C57BL/6 mice exposed to the radiation-attenuated schistosome vaccine exhibit high levels of protective immunity. The cell-mediated pulmonary effector mechanism involves IFN-γ-producing CD4+ T cells in a focal response around challenge larvae. IFN-γ can promote production of TNF and can synergize with this cytokine in its actions on responder cells. We have examined whether TNF plays a role in lung phase immunity to schistosomes using mice with a disrupted gene for TNFRI (TNFRI−/−). The most dramatic finding was that the schistosome vaccine elicited no protection whatsoever in these mice. However, this could not be attributed to a lack of responder cells, because more lymphocytes were lavaged from the airways of TNFRI−/− than wild-type mice. Furthermore, CD4+ T cells were equally represented in airway populations from the two groups and produced IFN-γ upon Ag stimulation in vitro. In contrast, pulmonary macrophage function was defective in TNFRI−/− mice, as indicated by a failure to up-regulate inducible NO synthase mRNA. Histopathological analysis revealed that focal infiltrates were of similar size and cell composition in the two groups but that more parasites were free of foci in the TNFRI−/− mice. These animals had a greatly impaired IgG response to schistosomes, which may explain their lack of residual protection due to Ab in a situation where cell-mediated immunity is disabled. We suggest that the absence of protective immunity could result from a retarded build-up of leukocytes around migrating lung worms and/or a deficit in accessory cell function within a focus, both of which would permit parasite escape.
Ahigh level of protective immunity to Schistosoma mansoni can be induced by a single exposure of C57BL/6 mice to radiation-attenuated cercariae (1). Normal challenge larvae provoke a focal inflammatory response in the lungs of vaccinated animals that leads to their elimination. This response is dependent on CD4+ T cells with Th1 characteristics that produce abundant IFN-γ upon antigenic stimulation in vitro. Administration of cytokine-neutralizing Ab to vaccinated mice after challenge results in 90% abrogation of immunity, revealing the crucial role of IFN-γ in the effector response (2). Furthermore, vaccination of mice with a disrupted IFN-γ receptor gene (IFN-γR−/−) elicits only low levels of protection (3), most probably mediated by Ab (P. S. Coulson, unpublished data).
IFN-γ has a well-characterized role in the priming of macrophages/monocytes to produce TNF in response to LPS stimulation (4). These two cytokines can act synergistically in host responses to infectious agents and are also involved in leukocyte recruitment to sites of inflammation. A possible link between IFN-γ and TNF in the schistosome vaccine model is provided by the observation of a lower overall level of TNF-α mRNA in the lungs of poorly protected IFN-γR−/− mice after challenge compared with highly protected C57BL/6 mice (3). Direct killing of schistosome larvae by cytotoxic agents such as NO (3) is one suggested effector mechanism in which both IFN-γ and TNF-α could participate. The two cytokines synergize in the production of inducible NO synthase (iNOS)3 and NO will mediate cytotoxic killing of parasites in vitro (5). Although increased iNOS mRNA has been demonstrated in the lungs of C57BL/6 mice after challenge (6, 7), a significant level of protection is elicited by vaccination of mice with a disrupted iNOS gene (7, 8), suggesting that NO is not a major factor in challenge parasite elimination. An alternative effector mechanism, in which both IFN-γ and TNF could participate, might involve the aggregation of leukocytes into a tight focus that simply blocks parasite migration through the vascular bed of the lungs. However, the adhesive interactions so far examined between pulmonary leukocytes (ICAM-1/LFA-1 and CD2/CD48) do not appear to influence the cohesiveness of effector foci or be important for protection (9).
The biological activities of TNF are mediated by two structurally related, but functionally distinct receptors, TNFRI (p55) and TNFRII (p75), which are coexpressed on most cell types. The derivation of mice with disrupted genes for the two receptors has enabled their respective functions in vivo to be dissected (10, 11, 12). The proinflammatory activities of TNF-α are mediated primarily via TNFRI (10, 12). In contrast, the role of TNFRII is less well defined, but it appears to be involved in lymphocyte proliferation and the induction of apoptosis by a novel pathway (4). We have investigated the involvement of TNF in the pulmonary inflammatory response to schistosome larvae using mice with a disrupted TNFRI gene. When these animals were vaccinated with attenuated parasites, we observed a complete absence of protective immunity. This could not be attributed to a lack of responder T cells in the lungs after challenge, or a switch to Th2 cytokine production. Although effector foci formed in TNFRI−/− mice, more parasites were free of surrounding infiltrates than in comparable wild-type (WT) animals, and iNOS mRNA was not up-regulated, suggesting that defects in macrophage function might explain the lack of protection.
