The elimination of lymphocytes within inflammatory lesions is a critical component in the resolution of disease once pathogens have been cleared. We report here that signaling through the TNF receptor p55 (TNFRp55) is required to eliminate lymphocytes from lesions associated with intracellular pathogens. Thus, TNFRp55−/− mice, but not Fas-deficient mice, maintained inflammatory lesions associated with either Leishmania major or Rhodococcus equi infection, although they developed a Th1 response and controlled the pathogens. Inflammatory cells from either L. major- or R. equi-infected C57BL/6 mice were sensitive to TNF-induced apoptosis, and conversely the number of apoptotic cells in the lesions from TNFRp55−/− mice was dramatically reduced compared with wild-type mice. Furthermore, in vivo depletion of TNF in wild-type mice blocked lesion regression following R. equi infection. Taken together, our results suggest that signaling through the TNFRp55, but not Fas, is required to induce apoptosis of T cells within inflammatory lesions once pathogens are eliminated, and that in its absence lesions fail to regress.

Apoptosis of T cells is the principal mechanism of deletion of self-reactive T cells in the thymus, and of activated peripheral T cells. Fas ligand (FasL)3 and TNF promote apoptosis of T cells by binding to cell surface expressed TNF receptors or Fas. It has been suggested that apoptosis induced by Fas-FasL interaction controls T cell clonal size after expansion in response to antigenic stimulation, and thus is critical in maintaining immune homeostasis (1, 2). Consistent with this idea are the findings that lpr and gld mice, which are defective in Fas or FasL, respectively, develop peripheral lymphadenopathy, presumably due to defective T cell apoptosis (3, 4). Nevertheless, while Fas-mediated apoptosis is defective in lpr mice, abnormal numbers of T cells are seen in these animals only as they begin to age, implicating other pathways of apoptosis that can control T cell numbers. One other pathway may involve TNF, since mice lacking both Fas and the p55 TNF receptor (also known as the type I TNFR) develop an accelerated lymphadenopathy and autoimmune disease (evident by 3 to 4 wk of age) compared with mice that lack only Fas (5). While a great deal of work has been done on the role of defective T cell apoptosis in autoimmune diseases, little is known about the relative importance of Fas-FasL and TNF in mediating apoptosis that is associated with clearance of inflammatory cells at a site of infection.

Resolution of infectious disease depends not only on elimination of pathogens, but also on the down-regulation of the inflammatory response that has been invoked during infection. While several cytokines that inhibit T cell activation or function, such as IL-10 and TGF-β, play important roles in down-regulating the immune response, apoptosis of lymphocytes that have accumulated at sites of infection may be another critical step in resolving disease. In the current study, we used two intracellular pathogens, Leishmania major and Rhodococcus equi, to investigate the role that TNF and Fas may play in lesion resolution. Leishmania is a protozoan parasite that causes a spectrum of diseases in man, including self-healing cutaneous lesions, chronic nonhealing cutaneous lesions, and fatal visceral infections. Leishmania infections also exhibit a spectrum of diseases in mice, which are in part associated with differential development of Th cell subsets (6, 7). Thus, C3H and C57BL/6 mice infected with L. major develop a strong Th1-type immune response and heal, while BALB/c mice develop a Th2 response and are unable to control the parasites. R. equi is a facultative intracellular bacterium that is associated with pulmonary disease in AIDS patients and young horses (8). In mice, the infection is normally associated with the development of a strong Th1 response and subsequent elimination of the bacteria, although depletion of IFN-γ leads to an uncontrolled, and eventually fatal, infection (9, 10).

Recently, we found that TNFRp55−/− and TNFRp55p75−/− mice inoculated with L. major were unable to resolve the inflammatory lesions in the footpad associated with infection, although they developed a Th1 response and cleared the parasites (11, 12). Here, we expand these observations by studying the course of R. equi infection following intratracheal inoculation in both TNFRp55−/− and B6-lpr mice, and compare our findings with the course of L. major infection in TNFRp55−/− and B6-lpr mice. Our results indicate that the absence of lesion resolution is not unique to the site of infection, or to one particular pathogen, and further suggest that TNF plays a unique role in the resolution of pathogen-induced lesions.

TNFRp55−/− (originally obtained from Klaus Pfeffer and Tak Mak, University of Toronto, Toronto, Canada) and wild-type mice were bred at the University of Pennsylvania and used at 4 to 5 wk of age (13). C57BL/6J and B6MRL-Faslpr mice, 4–5 wk old, were purchased from The Jackson Laboratory (Bar Harbor, ME). The animal colony was screened regularly for murine pathogens.

L. major (WHO MHOM/IL/Friedlin) was used for these studies. Parasites were grown in Grace’s insect cell culture medium (Life Technologies, Grand Island, NY) with 20% FBS (Hyclone, Logan, UT) in 2 mM glutamine. Stationary phase promastigotes were harvested and metacyclic stage parasites were selected using Arachis hypogae agglutinin (Sigma, St. Louis, MO) as described previously (14). Parasites were washed four times in PBS (Whittaker Bioproducts, Baltimore, MD). Mice were injected in the hind footpad with 2 × 106 metacyclic stage L. major parasites.

R. equi American Type Culture Collection (ATCC) 33701 is a virulent strain that possesses the 82-kb plasmid and expresses the 15- to 17-kDa proteins (vapA) associated with virulence in mice. Bacteria were maintained at −70°C prior to use. Mice were infected intratracheally with 2 × 107 bacteria in a total volume of 25 μl of PBS.

