Introduction

Abstract Introduction References

Multinucleated giant cells (MGCs) are a notable feature of the pulmonary immune response to many clinical or experimental infections such as measles (1), Pneumocystis carinii (2), Nippostrongylus brasiliensis (3), and Schistosoma mansoni (4). They are also detected in abundance in lung tissue following the administration or inhalation of various chemical agents, including talc, silica, and asbestos (e.g., 5, 6). From these studies, it appears that MGCs are associated with chronic inflammation in the lungs and occur where there is the persistence of pathogens, or nonphagocytosable foreign material. However, although MGCs have been frequently reported in the literature for many years, their exact etiology and role in the immune response remain unclear (7).

MGCs arise from the fusion of two or more macrophages (8) and are thought to represent the final stage of macrophage differentiation (9). Nevertheless, the physiologic basis behind their formation is not well understood. It is probable that cytokine mediators of the host's immune response, rather than some intrinsic feature of the foreign antigen, play an important part in the development of MGCs. Some workers have suggested that MGC formation is associated with T helper (Th) 1-type cell-mediated immune responses and the production of cytokines that upregulate macrophage activation, such as interferon- (IFN-; 10, 11). In contrast, others have proposed that cytokines produced by Th2 cells, in particular interleukin (IL)-4 and IL-13, are responsible for macrophage fusion and the creation of MGCs (12).

In this context, one of the key factors that affect the development of Th1 versus Th2 cell populations during priming of the immune response is the monokine IL-12 (15). It has a crucial role in the generation of Th1 cells via the stimulation of natural killer and T cells to secrete abundant IFN- (16). Conversely, the absence of IL-12 is associated with the development of Th2-biased responses (15, 17). The eventual development of MGCs may therefore depend to a large extent on the initial role of IL-12 in creating an appropriate Th1 or Th2 cytokine environment.

Our study investigates the influence of IL-12 in the formation of pulmonary MGCs using mice that are genetically deficient for the p40 subunit of IL-12 and that have a biased Th2 phenotype, and wild-type (WT) mice supplemented with exogenous recombinant IL-12 that have a strong Th1 phenotype. We have analyzed as a model the immune effector responses to challenge parasites of the helminth S. mansoni in the lungs of mice previously exposed to radiation-attenuated larvae. Normal parasites enter the host via the skin and then migrate intravascularly to reach the lungs and eventually the hepatic portal system, where they mature up to Day 35. In contrast, attenuated larvae have a truncated migration and progress no farther than the lungs, where they eventually die by Day 21 (18). The immune response primed by these attenuated larvae in WT mice is mediated by Th1 cells and abundant IFN-, and acts against challenge parasites as they arrive in the lungs (19). In the current study, we conclude that the formation of MGCs in the lungs in response to challenge larvae is greatly enhanced in the absence of IL-12 and is linked instead to the abundance of Th2 cytokines, such as IL-4 and IL-13.

	Materials and Methods

Mice and Parasites

Mice with a targeted deletion (exon 3) in the p40 subunit of the IL-12 heterodimer, on a C57BL/6 background, were obtained from Hoffman-La Roche (Nutley, NJ) (17). These knockout (KO) mice were subsequently bred and maintained in isolators at the University of York animal facility alongside WT C57BL/6 mice. A Puerto Rican isolate of the parasitic helminth S. mansoni was routinely maintained by passage through MF1 strain mice and Biomphalaria glabrata snails.

Experimental Protocol

WT and KO mice were exposed to 500 -irradiated (20 krad) cercariae of S. mansoni via the shaved abdomen. Half of the mice in each group were treated with murine recombinant IL-12 (rIL-12; Genetics Institute, Cambridge, MA) at 0.5 µg/dose in a volume of 25 µl, delivered intra-dermally over the sternum on Days 1, 2, and 4, and in a volume of 100 µl intraperitoneally on Days 8 and 11 after exposure. After 5 wk, all four groups of mice (i.e., WT, WT + IL-12, KO, and KO + IL-12) were challenged with 200 normal parasites via the tail. Fourteen days later, at the peak of the pulmonary effector response, the lungs were sampled from the four groups of mice as described subsequently. In some experiments, lungs of KO mice were also sampled at Days 8, 11, 14, 21, and 28 to assess the timing of MGC formation.

