/c mice 9.Administration of monoclonal anti-IL-10 antibodies (of IgM class) to T. congolense-infected BALB/c mice moderately prolonged their survival time 9.

IL-10 is a pleiotropic cytokine 10. Among other functions, IL-10 plays a major role in inhibiting the production of monokines such as IL-1, IL-6, IL-12, TNF-α and in down-regulating IFN-γ synthesis by NK cells and T cells 10. The IL-10 receptor (IL-10R) is a member of the class II interferon receptor-like cytokine receptor family 11. An mAb blocking murine IL-10R was recently generated 12.

Here we report a detrimental effect of blocking the IL-10R in mice infected with African trypanosomes. To our surprise and seemingly in conflict with our previously results obtained with the use of anti-IL-10 antibodies 9, blocking the IL-10R significantly shortened the survival time of BALB/c and C57BL/6 mice infected with T. congolense as well as T. brucei-infected BALB/c and outbred CD1 mice. The acute death of T. congolense-infected mice treated with anti-IL-10R in vivo was associated with focal liver necrosis, elevated levels of IL-6, IL-10, IL-12p40 and IFN-γ in the plasma, and enhanced synthesis of IL-6, IL-12p40 and IFN-γ by spleen cell cultures. Administration of anti-IFN-γ antibody prevented the acute death induced by blocking the IL-10R. Thus, we showed, for the first time, that blocking of the IL-10R in trypanosome-infected mice results in a fatal outcome. More importantly, this is the first direct evidence that IFN-γ mediated the early mortality of infected mice when the functioning of IL-10 was prevented. We have shown that overproduction of IFN-γ killed the host infected with African trypanosomes and that overproduction of IFN-γ was held in check by IL-10 in the resistant, but not the susceptible host. We have demonstrated, for the first time, that abundant parasite antigen was located in the cytoplasm of Kupffer cells in the liver of infected BALB/c mice. Its presence was associated with an apparent state of macrophage activation. In the discussion, we present a synopsis indicating a proposed role of trypanosome-pulsed macrophages, cytokines and trypanosome-specific T cells in the fatal outcome.

2 Results

2.1 Kupffer cells of mice infected with T. congolense were loaded with parasite antigen and were greatly enlarged

Immunohistochemical staining showed that, on day 6 post-infection, parasite antigen was located in the cytoplasm of large elongated mononuclear cells in livers of infected BALB/c mice treated with PBS (Fig. 1B, 1H) or anti-IL-10R (data not shown). Parasite antigen was also observed in the cytoplasm of elongated mononuclear cells in the livers of infected C57BL/6 mice treated with PBS or anti-IL-10R (data not shown). By immunofluorescent staining, the parasite antigen-containing cells were shown to carry the macrophage marker F4/80 (Fig. 1J) and thus were identified to be macrophages of the liver (Kupffer cells). In contrast, there was no positive staining for T. congolense antigen in livers of uninfected BALB/c mice (Fig. 1A, E, G). Immunohistochemical staining also revealed that the macrophages in the livers of infected BALB/c mice were greatly enlarged, in some areas almost clogging the capillaries (Fig. 1B, D). In contrast, there were less macrophages in the livers of uninfected normal mice and the macrophages in the livers of normal mice were much smaller compared to those of infected mice (Fig. 1C, F).

2.2 Acute death of T. congolense-infected BALB/c mice was linked to a significant decrease of blood pressure

As described below, enlarged capillaries were observed in the liver sections of T. congolense-infected BALB/c mice. Based on this observation, we speculated that the blood pressure of infected mice during the acute death might decrease. Thus, we measured the blood pressure of uninfected and T. congolense-infected BALB/c mice. The blood pressure of infected BALB/c mice decreased significantly on days 7, 8 and 8.5. Thus the blood pressure started to decline about 1 to 2 days prior to the death of the mice (Fig. 2). Concomitantly with development of hypotension, the infected mice showed a drop of body temperature (not shown), hypomotility and piloerection, clinical signs also observed in bacterial endotoxic shock 13.