Wild-type mice were infected with R. equi as described (10). At 12 days postinfection, mice were treated i.p. with 3 mg of anti-TNF mAb XT22.11 (DNAX, Palo Alto, CA). Mice were also treated every 8 h with 25 mg/kg of vancomycin (Abbott Labs., N. Chicago, IL) from day 12 to day 21 postinfection. Some mice were treated only with vancomycin.

Cells were harvested from the lungs of R. equi-infected TNFRp55−/− and wild-type mice at 18 days postinfection. Following disruption in a tissue grinder, total lung cells from individual TNFRp55−/− mice, or three pooled wild-type mice, were cultured in complete tissue culture medium with 25 ng/ml of TNF for 1 to 3 days, or with 10 μg/ml of anti-Fas (Jo2) Ab for 3 to 5 h to induce apoptosis. Similarly, cells were harvested from the footpads of L. major-infected TNFRp55−/− and wild-type mice 6 wk postinfection. The footpads were removed, minced, and disrupted in a tissue grinder. Single cell suspensions were obtained by differential centrifugation and the cells were pooled from three feet from individual mice and were cultured in complete tissue culture medium with 25 ng/ml of TNF for 1 to 3 days, or with 10 μg/ml of anti-Fas Ab (Jo2) for 3 to 5 h to induce apoptosis. As a control, lymphocytes from uninfected mice were activated and exposed to anti-Fas Ab or TNF, as previously described (15). Cells from wild-type mice were induced to undergo apoptosis with both anti-Fas Ab and TNF, while cells from TNFRp55−/− failed to undergo apoptosis in response to TNF (data not shown).

Apoptosis of individual lymphocytes was demonstrated by detection of cell surface expression of phosphatidylserine using FITC-labeled annexin V (PharMingen, San Diego, CA) and flow cytometry (16). Briefly, following culture with TNF or anti-Fas mAb, cells were washed, incubated with 10 μl of FITC-labeled annexin V for 25 min at room temperature, and analyzed by flow cytometry. In addition, lymphocytes were also analyzed for apoptosis by TUNEL, which detects DNA strand breaks in cells undergoing apoptosis (17). Briefly, viable cells were quantitated by trypan blue exclusion and resuspended at 2 × 106 cells in 0.5 ml of 10 mM sodium phosphate, pH 7.2, 150 mM sodium chloride (PBS). The cell suspension was added to 5 ml of 1% (w/v) paraformaldehyde in PBS and placed on ice for 15 min. Following centrifugation, cells were washed and resuspended in 0.5 ml of PBS. Cells were added to 5 ml of ice-cold 70% (v/v) ethanol (for permeabilization) and stored at −20°C overnight. Cells were then washed in TdT buffer (0.5% M cacodylate, pH 6.8, 1 mM CoCl2, 0.5 mM DTT, 0.05% BSA, and 0.15% M NaCl), incubated for 60 min at 37°C with 2–4 M FITC-conjugated dUTP (PharMingen) and 5–10 U TdT in 25 μl TdT buffer. Cells were washed twice, resuspended in 1 ml of phosphatidylinositol/RNase solution (PharMingen), incubated for 30 min at room temperature, then analyzed in phosphatidylinositol/RNase solution by flow cytometry.

Apoptosis was detected in the footpad or lung lesions from mice infected with L. major or R. equi, respectively, as described (18). Briefly, formalin-fixed, paraffin-embedded sections of footpads or lungs were stained by the TUNEL technique (19) using a FITC-conjugated APO-TAG kit (Oncor, Gaithersburg, MD), according to the manufacturer’s recommendations. The number of apoptotic cells per 100 nuclei was determined in a blind fashion by another pathologist. For each mouse, 10 fields/slide were counted. Groups were analyzed statistically with a one-way ANOVA and a Dunnett’s post test.

We previously reported that L. major infection in TNFRp55−/− or TNFRp55p75−/− mice was associated with a nonhealing infection, but surprisingly that these animals were able to eliminate the parasites at the site of infection (11, 12). In order to determine if this phenotype was unique to either L. major or to the site of infection, we infected TNFRp55−/− mice intratracheally with the bacteria, R. equi. Although the peak numbers of R. equi recovered from TNFRp55−/− mice were significantly higher than in wild type, the bacterial numbers were reduced three logs by 18 days postinfection (Fig. 1). However, during the same period when bacterial numbers were decreasing, there was a progressive cellular infiltration into the lungs of TNFRp55−/− mice. Interestingly, in contrast to L. major infection in TNFRp55−/− mice, R. equi infected TNFRp55−/− mice began to die by 21 days of infection, presumably due to the massive infiltration of lymphocytes into the lungs (Fig. 1, also see Fig. 4). Lesions in R. equi-infected mice were assessed by recovery of cells from the lungs, which progressively increased during the course of infection. The cells infiltrating the lungs of R. equi-infected TNFRp55−/− mice were 25% CD4+ T cells, 24% γδ T cells, 5% CD8 T cells, and 30 to 40% macrophages as assessed by flow cytometry (data not shown). Thus, the course of R. equi infection in TNFRp55−/− mice exhibited some important similarities to that seen with L. major (11, 12). First, in both cases pathogen numbers were greater in the absence of the TNFRp55. Macrophage activation is the principal immune effector mechanism responsible for controlling these pathogens, and these results reflect the decreased efficiency of macrophage activation in the absence of TNF (12). More importantly, however, in both cases TNFRp55−/− mice were eventually able to control pathogen replication, but were unable to control the inflammatory response associated with the infections. From these results we hypothesized that the TNFRp55 was required for T cell apoptosis once the infection was cleared.