Recovery and Characterization of Pulmonary Leukocytes

Pulmonary leukocytes were recovered from the airways by bronchoalveolar lavage (BAL; 20) using 10 ml of sterile phosphate-buffered saline containing 12 mM lidocaine chloride, 20 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid, and 1% bovine serum albumin (Sigma Chemical Co., Poole, UK) warmed to 37°C. The cells were chilled on ice, washed, and brought to 1 ml in GMEM plus 10% fetal calf serum (FCS; GIBCO BRL, Paisley, UK), 200 U/ml penicillin, and 100 µg/ml streptomycin (GMEM+; Sigma). Total cell number was estimated using an improved Neubauer hemocytometer. The proportions of the major leukocyte classes within each BAL sample were determined on the basis of relative size and granularity, using a Coulter XL flow cytometer (Coulter Electronics, Luton, UK) equipped with a 488-nm emission argon-ion laser (20). Absolute numbers for each cell population were then calculated using the hemocytometer counts. The mean forward light scatter (FLS) signal of the gated macrophage population was used as an indicator of cell size (21). Cytospin preparations of BAL samples were stained using Diff-Quik (Baxter Diagnostics AG, Dudingen, Switzerland) to distinguish eosinophils.

Culture of Pulmonary Leukocytes and Cytokine Enzyme-Linked Immunosorbent Assays

BAL cells (2 × 10 ⁵ cells/well) in GMEM+ were cultured in 96-well plates in the presence or absence of soluble lung-stage parasite antigen (SLAP; 50 µg/ml) (22). Culture supernatants were removed at 48 and 72 h for cytokine measurement. Cytokine-specific double-antibody enzyme linked immunosorbant assays (ELISAs) were used to measure the amounts of IFN-, IL-4, IL-5, and IL-10 present in the culture supernatants as described previously (20, 21). IL-13 was measured by ELISA (Quantikine^TM; R&D Systems, Oxford, UK) according to the manufacturer's instructions relative to a recombinant IL-13 standard curve.

Reverse Transcription-Polymerase Chain Reaction of Cytokine Messenger RNA (mRNA)

A single lung lobe was removed from individual mice after BAL, placed in 0.5 ml RNAzol (Biogenesis, Ltd., Poole, UK) and stored at 80°C before processing. Thawed tissues were homogenized in RNAzol using a tissue shearer (Ultra-Turrax; IKA, Staufen, Germany) fitted with a 5-mm blade, and RNA was extracted according to the manufacturer's protocol. Following phenol extraction and ethanol precipitation, the RNA yield was measured at 260/280 nm (DU 640B spectrophotometer; Beckman, High Wycombe, UK). Reverse transcription (RT) was performed on 1 µg mRNA exactly as described previously (23). Primers were based on previously described work: IL-4 and IFN- (23), IL-5 (24), IL-13 (25), IL-6 (26); hypoxanthine guanine phosphoribosyl transferase (HPRT; 27), and tumor necrosis factor- (TNF-) (27). Diluted complementary DNA (cDNA) (1:15) was amplified for 1 min at 94°C; 1 min at 55, 60, or 65°C; and 2 min at 72°C over 27 cycles (HPRT), or 30 cycles (IFN-, IL-4, IL-5, IL-6, IL-13, and TNF-) using a 480 Thermal cycler (Perkin Elmer, Norwalk, CT). A total of 5 µl of each polymerase chain reaction (PCR) product was denatured, slot-blotted onto Zeta probe GT membrane (Bio-Rad, Hemel Hempsted, UK), and then baked at 80°C for 1 h under vacuum. The membranes were hybridized with ³²P end-labeled oligonucleotide probes as described previously: IFN-, IL-4, and IL-5 (24); HPRT (27); IL-6 (26); and IL-13 (25). Finally, membranes were then washed at 1°C below the melting temperature of the probe, exposed to a phosphor screen, and analyzed on a phosphorimager (Molecular Dynamics, Sunnyvale, CA) to estimate the bound radioactivity. Counts were normalized between samples relative to levels of the housekeeping gene HPRT.