2.3 Administration of anti-IL-10R shortened the survival time of highly susceptible T. congolense-infected BALB/c mice

The infected BALB/c mice received either 200 μg purified anti-IL-10R mAb or 200 μg rat IgG in PBS on days 3 and 5 post-infection. Statistical analysis showed that the survival time of the T. congolense-infected mice treated with anti-IL-10R was significantly (Fig. 3A; p<0.05) shorter than that of infected control BALB/c mice treated with rat IgG, i.e. 7.3±0.2 days versus 8.5±0.3 days. Uninfected mice given anti-IL-10R did not exhibit any detrimental effects.

2.5 Administration of anti-IL-10R shortened the survival time of T. brucei-infected mice

We next infected BALB/c and outbred CD1 mice with 10³ organisms of T. brucei, strain 10–26, and then treated the mice with 100 μg of anti-IL-10R or equal amounts of rat IgG (as control) on days 3 and 5 post-infection. The survival time of both BALB/c and CD1 mice infected with T. brucei was significantly shortened by anti-IL-10R treatment (Fig. 4).

2.8 Acute death of T. congolense-infected mice treated with anti-IL-10R was associated with enhanced cytokine synthesis in vivo

We next asked the question how the anti-IL-10R treatment might influence the balance of cytokine levels in T. congolense-infected mice. Thus, we measured the plasma levels of TNF-α, IL-6, IL-10, IL-12p40 and IFN-γ in infected BALB/c (on day 7 post-infection) and C57BL/ 6 mice (on day 8 post-infection) treated with anti-IL-10R on days 3 and 5 post-infection. There were no significant differences between TNF-α plasma concentrations of anti-IL-10R-treated and PBS-treated BALB/c mice (60.0±31.8 pg/ml versus 64.1±37.5 pg/ml for mice treated with anti-IL-10R and PBS, respectively). Anti-IL-10R treatment of T. congolense-infected BALB/c and C57BL/6 mice significantly enhanced the plasma levels of IL-6, IL-10, IL-12p40 and IFN-γ by 3- to 12-fold compared to T. congolense-infected mice treated with PBS (Fig. 5).

2.9 Acute death of T. congolense-infected BALB/c and C57BL/6 mice treated with anti-IL-10R in vivo was associated with increased cytokine synthesis by spleen cells cultured in vitro

Splenocytes were collected from T. congolense-infected BALB/c (day 7 post-infection) and C57BL/6 (day 8 post-infection) mice treated with anti-IL-10R or PBS, and cultured in vitro for 48 h. The levels of IL-6, IL-10, IL-12p40 and IFN-γ in the culture supernatant fluids were determined using a sandwich ELISA. The production of IL-6, IL12p40 and IFN-γ by splenocytes cultured in vitro were significantly elevated by 3- to 35-fold for infected BALB/ c and C57BL/6 mice following treatment with anti-IL-10R mAb (Fig. 6). IL-10 levels were not significantly enhanced (data not shown).

2.10 Acute death of T. congolense-infected C57BL/6 mice treated with anti-IL-10R in vivo was prevented by administration of anti-IFN-γ antibody

We next tried to elucidate the mechanism(s) that may be involved in the early death of infected mice treated with anti-IL-10R. We did find that the acute death of infected mice treated with anti-IL-10R was associated with elevated production of IL-6, IL-10, IL-12p40 and IFN-γ. Of these cytokines, IL-10 could not exert any function since the IL-10R was blocked. We have previously shown that treatment of T. congolense-infected BALB/c mice with anti-IFN-γ could increase the survival of those mice 6. Therefore, we speculated that the excessively high levels of IFN-γ in the plasma, after blocking of the IL-10R, might be a major factor mediating the acute death of infected mice treated with anti-IL-10R. Accordingly, we administered anti-IFN-γ mAb to T. congolense-infected C57BL/6 mice whose IL-10R were blocked by anti-IL-10R. As control, we also treated infected C57BL/6 mice with anti-IL-10R only or with equivalent amounts of rat IgG. The infected C57BL/6 mice treated with anti-IL-10R alone succumbed to the infection between day 9 and day 13 post-infection, with an average survival time of 11.8±0.95 days (Fig. 7). In contrast, all of the infected mice treated with both anti-IL-10R and anti-IFN-γ mAb, as well as the infected mice treated with rat IgG alone survived (Fig. 7). These mice were killed on day 39 post-infection. Thus, the condition leading to acute death of T. congolense-infected C57BL/6 mice treated with anti-IL-10R in vivo could be reversed by administration of anti-IFN-γ mAb.