FIGURE 1.

Pathogen burden and lesion development in TNFRp55−/− and B6-lpr mice infected with R. equi or L. major. A, CFU recovered from the lungs of TNFRp55−/− and B6-lpr mice infected with R. equi. B, Total cells recovered from the lungs of R. equi-infected mice. * TNFRp55−/− mice died between days 18 and 21. C, Lesion size during the course of infection with L. major infection in TNFRp55−/− and B6-lpr mice. There are at least five mice per time point in each group. Significantly higher (p < 0.05) CFUs were measured at all time points in TNFRp55−/− mice compared with wild-type mice, and at 14 and 18 days postinfection compared with B6-lpr mice. Significantly higher (p < 0.05) cell numbers were recovered from lungs of TNFRp55−/− mice at 18 days postinfection. At 15 wk, all L. major-infected mice contained ≤102 parasites in their lesions.

FIGURE 1.

Pathogen burden and lesion development in TNFRp55−/− and B6-lpr mice infected with R. equi or L. major. A, CFU recovered from the lungs of TNFRp55−/− and B6-lpr mice infected with R. equi. B, Total cells recovered from the lungs of R. equi-infected mice. * TNFRp55−/− mice died between days 18 and 21. C, Lesion size during the course of infection with L. major infection in TNFRp55−/− and B6-lpr mice. There are at least five mice per time point in each group. Significantly higher (p < 0.05) CFUs were measured at all time points in TNFRp55−/− mice compared with wild-type mice, and at 14 and 18 days postinfection compared with B6-lpr mice. Significantly higher (p < 0.05) cell numbers were recovered from lungs of TNFRp55−/− mice at 18 days postinfection. At 15 wk, all L. major-infected mice contained ≤102 parasites in their lesions.

Close modal
FIGURE 4.

Pathology of lungs, 18 days postinfection, from mice infected intratracheally with R. equi. A, TNFRp55−/−. B, Anti-TNF-treated wild-type. C, B6-lpr. D, B6 wild-type. Lung sections were fixed in 10% buffered formalin, embedded in paraffin, and stained with hematoxylin and eosin. Bar = 125 μ.

FIGURE 4.

Pathology of lungs, 18 days postinfection, from mice infected intratracheally with R. equi. A, TNFRp55−/−. B, Anti-TNF-treated wild-type. C, B6-lpr. D, B6 wild-type. Lung sections were fixed in 10% buffered formalin, embedded in paraffin, and stained with hematoxylin and eosin. Bar = 125 μ.

Close modal

Because a major pathway associated with T cell apoptosis involves Fas-FasL interactions, we also investigated whether lpr mice, which are Fas deficient, would exhibit a phenotype similar to TNFRp55−/− mice. B6-lpr mice were infected in the footpad with L. major or intratracheally with R. equi. L. major-infected lpr mice developed lesions that were somewhat prolonged in their duration, but eventually the mice healed (Fig. 1). In addition, we also found that IL-12 treatment of L. major infected B6-lpr mice accelerated lesion resolution (data not shown). These results contrast with TNFRp55−/− mice, which fail to heal their lesions. When B6-lpr mice were infected with R. equi, they exhibited a course of infection similar to controls, demonstrating that the inflammatory response associated with this bacterial infection can be resolved without Fas-FasL. Since lpr mice exhibited a delay in healing a leishmanial infection, it is possible that Fas-FasL may play some protective role in leishmaniasis. In fact, recent studies with L. major-infected lpr mice indicated that they were unable to control parasites to the same extent as wild-type mice (20, 21, 22). Nevertheless, our results indicate that this effector mechanism is not required for parasite control or lesion resolution. While the reasons for the differences in the results with lpr mice are not known, it is important to mention that the reported inability of lpr mice to heal was related to an inability to clear parasites, which is quite different from the phenotype we see in TNFRp55−/− mice. Thus, it appears that TNFRp55, but not Fas, is required to eliminate chronic inflammatory lesions associated with these infections.

Following L. major infection in TNFRp55−/− mice, the parasite burden is reduced to that seen in wild-type mice by 10 to 15 wk (11). However, although TNFRp55 mice controlled R. equi replication after 3 days, R. equi levels remained relatively high (Fig. 1). To determine if the presence of large numbers of bacteria were required for the maintenance of the inflammatory response in the lung, we treated TNFRp55−/− mice with the antibiotic, vancomycin, starting on day 12 of infection. Antibiotic treatment resulted in a dramatic reduction of bacterial burden, and by 18 days, no bacteria could be cultured from the lung. Nevertheless, these mice continued to show an increase in the number of cells within the lungs (Fig. 2). Taken together with the results seen in TNFRp55−/− mice infected with L. major, where the parasite burdens are reduced to wild-type levels by 12 wk of infection (11, 12), these results argue that a high pathogen burden is not required to maintain the inflammatory response observed in mice lacking the TNFRp55.

FIGURE 2.