Processing and Histologic Examination of Lung Tissue

One lung lobe was taken from individual mice after BAL and immediately fixed in 4% formal saline. Tissue was paraffin-embedded, serially sectioned at 7 µm, and stained with Mayer's hemalum-eosin (BDH Laboratory Supplies, Poole, UK). Lung sections were examined using a Nikon Labophot microscope. Information on the presence, size, character, and composition of any associated cellular aggregates was recorded (21).

Statistical Analysis

All data comparisons were tested for significance by using Student's t test. Arithmetic means are shown ± SEM.

	Results

Histologic Examination of Lung Tissue

Fourteen days after challenge, numerous cellular foci were observed in the sections of lung tissue obtained from WT mice. The foci were distributed throughout the lungs and are thought to be responsible for parasite elimination (19). The majority surrounded a challenge parasite, although it was not always detected at the center of the focus (Figure 1A). In WT mice, the foci were largely composed of mononuclear cells, with few granulocytes and only an occasional MGC. In comparison, foci in the lungs of KO mice were larger and more profuse (Figure 1B). These foci were similar in numbers to those in WT mice and were usually associated with a challenge parasite (although none is present in the section shown in Figure 1B). Cellular infiltrates were more widely distributed in the pulmonary parenchyma of KO mice, with additional smaller aggregates not associated with a parasite (Figure 1B). The foci in KO mice contained many large macrophages and MGCs, and eosinophils were also extremely common (Figure 1C). Some of the MGCs in KO mice were enormous (e.g., 50 × 60 µm; Figure 1D), with more than 30 nuclei counted in a sequence of sections. Their shape was highly variable, some being approximately ovoid whereas others were irregular in outline (Figure 1B). Most of the MGCs from the KO mice showed evidence of striations in their cytoplasm, which seemed to represent rodlike inclusion bodies (Figures 1C and 1D). Another feature of the foci in KO mice was that many of the MGCs contained large numbers of clearly discernible eosinophils within vacuoles in the cytoplasm (Figure 1E). Morphometric analysis of representative sections through the pulmonary foci revealed a mean of 10.6 ± 0.8 MGCs per focus area. Furthermore, although MGCs represented approximately 1% of the cells in each focus, they occupied between 5.2 and 11.4% (mean = 7.4 ± 0.86%) of the area. On Day 8 no MGCs were detected, but on Day 11 a limited number of small MGCs were noted at the periphery of some of the foci in KO mice (data not shown). Large MGCs of the type described in detail above were present from Day 14 to Day 28, although at this later time point they were clearly vacuolated (data not shown).

View larger version (151K): [in this window] [in a new window]   Figure 1.   Inflammatory foci surrounding challenge parasites in the lungs of mice exposed to irradiated schistosomes at Day 14 after challenge infection. In WT mice, a compact cellular focus (F ) composed largely of macrophages and lymphocytes surrounds a challenge parasite (P) (panel A). This contrasts with KO mice, in which the main focus (F ) is much larger with a looser composition (panel B); several additional but smaller foci are also seen in this section. The foci in KO mice contain numerous eosinophils (E), large macrophages (M), and large giant cells (MGC) (panels C, D, and E). Some of the MGCs were enormous (50 to 60 µm) with many nuclei (panel D), whereas in others there was clear evidence of phagocytosed eosinophils (indicated by arrows; panel E ). A number of striations (S) are visible in the cytoplasm of the MGCs ( panels C and D). Bars are 40 µm.

The administration of rIL-12 to WT mice after exposure to irradiated parasites did not substantially alter the size or cellular composition of the foci around challenge larvae (data not shown). In contrast, administration of IL-12 to KO mice had dramatic effects on the foci. The abundant MGCs and eosinophils detected in the untreated KO mice were almost totally absent after treatment with rIL-12 (data not shown). Indeed, the small, dense foci observed in the proximity of parasites were similar in composition (predominantly macrophages and lymphocytes) and size to those observed in WT mice.