3 Discussion

How can we put the above data into perspective? We have tried to demonstrate a synopsis by two diagrams shown in Fig. 8. We believe that these diagrams, although simplistic and incomplete, represent a view that is consistent with our present knowledge.

The events of steps 1–2 (Fig. 8 A, B) are well documented: trypanosomes opsonized by antibody to variant surface glycoprotein (VSG) are phagocytosed by macrophages. Antibodies to the VSG do mediate phagocytosis by macrophages in vitro 14. Anti-VSG antibody of IgM class is as efficient as antibody of IgG2a isotype in causing phagocytosis of T. congolense in vitro 15. In mice infected with T. brucei, 60–80 % of the opsonized T. brucei end up in the liver and less, but a considerable number, in the spleen 16, 17. We have shown here that Kupffer cells of T. congolense-infected BALB/c mice are loaded with trypanosomal antigen (Fig. 1). Thus, we provided direct evidence in vivo that macrophages engulf trypanosomes in the liver of infected mice.

There is good evidence for step 3 in vivo and in vitro: macrophages that have engulfed trypanosomes are highly activated. Macrophages isolated from T. brucei-infected mice during the first wave of parasitemia have an increased ability to secrete plasminogen activator, anion superoxide, H₂O₂ 18, nitric oxide (NO) 19, and IL-10 20. Even stimulation of macrophages with homogenates of T. brucei or trypanosomal glycosylphosphatidylinositol (GPI) induces synthesis of cachectin/TNF-α and IL-1 21–23. Although not shown to be derived from macrophages, T. congolense-infected mice have elevated plasma levels of TNF-α, IL-6, IL-10 and IL-12p40 (Fig. 5 and 6, 9) at a time when Kupffer cells are pulsed with trypanosomal antigen and when these cells are in much greater numbers and their volumes much more enlarged than Kupffer cells in livers of uninfected mice (Fig. 1). Macrophages that have engulfed T. congolense or have been stimulated with T. congolense homogenates in vitro are induced to secrete TNF-α, IL-6, IL-10, IL-12 and NO 15, 24, 25.

Although we presently do not know whether there is any significant interaction between T cells and Kupffer cells in the liver, we have provided evidence for steps 4 and 5 to occur in the spleen: the synergistic action of trypanosome-specific T cells and trypanosome-pulsed macrophages induce IFN-γ production in T. congolense-infected BALB/c mice. We have shown that an unusual population of matrix-adherent T cells (adherent to plastic, nylon wool and Sephadex G-10) in synergy with trypanosome-pulsed macrophages produced a great amount of IFN-γ. The T cells were trypanosome-specific and have not been found in normal mice 26. We have extended this study by direct immunocytochemistry with double labeling. These IFN-γ-producing T cells are present at very low prevalence (<1% of total spleen cells) and are TCRγ δ^– and NK 1.1^–. They are Thy 1.2⁺, CD3⁺, TCRβ⁺, about 75% CD4⁺, 21% CD4^–8^– and 4% CD8⁺ (Shi et al, unpublished). We also have in vitro evidence that cytokine production by trypanosome-pulsed macrophages is enhanced by addition of IFN-γ 24, 25. In context of steps 4 and 5, the IFN-γ proposed to increase the activation of trypanosome-pulsed Kupffer cells might be predominantly produced in the spleen (Fig. 6) and disseminated via the circulation (Fig. 5). Because of present lack of knowledge, the potential production of IFN-γ by trypanosome-specific T cells in the liver cannot be excluded.