Bacterial burden and lesion development in R. equi-infected mice treated with antibiotics. Mice were treated every 8 h with 25 mg/kg of vancomycin (40 ) from day 12 to day 21 postinfection, and the bacterial burden (A) and number of cells infiltrating the lungs (B) assessed at various time points. There are at least five mice per time point in each group. There were significantly higher (p < 0.05) CFUs in TNFRp55−/− mice compared with antibiotic-treated TNFRp55−/− mice and wild-type mice at 18 days postinfection.

FIGURE 2.

Bacterial burden and lesion development in R. equi-infected mice treated with antibiotics. Mice were treated every 8 h with 25 mg/kg of vancomycin (40 ) from day 12 to day 21 postinfection, and the bacterial burden (A) and number of cells infiltrating the lungs (B) assessed at various time points. There are at least five mice per time point in each group. There were significantly higher (p < 0.05) CFUs in TNFRp55−/− mice compared with antibiotic-treated TNFRp55−/− mice and wild-type mice at 18 days postinfection.

Close modal

To determine if the phenotype observed in TNFRp55−/− mice could be recapitulated in wild-type mice when TNF was neutralized, we treated C57BL/6 mice with anti-TNF mAb starting at 12 days after R. equi infection and monitored the course of infection. Similar to TNFRp55−/− mice, C57BL/6 mice treated with anti-TNF mAb after the peak of the infection controlled the bacteria (Fig. 3). However, in contrast to control mice, mice treated with anti-TNF mAb maintained a cellular infiltrate in the lungs at 18 days. Histologically, lung lesions from R. equi-infected TNFRp55−/− mice were characterized by large numbers of lymphocytes and macrophages, which replaced and obliterated much of the pulmonary architecture (Fig. 4). Similarly, mice treated with anti-TNF exhibited large numbers of cells in the lung, although the inflammatory lesion was less severe than in TNFRp55−/− mice (Fig. 4). In contrast, B6-lpr and wild-type mice had minimal peribronchiolar and perivascular inflammation in the lungs at 18 days postinfection (Fig. 4). These results show that the phenotype observed in TNFRp55−/− mice can be mimicked in conventional animals by TNF depletion, directly implicating a role for TNF in lesion regression.

FIGURE 3.

Bacterial burden and lesion development in R. equi-infected mice treated with anti-TNF mAb. At 12 days postinfection, TNFRp55−/− mice were treated i.p. with 3 mg of anti-TNF mAb XT22.11 (DNAX, Palo Alto, CA), and the bacterial burden (A) and number of cells infiltrating the lungs (B) assessed at various time points. There are at least five mice per time point in each group. The CFUs and the number of cells in the lung were significantly higher (p < 0.05) in TNFRp55−/− and anti-TNF-treated mice, compared with wild-type mice at 18 days postinfection.

FIGURE 3.

Bacterial burden and lesion development in R. equi-infected mice treated with anti-TNF mAb. At 12 days postinfection, TNFRp55−/− mice were treated i.p. with 3 mg of anti-TNF mAb XT22.11 (DNAX, Palo Alto, CA), and the bacterial burden (A) and number of cells infiltrating the lungs (B) assessed at various time points. There are at least five mice per time point in each group. The CFUs and the number of cells in the lung were significantly higher (p < 0.05) in TNFRp55−/− and anti-TNF-treated mice, compared with wild-type mice at 18 days postinfection.

Close modal

Resting T cells do not undergo apoptotic death following treatment with anti-Fas or TNF, although following in vitro activation, T cells are sensitive to both pathways for apoptosis (15). To test whether cells from chronic lesions had been primed in vivo to undergo apoptosis following exposure to either anti-Fas or TNF, we harvested cells from either R. equi-infected lungs or L. major-infected footpads and measured Fas and TNF-mediated apoptosis without any additional stimulation. Cells harvested from R. equi-infected lungs of wild-type mice at 18 days postinfection, or from the footpads of L. major-infected wild-type mice at 6 wk postinfection, were sensitive only to TNF-mediated apoptosis (Fig. 5). In contrast, cells harvested from TNFRp55−/− mice, infected with either R. equi or L. major were not sensitive to either TNF or Fas-mediated apoptosis. On the other hand, cells from lymph nodes draining the site of either R. equi or L. major infection in wild-type mice were sensitive to both Fas and TNF, while cells from TNFRp55−/− mice were sensitive only to Fas-mediated apoptosis (data not shown). These results are consistent with the fact that lymph nodes in TNFRp55−/− were not enlarged relative to wild-type mice, and suggest that apoptosis may be regulated differently in the draining lymph node and at the site of inflammation.

FIGURE 5.

Induction of apoptosis by TNF or anti-FAS Ab of cells harvested from R. equi-infected lungs 18 days post-infection (A) or L. major-infected footpads 6 wk postinfection (B). Similar results were obtained when lymphocytes were analyzed for apoptosis by TUNEL. There were at least six mice per time point in each group, and error bars represent the SD of the mean. There were significantly higher (p < 0.05) numbers of apoptotic cells in TNF-treated wild-type cells compared with TNFRp55−/− cells.

FIGURE 5.

Induction of apoptosis by TNF or anti-FAS Ab of cells harvested from R. equi-infected lungs 18 days post-infection (A) or L. major-infected footpads 6 wk postinfection (B). Similar results were obtained when lymphocytes were analyzed for apoptosis by TUNEL. There were at least six mice per time point in each group, and error bars represent the SD of the mean. There were significantly higher (p < 0.05) numbers of apoptotic cells in TNF-treated wild-type cells compared with TNFRp55−/− cells.