Materials and Methods
Mice of the 129 strain with a targeted disruption of the gene for the 55-kDa TNFR (TNFRI−/−; Ref. 10) were obtained from Hoffmann-LaRoche (Basel, Switzerland) and backcrossed six times to the C57BL/6 background. Homozygous mutants and WT mice from the sixth cross were bred in isolators at the University of York and maintained in laminar flow cabinets.
Parasite exposure regimens and experimental design
Mice were exposed to a Puerto Rican isolate of S. mansoni, routinely maintained in albino Biomphalaria glabrata and outbred MF1 mice. Groups of test mice were vaccinated with 500 radiation-attenuated cercariae via the shaved abdomen (13). They were challenged 5 wk later, together with control groups of age-matched naive mice, by tail skin exposure to 200 normal cercariae. The level of protection was determined 5 wk later by a comparison of adult worm burdens in test and control groups. Pulmonary immune responses were investigated in additional groups of vaccinated mice at 14 days postchallenge, the time of peak inflammation in the lungs (14).
Analysis of pulmonary leukocytes, cytokines, and iNOS expression
Airway leukocytes were recovered from the perfused lungs by bronchoalveolar lavage (BAL) and characterized as previously described (13). Individual BAL cell suspensions were adjusted to 1 ml in Glasgow MEM, containing 10% FCS (Globefarm, Esher, Surrey, U.K.), 200 U/ml penicillin, and 100 μg/ml streptomycin (Sigma, St. Louis, MO). Total cell number was estimated by hemocytometer, and analyses were performed on aliquots of fresh cells using an EPICS XL flow cytometer (Coulter Electronics, Hialeah, FL). CD4+ T lymphocytes were phenotyped using fluorochrome-conjugated rat mAb (clone YTS 191.1; Caltag, Burlingame, CA).
Ag-stimulated production of cytokines by the remaining BAL leukocytes was measured after culture in 96-well plates (Nunclon, Life Technologies, Paisley, Scotland) with or without 50 μg/ml soluble 18-h schistosomular Ag (SSP; Ref. 15), for 72 h at 37°C, 5% CO2/95% air. The amounts of IFN-γ and IL-4 in culture supernatants were measured by double Ab ELISAs as described in detail elsewhere (3, 14).
RT-PCR analysis of total RNA extracted from whole lung was conducted for iNOS and the housekeeping gene hypoxanthine-guanine phosphoribosyl transferase, using the sense and antisense primers previously described (7). Semiquantitative measurement of enzyme PCR product was performed on slot-blots by hybridizating end-labeled oligonucleotide probes, and bound radioactivity was estimated using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Lungs were excised 14 days postchallenge and processed as previously described (3). A total of 140 consecutive sections was scanned from a single lobe for each of five vaccinated and challenged TNFRI−/− and WT mice (1400 sections in all). The number of parasites in each sequence of the sections was recorded and any association with a focus was noted. The cellular composition of foci was determined, and each was photographed in the section displaying the largest profile for estimation of maximum dimension.
Serum Ab assays
Serum samples were collected on the day before challenge and again at day 14 postchallenge. The levels of Ag-specific IgG isotypes were determined by ELISA according to a previously described protocol (16). Briefly, microtiter plates were coated overnight at 4°C with 2.5 μg/ml 18 h SSP in carbonate buffer (pH 9.6). Plates were washed and probed with 1:150 dilution of the test sera for 1.5 h. The plates were then probed with HRP-conjugated rabbit Abs to mouse IgG and IgG1, IgG2a, IgG2b, or IgG3 isotypes (Zymed, San Francisco, CA). Ab binding was assessed following addition of peroxidase substrate (Sigma) and measurement of color development at 490 nm.
Sample groups were composed of five mice at all times in all experiments, and results are expressed as the mean ± SE unless the samples were pooled. Data were analyzed for statistical significance using Student’s t test.
Protective immunity is not induced by vaccination of TNFRI−/− mice
To determine whether TNF is required for the protective response induced by the radiation-attenuated schistosome vaccine, a series of vaccination and challenge experiments was performed using mice genetically deficient in TNFRI. Exposure of WT mice to attenuated cercariae resulted in the anticipated significant reduction in challenge worm burden relative to control animals (p < 0.02; Fig. 1); this amounted to a mean of 53.4% protection in three experiments. In striking contrast, no significant reduction in worm burden was observed after challenge of vaccinated TNFRI−/− mice compared with each control group (p > 0.25); the mean protection of the three vaccinated groups was −1.2%.