There is also evidence for step 6: trypanosome-pulsed macrophages produce IL-10. T. congolense-pulsed macrophages produce IL-10 24, 25 and macrophages isolated from T. brucei-infected mice have been shown to produce IL-10 20.

We believe that the scenario shown in Fig. 8A is sufficient for leading to the death of highly susceptible BALB/c mice due to a systemic inflammatory response syndrome (SIRS) 27. Anti-IL-10R treatment simply accelerates the process. The similarity of SIRS produced by bacterial endotoxin 13 to the acute death of T. congolense-infected BALB/c mice is shown by the involvement of macrophages (Fig. 1), cytokine release syndrome (Fig. 5), enlarged capillary bed (Fig. 1), decrease of blood pressure (Fig. 2), and drop of body temperature (not shown), hypomotility and piloerection at the terminal stage. In the scenario of Fig. 8B, blocking of the IL-10R leads to lack of any function of IL-10. Thus, anti-IL-10R treatment of infected C57BL/6 mice results in enhanced synthesis of IFN-γ by trypanosome-specific T cells (Fig. 6). It also results in an absence of a down-regulatory effect of IL-10 on the trypanosome-pulsed macrophages, leading to a tremendous enlargement of Kupffer cells and an up-regulation of synthesis of inflammatory cytokines, such as IL-6 and IL-12p40 (Fig. 5, 6). Altogether, the outcome is an early death of the otherwise relatively resistant C57BL/6 mice (Fig. 3).

Although we believe our view is consistent with the presently known facts, we are very aware that the painted picture is incomplete. How do the high levels of IL-10 observed in infected BALB/cmice (Fig. 5) fit into the synopsis discussed above? This observation could be interpreted to mean that our synopsis is incorrect. Our present answer, however, is this. Anti-IL-10R treatment very profoundly elevates IFN-γ levels in T. congolense-infected BALB/c mice (Fig. 5, 6), indicating that IL-10 does have a down-regulatory effect on IFN-γ synthesis of highly susceptible BALB/c mice. However, IL-10 appears to be unable to sufficiently down-regulate IFN-γ in infected BALB/c mice even without anti-IL-10R treatment (Fig. 5, 6, 9). We think there are genetic factors, which are still unknown, that make this to happen. In genetic analyses of T. congolense infections of BALB/c versus C57BL/6 mice, five different quantitative trait loci that control resistance/susceptibility have been identified 28, 29. The genes in these loci are still unknown. We suggest that the function(s) of one or more than one of these still unknown gene products in T. congolense-infected BALB/c mice is exerting an IFN-γ-inducing effect that is overriding the down-regulatory function of IL-10. Thus, in T. congolense-infected BALB/c mice, IL-10 is down-regulating the production of IFN-γ (Fig. 5), albeit insufficiently. Thus, a critical level of IFN-γ in trypanosome-infected mice appears to be lethal.

What is the nature of the interaction of T cells and trypanosome-pulsed macrophages that results in the crucial IFN-γ production? Considering the observed cytokine pattern discussed above,we think this interaction might not be MHC-restricted. Although we have no direct evidence, we speculate that trypanosomal GPI, the membrane anchor of the VSG, might be the major antigen involved. If this is correct, a CD1-dependent antigen presentation is conceivable 30, 31.

We suggest that early mortality of mice infected with T. congolense is due to SIRS 27 in which IFN-γ produced by trypanosome-specific T cells (26, Fig. 6) is a major mediator, similar to mechanisms observed in bacterial endotoxemia 32. Although most, but not all, of our studies have been carried out with T. congolense infections, we suggest that similar mechanisms operate in mice infected with T. brucei (Fig. 4).

In summary, we have presented a coherent, although still incomplete, hypothesis to explain the early death of mice infected with African trypanosomes.

4 Materials and methods