Close modal

We next examined whether the persistent lesions seen in TNFRp55−/− mice infected with R. equi were associated with decreased numbers of apoptotic cells when compared with lesions from wild-type mice. During the course of infection with R. equi the number of apoptotic cells paralleled the inflammatory response to the pathogen in wild-type mice (Fig. 6). By 18 days, apoptotic cells were rare in lungs from TNFRp55−/− mice, but were clearly evident within the cuffs of lymphocytes around bronchioles and small blood vessels of wild-type mice. A quantitative analysis of these results indicated that there were significantly more apoptotic cells in the wild-type mice than in the TNFRp55−/− mice (Table I). By immunohistochemistry, most of the apoptotic cells in wild-type mice were CD3+ (data not shown). Finally, to determine if apoptosis also occurred during lesion resolution in the absence of Fas, apoptosis was evaluated in lungs of R. equi-infected B6-lpr mice (Fig. 6). Similar to wild-type mice, apoptotic cells were prominent in lymphocyte cuffs around bronchioles in lpr mice. The number of apoptotic cells were similar to those observed in wild-type mice (Table I). Thus, it appears that the cells from these inflammatory lesions become sensitized to undergo TNF-mediated apoptosis, and the p55 receptor is required to mediate the apoptotic signal.

FIGURE 6.

Demonstration of apoptosis in situ in lungs from mice 18 days after infection with R. equi. A, TNFRp55−/−. B, B6 wild type. C, B6-lpr. Formalin-fixed, paraffin-embedded sections of lungs were stained by the TUNEL technique (19 ) using a FITC-conjugated APO-TAG kit (Oncor), according to the manufacturer’s recommendations. Bar = 50 μ.

FIGURE 6.

Demonstration of apoptosis in situ in lungs from mice 18 days after infection with R. equi. A, TNFRp55−/−. B, B6 wild type. C, B6-lpr. Formalin-fixed, paraffin-embedded sections of lungs were stained by the TUNEL technique (19 ) using a FITC-conjugated APO-TAG kit (Oncor), according to the manufacturer’s recommendations. Bar = 50 μ.

Close modal
Table I.

Number of apoptotic cells/100 nucleia

Wild typelpr/lprTNFRp55−/−
Mean 20.78 17.30 6.93 
SD 2.65 1.59 1.53 
Wild typelpr/lprTNFRp55−/−
Mean 20.78 17.30 6.93 
SD 2.65 1.59 1.53 
a

Data represent the average of four mice with 10 fields examined per mouse. TNFRp55−/− compared with wild type, p < 0.01; lpr/lpr was not significantly different from wild type.

Our results demonstrate an absolute requirement for TNFRp55 to control the inflammatory response associated with two different pathogens, which occupy different anatomic locations. In the case of L. major, lack of lesion resolution leads to a chronic, but not fatal, infection. In contrast, following R. equi infection, the uncontrolled inflammatory response in the lung leads to death of the mice within 3 wk. We found that this phenotype was unique to TNFRp55−/− mice, and not observed in mice lacking Fas, one of the other major death receptors. Moreover, we correlated the inability to resolve pathogen-induced lesions in TNFRp55−/− mice with a lack of apoptosis at the site of infection. Finally, we were able to recapitulate this phenotype in wild-type mice by treating the animals with a neutralizing anti-TNF mAb.

The most straightforward interpretation of these results is that TNFRp55-dependent apoptosis is required to eliminate lymphocytes within inflammatory lesions. Activation-induced cell death (AICD) probably plays an important role in down-regulating the immune response after pathogens have been cleared, as well as playing a role in normal T cell development via deletion of self-reactive cells in the thymus. Recently, much effort has been directed at defining mechanisms of AICD, and from these studies a central role for Fas-FasL signaling in T cell deletion has emerged (1). However, several recent studies also suggest that in certain circumstances TNF may play a role in AICD of peripheral lymphocytes, which is consistent with the fact that the p55 TNF receptor shares a common apoptotic pathway with Fas (1). For example, it is known that lpr mice develop a lymphoproliferative disease after 10 to 12 wk of age, which is presumed to be due to defective Fas-dependent T cell deletion. However, when lpr mice were crossed with TNFRp55−/− mice the animals developed lymphoproliferative disease significantly earlier, demonstrating that TNF signaling can be important for cell deletion in the absence of Fas (5). Other studies have been done with lpr mice that have been crossed with an influenza hemagglutinin (HA) TCR transgenic mouse, producing an animal in which the influence of Fas on both thymic and peripheral Ag-specific T cell deletion can be assessed (23). Thus, injection of TCR-specific Ag into a TCR transgenic mouse leads to deletion of both thymic and peripheral T cells. Using the TCR transgenic/lpr mouse, it was found that injection of hemagglutinin could still induce apoptosis of peripheral T cells, suggesting that Fas was not required for T cell AICD. However, when animals were treated with a neutralizing anti-TNF mAb, apoptosis of T cells was abrogated, suggesting that TNF contributes to T cell AICD. It should be pointed out, however, that in animals that have Fas, anti-TNF mAb did not abrogate apoptosis, and that in other TCR transgenic/lpr mice a defect in apoptosis was seen without injecting anti-TNF mAb (24). Nevertheless, these results argue that, under certain circumstances, TNF may contribute to peripheral T cell deletion. Our data suggest that one of these special circumstances may involve pathogen-induced inflammation.