More leukocytes are recovered from airways of TNFRI−/− than WT mice
Vaccination of mice with attenuated cercariae results in the recruitment of leukocytes to the pulmonary parenchyma and airways, culminating in a persistent CD4+ T cell population that participates in the effector response (1). Because TNF is considered an important cytokine in the tissue recruitment of leukocytes, we examined whether there was a reduced cell infiltration into the lungs of TNFRI−/− mice, which might explain their lack of immunity. When airway leukocytes were recovered at day 14 postchallenge by BAL for flow cytometric analysis, the total cell yield from TNFRI−/− mice was double that from their WT counterparts (Fig. 2). In both groups of mice, macrophages and lymphocytes were the dominant cell population, with granulocytes (eosinophils) a minor component. The cell increment in TNFRI−/− mice was evenly distributed between the three leukocyte subpopulations, showing that the overall increase was not due to differential recruitment. CD4+ T cells represented 64% of total lymphocytes in the TNFRI−/− mice vs 67% in the WT animals.
BAL lymphocytes show a dominant Th1 cytokine profile after vaccination and challenge
In other gene-disrupted mice (e.g., IFN-γR−/−), where we have observed low levels of protection after exposure to the radiation-attenuated vaccine, there was a marked shift to a Th2 cytokine profile in the Ag-stimulated BAL leukocyte cultures. We therefore compared the cytokines produced by BAL cells following recovery from WT and TNFRI−/− mice 14 days after challenge. In cultures from WT mice, Ag-stimulated IFN-γ production greatly exceeded that of IL-4 (Fig. 3); there was also a low level of endogenous IFN-γ release in the absence of added Ag. Cultures from TNFRI−/− mice showed, if anything, a stronger Th1 bias, with higher IFN-γ and lower IL-4 levels than WT animals. Indeed, when cytokine production was expressed per 105 lymphocytes in culture to take into account the differences in cell content of the BAL fluid from the two groups, the pattern of Th1 dominance was reinforced. Thus, the ratio of IFN-γ:IL-4 in culture supernatants from WT mice was 4.9:1, whereas that from TNFRI−/− mice was 14.5:1.
Pulmonary effector foci develop in the lungs after challenge
Challenge parasites in the lungs of vaccinated C57BL/6 mice attract a focal inflammation that brings about their elimination (1). We therefore undertook a histopathological study of lungs from TNFRI−/− and WT mice at day 14 postchallenge to determine whether focus formation was impaired. Fewer foci were detected in TNFRI−/− than WT mice, and the mean maximum dimension of each focus was also smaller in the former animals (Table I). In contrast, more parasites were found in TNFRI−/− than WT mice but fewer of these were actually associated with foci in the former animals. The pulmonary cellular infiltrates in both groups of mice had a predominantly mononuclear cell composition, indicative of a Th1-biased response (data not shown).
|.||WT .||TNFRI−/− .|
|Mean maximum dimension (μm)||255± 12.6||219± 13.7|
|% Parasites associated with a focus||82||45|
|.||WT .||TNFRI−/− .|
|Mean maximum dimension (μm)||255± 12.6||219± 13.7|
|% Parasites associated with a focus||82||45|
A sequence of 140 serial sections was examined from each of five WT and TNFRI−/− mice to avoid double counting of individual parasites and to identify the point where a focus reached maximum dimensions.
Pulmonary macrophage function is impaired after challenge
Because T cell function is apparently normal in TNFRI−/− mice, the lesion responsible for the lack of protection may be present in the accessory cells involved in the effector response. Elevated production of NO is a feature of pulmonary macrophage activation in C57BL/6 mice after vaccination and challenge (7). Furthermore, TNF-mediated synergism with IFN-γ for NO production by macrophages is thought to act via TNFRI signaling (19). We therefore measured the level of iNOS mRNA in whole lungs as an indicator of the functional state of macrophages after challenge. In WT animals, there was a significant increase (2.5-fold) in iNOS mRNA at day 14 compared with the naive level (Fig. 4). In contrast, although iNOS mRNA expression was evident in naive TNFRI−/− mice, no iNOS mRNA up-regulation was observed in mice at day 14 postchallenge.