While the experiments described above implicate TNFRp55 as a mediator of AICD, other studies have suggested that the p75 receptor delivers the apoptotic signal in lymphocytes. One study demonstrated that anti-CD3-induced T cell apoptosis was blocked by inhibiting TNF, and that the receptor involved in this apoptotic pathway was the p75 receptor (25). However, in leishmaniasis the p75 receptor is probably not required, since our studies found no defect in the resolution of leishmanial-induced lesions in TNFRp75−/− mice (12).

While TNF signaling through TNFRp55 is not absolutely required to control replication of R. equi or L. major, infections in TNFRp55−/− mice with Listeria monocytogenes or Mycobacterium tuberculosis, and in TNFRp55p75−/− mice with Toxoplasma, are fatal (13, 26, 27, 28). Differences in the ability of TNFRp55−/− mice to control these pathogens may be due to the greater virulence of Listeria and M. tuberculosis in mice compared with R. equi, which is avirulent in immunocompetent mice infected intratracheally, and Leishmania, which is controlled in mice on the C57BL/6 background. In the case of Toxoplasma, it appears that a TNF-dependent, inducible nitric oxide synthase independent effector mechanism may be lacking in TNFR-deficient mice (26). Therefore, by infecting with pathogens that are typically controlled by normal mice, we were able to observe the defect in lesion resolution in TNFRp55−/− mice. Recent studies in TNF-deficient mice support this hypothesis. Thus, i.p. injection of heat-killed Corynebacterium parvum led to an uncontrolled, and eventually fatal, inflammatory response (29). Similarly, we found that i.p. injection with the slow-growing bacterium, Mycobacterium paratuberculosis, leads to an uncontrolled inflammatory response in the peritoneal cavity of TNFRp55−/− mice (P. Chilton and P. Scott, unpublished observations). Thus, the phenotype we see in TNFRp55−/− mice may be masked by an uncontrolled infection in mice infected with more virulent pathogens, or when TNF is required for a protective immune effector mechanism.

TNF plays a central role in macrophage activation, and hence it is not surprising that we observed an increase in the pathogen load in mice lacking TNFRp55 (11). Our results are similar to those previously reported, in which it was found that neutralization of TNF with mAbs, or infection of transgenic mice expressing TNFRp55 fusion protein, enhanced susceptibility to L. major (30, 31, 32, 33). However, the ability of TNFRp55−/− mice to eventually control L. major parasites suggests that other mechanisms can participate in parasite elimination. At present, we believe these are inducible nitric oxide synthase dependent, and current studies in our laboratory suggest that a T cell-dependent mechanism of macrophage activation may compensate for the absence of TNF in these knockout mice.

In our studies, we found that B6-lpr mice are able to heal following L. major (FN strain) infection, although they were slightly delayed in healing compared with control mice. These results contrast with reports that B6-lpr mice and MRL/lpr mice are susceptible to L. major (LV39) (20, 21, 22). We excluded the possibility that these differences were due to the parasite strain used, since we found that L. major (LV39)-infected B6-lpr mice were also able to heal (data not shown). In addition, B6-lpr mice were also found to be resistant to another L. major strain (NIH 173) (Dr. S. Reiner, personal communication). In contrast to B6-lpr mice, however, we found that MRL/lpr mice were susceptible to L. major (data not shown). We believe this susceptibility is not directly related to the absence of Fas-induced apoptosis, since MRL/lpr mice developed a severe lymphoproliferative disease and immunosuppression during infection. It is for this reason that we studied B6-lpr mice, which do not develop lymphoproliferative disease until later in life. Thus, the discrepancy in the results with B6-lpr mice remain unexplained, but since we found that the B6-lpr mice exhibited a delay in healing, our data might also be interpreted as an indication that Fas-FasL plays some role, albeit minor, in lesion resolution.

The question these results raise is what factors influence whether TNF signaling will lead to proinflammatory responses or to apoptosis of lymphocytes during infections with intracellular pathogens. Signaling via the p55 receptor can either activate NF-κB or apoptosis (34). It is likely that both the cytokine milieu and the activation status of the cell influence whether TNF activates NF-κB or the apoptotic cascade. For example, it has been shown that TNF-induced apoptosis is suppressed by NF-κB activation (35, 36), suggesting the possibility that apoptosis of inflammatory cells in vivo may be preceded by decreased levels of NF-κB inducing cytokines at the site of infection. Indeed, an alternative explanation for our results is that TNF plays an inhibitory role in cytokine production, and that TNFRp55−/−-deficient mice exhibit an uncontrolled inflammatory response due to maintenance of cytokine production. For example, a recent study suggests that TNF inhibits IL-12 production (37), although in other studies TNF appears to augment IL-12 production (38, 39). In TNFRp55−/− mice, there is no defect in IFN-γ or IL-12 production, at least indicating that TNF is not absolutely required for the production of these cytokines (11, 12, 26).

Our results define a previously unrecognized role for TNF in regulating pathogen-induced inflammation, and provides an impetus to define the factors that protect cells from TNF-mediated apoptosis during an infection. One example in which these findings may be applicable is in leishmaniasis, in which some forms of disease, such as mucocutaneous leishmaniasis and leishmaniasis recidivans, are associated with strong cell-mediated immunity, elimination of most of the parasites associated with the infection, but maintenance of lesions. These studies expand our understanding of the factors involved in down-regulating pathogen-induced inflammation, which is the first step in designing new strategies to control such naturally occurring persistent inflammatory lesions.

We thank Klaus Pfeffer and Tak Mak (University of Toronto) for providing TNFRp55−/− mice, Jay Farrell and Chris Hunter (Department of Pathobiology, University of Pennsylvania) for critical reading of the manuscript, and Leslie Taylor for technical assistance.

1

This work was supported by National Institutes of Health Grants AI 01233 and AI 41880.

3

Abbreviations used in this paper: FasL, Fas ligand; AICD, activation-induced cell death.

1
Nagata, S..
1997
. Apoptosis by death factor.
Cell
88
:
355
2
Lenardo, M. J..
1996
. Fas and the art of lymphocyte maintenance.
J. Exp. Med.
183
:
721
3
Watanabe-Fukunaga, R., C. I. Brannan, N. G. Copeland, N. A. Jenkins, S. Nagata.
1992
. Lymphoproliferation disorder in mice explained by defects in Fas antigen that mediates apoptosis.
Nature
356
:
314
4
Takahashi, T., M. Tanaka, C. I. Brannan, N. A. Jenkins, N. G. Copeland, T. Suda, S. Nagata.
1994
. Generalized lymphoproliferative disease in mice, caused by a point mutation in the Fas ligand.
Cell
76
:
969
5
Zhou, T., C. K. Edwards, P. Yang, Z. Wang, H. Bluethmann, J. D. Mountz.
1996
. Greatly accelerated lymphadenopathy and autoimmune disease in lpr mice lacking tumor necrosis factor receptor I.
J. Immunol.
156
:
2661
6
Reiner, S. L., R. M. Locksley.
1995
. The regulation of immunity to Leishmania major.
Annu. Rev. Immunol.
13
:
151
7
Scott, P..
1996
. T helper cell development and regulation in experimental cutaneous leishmaniasis.
Chem. Immunol.
63
:
98
8
Prescott, J. F., A. M. Hoffman.
1993
.
Rhodococcus equi. Vet. Clin. North Am.—Equine Pract.
9
:
375
9
Hines, S. A., S. T. Kanaly, B. A. Byrne, G. Palmer.
1997
. Immunity to Rhodococcus equi.
Vet. Microbiol.
56
:
177
10
Kanaly, S. T., S. A. Hines, G. H. Palmer.
1995
. Cytokine modulation alters pulmonary clearance of Rhodococcus equi and development of granulomatous pneumonia.
Infect. Immun.
63
:
3037
11
Vieira, L. Q., M. Goldschmidt, M. Nashleanas, K. Pfeffer, T. Mak, P. Scott.
1996
. Mice lacking the TNF receptor p55 fail to resolve lesions caused by infection with Leishmania major, but control parasite replication.
J. Immunol.
157
:
827
12
Nashleanas, M., S. Kanaly, P. Scott.
1998
. Control of Leishmania major infection in mice lacking TNF receptors.
J. Immunol.
160
:
5506
13
Pfeffer, K., T. Matsuyama, T. M. Kundig, A. Wakeham, K. Kishihara, A. Shahinian, K. Weignmann, P. S. Ohashi, M. Kronke, T. W. Mak.
1993
. Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection.
Cell
73
:
457
14
Sacks, D. H., P. V. Perkins.
1984
. Identification of an infective stage of Leishmania promastigotes.
Science
223
:
1417
15
Lenardo, M. J..
1991
. Interleukin-2 programs mouse αβ T lymphocytes for apoptosis.
Nature
353
:
858
16
Vermes, I., C. Haanen, H. Steffens-Nakken, C. Reutelingsperger.
1995
. A novel assay for apoptosis: flow cytometric detection of phosphatidylserine expression on early apoptotic cells using fluorescein labelled Annexin V.
J. Immunol. Methods
184
:
39
17
Surh, C. D., J. Sprent.
1994
. T-cell apoptosis detected in situ during positive and negative selection in the thymus.
Nature
372
:
100
18
Milik, A. M., V. A. Buechner-Maxwell, J. Sonstein, S. Kim, G. D. Seitzman, T. F. Beals, J. L. Curtis.
1997
. Lung lymphocyte elimination by apoptosis in the murine response to intratracheal particulate antigen.
J. Clin. Invest.
99
:
1082
19
Gavrieli, Y., Y. Sherman, S. A. Ben-Sasson.
1992
. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation.
J. Cell Biol.
119
:
493
20
Boyer, M. H., C. L. Jaffe.
1984
. Increased susceptibility to Leishmania tropica infection in autoimmune MRL/MpJ-lpr mice.
J. Parasitol.
70
:
145
21
Conceicao-Silva, F., M. Hahne, M. Schroter, J. Louis, J. Tschoop.
1998