Primary Ab response is impaired and no secondary response is observed after challenge
In C57BL/6 mice after a single exposure to the attenuated schistosome vaccine, the pulmonary effector mechanism is dependent on CD4+ T cells, with little evidence for the involvement of B cells (17). However, TNFRI−/− mice have defects in the differentiation of follicular dendritic cell networks and the development of germinal centers in lymphoid tissue, which affect Ab responses (18). A plausible explanation for the lack of protection against schistosomes in such mice is that Ab is an unsuspected and essential component of the effector mechanism. We compared the levels of specific Ab in the serum of vaccinated WT and TNFRI−/− mice immediately before challenge and at the height of the effector response 14 days later. A significant primary IgG response was observed in WT animals following vaccination, with IgG1, IgG2a, and IgG2b production (Fig. 5) but no IgG3 production (data not shown); the total specific IgG and the levels of the three IgG isotypes were boosted by challenge exposure. In marked contrast, the total specific IgG level in TNFRI−/− mice at challenge was approximately one-third that of WT counterparts, and the same was true for all three isotypes. Furthermore, exposure of TNFRI−/− mice to a normal parasite challenge had no immunostimulatory effect. Indeed, with the exception of IgG2a, the Ab levels were lower than after the primary exposure.
C57BL/6 mice given a single exposure to the radiation-attenuated schistosome vaccine show a 60–70% reduction in challenge worm burden compared with naive controls. Although the protective immune mechanism that operates in the lungs is directed against a large extracellular parasite (120 μm long by 30 μm diameter in the contracted state; Ref. 20), it has the characteristics of a CD4+ T cell-mediated, delayed-type hypersensitivity response. In this respect it has more in common with responses to intracellular protozoa such as Leishmania major and Toxoplasma gondii (21) than nematode helminths, in which protective immunity is Th2 mediated (22). The role of IFN-γ in protection against schistosomes was first highlighted by in vivo cytokine neutralization experiments (2, 23). We have now demonstrated the requirement for a second cytokine, TNF. Given the involvement of TNF in the IFN-γ-mediated immunity both to pathogenic bacteria (24) and protozoa (25), this result is perhaps not surprising. However, what is unusual is the absolute nature of the requirement, because protection was completely absent when TNFRI−/− mice were vaccinated and challenged. Such a complete loss of function rarely occurs when gene-disrupted mice are used to probe immune responses because of the degree of redundancy among immune effector mechanisms (cf. vaccination of IFN-γR−/− and IL-12 −/− mice; Refs. 3 and 26). Protection is also absent in mice deprived of both the p55 and p75 receptors for TNF (mean of 0% protection in two experiments; our unpublished data), but signaling through TNFRI is clearly the key.
It is possible that the lack of protection in the absence of TNFRI signaling is due to TNF action at a site distant from the lungs. The role of TNF in the development of secondary lymphoid organs (reviewed in Ref. 27) would come into this category. Thus, primary specific Ab responses after exposure to attenuated schistosomes were poor in TNFRI−/− mice compared with controls. Furthermore, secondary responses to challenge larvae were negligible. This could be taken as evidence for a hitherto unsuspected role for Ab in the pulmonary effector response of mice that were vaccinated once. Two recent studies have addressed this question by vaccination and challenge of μMT mice lacking functional B cells (17, 28). Both studies reported that significant protection was induced in these mice, although, overall, somewhat less than in compatible WTs. In one study (28), restitution of the immunity to the WT level was achieved by administration of immune serum to vaccinated μMT mice, a result that implies a role for Ab- and cell-mediated mechanisms in WT animals that were vaccinated once.
In addition, we need to consider a local role for TNF in the effector mechanism. The cytokine has been implicated in the up-regulation of adhesion molecule expression on vascular endothelial cells and the production of several chemokines at sites of delayed-type hypersensitivity reactions (29). Thus, in the absence of signaling via TNFRI, we might anticipate reduced cell infiltration into the lungs. However, analysis of airway leukocyte populations after challenge revealed elevated cell numbers rather than impaired recruitment. Clearly, the failure to eliminate challenge parasites in the lungs of vaccinated TNFRI−/− mice is not due to a reduced frequency of responder cells. This is emphasized by the high level of IFN-γ and low level of IL-4 in Ag-stimulated BAL cultures from both WT and TNFRI−/− mice. The Th1 bias of the elicited immune response is thus maintained in the absence of signaling via TNFRI, in contrast to the marked switch to a Th2 cytokine profile, which we have previously observed in the lungs of IFN-γR−/− and IL-12−/− mice (3, 26). (This Th1 bias in the response of TNFRI−/− mice coupled with their defective germinal centers also means that, unlike the foregoing gene-disrupted mice, they fail to mount the alternative Th2 response that leads to a degree of protection mediated by Ab.)
The elevation in leukocyte numbers we observed in the lungs is similar to the situation reported following infection with L. major (30). Here, nonhealing skin lesions developed in TNFRI−/− mice, with marked lymphocyte infiltration of the epidermis and severe edema, in the absence of parasite-containing macrophages. It was subsequently suggested that the failure of such lesions to resolve was due to the absence of lymphocyte apoptosis, normally signaled via TNFRI (31). By analogy, the elevated lymphocyte numbers in the lungs, reported in the present study, could reflect diminished apoptosis and cell clearance.
Elimination of challenge parasites in the lungs of vaccinated WT mice requires the formation of a focal aggregate of cells around a migrating intravascular larva (32). Early ultrastructural observations suggested that qualitative aspects of this focal response were crucial to its success, with a predominantly mononuclear cell composition of paramount importance (33). This was borne out by subsequent observations in IFN-γR−/− mice where challenge elicits a more intense cellular infiltrate in which eosinophils predominate, although protection (i.e., effectiveness of foci) is greatly diminished (3). The precise way in which the effector focus achieves parasite elimination has not been satisfactorily explained. There are two contending hypotheses, direct cytotoxic killing by agents such as NO (34), and simple physical blocking of migration until the parasite expires (35). The failure of nonprotected TNFRI−/− mice to up-regulate the iNOS message in the lungs after challenge argues for NO involvement in the effector response. Conversely, the induction of protection in iNOS−/− mice (7, 8), which still mount a dominant Th1 response (8), and the insusceptibility of lung stage larvae (1–2 wk old) to NO-mediated killing (36), argue against NO involvement in the effector response.
Our histopathological observations in the present study reveal that foci still develop in the lungs after challenge of vaccinated mice. Superficially, their cellular composition is normal, yet they are obviously completely ineffective at eliminating parasites. One hypothesis we can offer is temporal; we have previously shown that in a C57BL/6 mouse it takes more than 2 but less than 4 days for an effector focus to form around an embolized larva in the lungs (37). This timing is crucial because it is virtually identical with the first transit time of a larva through the lung vascular bed (38). Thus, a small delay in cellular aggregation around larvae due to the absence of TNFRI signaling would allow their escape. We are currently undertaking a histopathological investigation of the kinetics of focus formation after i.v. challenge of vaccinated WT and TNFRII−/− mice with lung schistosomula. That a focus would form at all in the above circumstances requires explanation. However, “empty” foci are quite frequent events in the lungs of vaccinated C57BL/6 mice after challenge. We have always assumed, in the absence of a discriminatory test, that they represent either an inflammatory infiltrate in which a larva has died and disintegrated, or perhaps more likely a “footprint” left by the larva as it sheds Ag during passage through the pulmonary vasculature (this traverse is a very tight squeeze that brings the larval tegumental surface into close and prolonged contact with capillary endothelium; Ref. 39).
It is not possible to draw precise parallels between microbial infections that are more widely disseminated in the lungs and schistosome larvae, each of which must serve virtually as a point source of Ag. However, analogous situations of defective granuloma formation have been reported after exposure of TNF−/− mice to heat-killed Corynebacterium parvum (40) and TNFRI−/− mice to live Mycobacterium avium (41, 42). In the former animals, although little or no initial response was observed, at later times a vigorous but disorganized inflammation developed, leading to death. In the latter, granuloma formation was delayed by 2 wk and the inflammatory lesions were less compact and malorganized than in WT mice (41). The above studies suggest that TNF is required to maintain the integrity and organization of granulomatous responses.
TNF has been implicated in several aspects of accessory cell function in immune responses. This is evidenced by a lack of germinal centers, due to defective follicular dendritic cell networks, in TNFRI−/− mice (18). TNF is also required for dendritic cell migration from peripheral tissue to lymphoid organs, best exemplified by epidermal Langerhans cells in a process that involves altered expression of adhesion molecules (42). The failure to up-regulate iNOS mRNA expression in the lungs of TNFRI−/− mice in the present study is also indicative of defective macrophage function. Although we have found altered adhesion molecule expression on CD4+ T cells recovered from the lungs of IFN-γR−/− mice, we were unable to make a link between the reduced homotypic adhesion of such cells and the minimal immunity displayed by these mice (9). We are currently investigating whether there are any phenotypic or functional differences in dendritic cells and macrophages from the lungs of TNFRI−/− vs WT mice, which might explain why effector foci in the former animals fail to eliminate challenge parasites.
This work was supported by a Project Grant from the Wellcome Trust to R.A.W.
Abbreviations used in this paper: iNOS, inducible NO synthase; WT, wild type; BAL, bronchoalveolar lavage; SSP, soluble schistosomular Ag preparation.