. The resolution of lesions induced by Leishmania major in mice requires a functional Fas (APo-1, CD95) pathway of cytotoxicity.
Eur. J. Immunol.
28
:
237
22
Huang, F. P., D. Xu, E. O. Esfandiari, W. Sands, X. Q. Wei, F. Y. Liew.
1998
. Mice defective in Fas are highly susceptible to Leishmania major infection despite elevated IL-12 synthesis, strong Th1 responses, and enhanced nitric oxide production.
J. Immunol.
160
:
4143
23
Sytwu, H. K., R. S. Liblau, H. O. McDevitt.
1996
. The roles of Fas/Apo-1 (CD95) and TNF in antigen-induced programmed cell death in T cell receptor transgenic mice.
Immunity
5
:
17
24
Singer, G. G., A. K. Abbas.
1994
. The fas antigen in involved in peripheral but not thymic deletion of T lymphocytes in T cell receptor transgenic mice.
Immunity
1
:
365
25
Zheng, L., G. Fisher, R. E. Miller, J. Peschon, D. H. Lynch, M. J. Lenardo.
1995
. Induction of apoptosis in mature T cells by tumour necrosis factor.
Nature
377
:
348
26
Yap, G. S., T. Scharton-Kersten, H. Charest, A. Sher.
1998
. Decreased resistance to TNF receptor p55- and p75-deficient mice to chronic toxoplasmosis despite normal activation of inducible nitric oxide synthase in vivo.
J. Immunol.
160
:
1340
27
Rothe, J., W. Lesslauer, H. Loetscher, Y. Lang, P. Koebel, F. Koentgen, A. Althage, R. Zinkernagel, M. Steinmetz, H. Bluethmann.
1993
. Mice lacking the tumor necrosis receptor 1 are resistant to TNF-mediated toxicity but highly susceptible to infection by Listeria monocytogenes.
Nature
364
:
798
28
Flynn, J. L., M. M. Goldstein, J. Chan, K. J. Triebold, K. Pfeffer, C. J. Lowenstein, R. Schreiber, T. W. Mak, B. R. Bloom.
1995
. Tumor necrosis factor-α is required for the protective immune response against Mycobacterium tuberculosis in mice.
Immunity
2
:
561
29
Marino, M. W., A. Dunn, D. Grail, M. Inglese, Y. Noguchi, E. Richards, A. Jungbluth, H. Wada, M. Moore, B. Williamson, S. Basu, L. J. Old.
1997
. Characterization of tumor necrosis factor-deficient mice.
Proc. Natl. Acad. Sci. USA
94
:
8093
30
Titus, R. G., B. Sherry, A. Cerami.
1989
. Tumor necrosis factor plays a protective role in experimental murine cutaneous leishmaniasis.
J. Exp. Med.
170
:
2097
31
Kossodo, S., G. E. Grau, J. A. Louis, I. Muller.
1994
. Tumor necrosis factor α (TNF-α) and TNF-β and their receptors in experimental cutaneous leishmaniasis.
Infect. Immun.
62
:
1414
32
Liew, F. Y., Y. Li, S. Millott.
1990
. Tumour necrosis factor (TNF-α) in leishmaniasis. II. TNF-α-induced macrophage leishmanicidal activity is mediated by nitric oxide from l-arginine.
Immunology
71
:
556
33
Garcia, I., Y. Miyazaki, K. Araki, M. Araki, R. Lucas, G. E. Grau, G. Milon, Y. Belkaid, C. Montixi, W. Lesslauer, P. Vassalli.
1995
. Transgenic mice expressing high levels of soluble TNF-R1 fusion protein are protected from lethal septic shock and cerebral malaria, and are highly sensitive to Listeria monocytogenes and Leishmania major infections.
Eur. J. Immunol.
25
:
2401
34
Hsu, H., H.-B. Shu, M.-G. Pan, D. V. Goeddel.
1996
. TRADD-TRAF2 and TRADD-FADD interactions define two distinct TNF receptor 1 signal transduction pathways.
Cell
84
:
299
35
Beg, A. A., D. Baltimore.
1996
. An essential role for NF-κB in preventing TNF-α-induced cell death.
Science
274
:
782
36
Liew, F. Y., Y. Li, A. Severn, S. Millot, J. Schmidt, M. Salter, S. Moncada.
1991
. A possible novel pathway of regulation by murine T helper type-2 (Th2) cells of a Th1 cell activity via the modulation of the induction of nitric oxide synthase on macrophages.
Eur. J. Immunol.
22
:
1143
37
Hodge-Dufour, J., M. W. Marino, M. R. Horton, A. Jungbluth, M. D. Burdick, R. M. Strieter, P. W. Noble, C. A. Hunter, E. Pure.
1998
. Inhibition of interferon γ induced interleukin 12 production: a potential mechanism for the anti-inflammatory activities of tumor necrosis factor.
Proc. Natl. Acad. Sci.
95
:
13806
38
Flesch, I. E. A., J. H. Hess, S. Huang, M. Aguet, J. Rothe, H. Bluethmann, S. H. E. Kaufmann.
1995
. Early interleukin 12 production by macrophages in response to mycobacterial infection depends on interferon γ and tumor necrosis factor α.
J. Exp. Med.
181
:
1615
39
Becher, B., M. Blain, P. S. Giacomini, J. P. Antel.
1999
. Inhibition of Th1 polarization by soluble TNF receptor is dependent on antigen-presenting cell-derived IL-12.
J. Immunol.
162
:
684
40
Nordmann, P., J. J. Kerestedjian, E. Ronco.
1992
. Therapy of Rhodococcus equi disseminated infections in nude mice.
Antimicrob. Agents Chemother.
36
:
1244