Composition of Cell Populations Recovered by BAL

BAL revealed that at Day 14 after challenge, the cellular content of the airways from KO mice was just over twice as large as that in the WT group (P < 0.05; Figure 2a). In WT mice, the BAL cell population was composed mainly of macrophages (Figure 2b), with lower numbers of lymphocytes (Figure 2c) and granulocytes (Figure 2d). However, in KO mice the major cell population was granulocytic in nature and was largely responsible for the increase in total cell number. Indeed, there were 8.2-fold more granulocytes in the KO than in the WT cell population (Figure 2d; P < 0.01). Cytospin preparations established that more than 95% of these granulocytes were eosinophils. In addition, there was a significant increase (2.3-fold; P < 0.05) in the number of lymphocytes in KO lungs, although there were slightly fewer macrophages (Figure 2c; P < 0.01) than in WT mice.

View larger version (12K): [in this window] [in a new window]   Figure 2.   BAL cell populations at Day 14 after challenge in WT, KO, WT + IL-12, and KO + IL-12 mice, showing the total cell number (a) per mouse, and total numbers of lymphocytes (b), macrophages (c), and granulocytes (d). Cell populations were distinguished using a flow cytometer on the basis of their FLS and granularity, and then multiplied by hemocytometer cell counts. Bars are means of individual mice (n = 5) ± SEM.

Addition of rIL-12 to WT mice up to Day 11 after initial parasite exposure had little effect on the numbers and proportions of cells recovered by BAL on Day 14 after challenge (38 days after rIL-12 treatment had ceased) compared with untreated WT mice (Figure 2a). However, in KO mice the effect of this early rIL-12 administration was to reduce the number of lymphocytes (Figure 2b) and to ablate almost completely the large granulocyte population detected previously in the untreated KO group (Figure 2d). Thus, the resulting BAL cell population in KO mice treated with rIL-12 had much the same profile as that seen in WT mice.

Although fewer macrophages were detected in KO than in WT mice (Figure 2c), they were significantly larger (P < 0.001), as estimated from the mean FLS value (Figure 3). Administration of rIL-12 had no effect on macrophage size in WT mice, whereas in KO mice it significantly decreased the mean value below that recorded for both untreated KO (P < 0.001) and untreated WT mice (P < 0.05).

View larger version (19K): [in this window] [in a new window]   Figure 3.   Mean size of alveolar macrophages recovered by BAL on Day 14 after challenge. The mean FLS channel of the gated macrophage population, analyzed by flow cytometry, was used as an indicator of cell size. Bars are means of individual mice (n = 5) ± SEM. (*P < 0.05, ***P < 0.001 compared with WT value).

Production of Cytokines by BAL Cells

BAL cells recovered from the lungs of WT mice 14 d after challenge, at the peak of parasite elimination, secreted abundant IFN- (~ 7,000 pg/ml) when cultured in the presence of SLAP antigen (Figure 4a). Administration of rIL-12 to WT mice increased the amount of IFN- detected, although this was not significant (P > 0.05). In contrast, negligible quantities of this cytokine were secreted by BAL cells from KO mice, compared with WT mice (P < 0.001). However, following treatment of KO mice with rIL-12, significant quantities of IFN- (~ 6,000 pg/ ml) were detected compared with untreated KO mice (P < 0.001) and were similar to those detected for untreated WT mice.

View larger version (21K): [in this window] [in a new window]   Figure 4.   The effect of rIL-12 administration to WT and KO mice on SLAP-specific cytokine production by BAL cells taken on Day 14 after challenge. IFN- (a), IL-4 (b), IL-5 (c), and IL-13 (d) production in 48-h BAL cell culture supernatants were determined by ELISA. Cells were pooled within groups of mice (n = 5) and error bars are for replicate ELISA wells. This data is representative of repeat experiments.

IL-4 secretion was undetectable in BAL cell cultures from WT mice, whereas > 200 pg/ml was produced by KO cultures (Figure 4b). The Th2-biased response in the lungs of KO mice after challenge was confirmed by the detection of significant quantities of IL-5 (1.96 ng/ml) in BAL cell cultures (Figure 4c). However, treatment of KO mice with rIL-12 inhibited production of both IL-4 and IL-5 (Figures 4b and 4c), with the net result that the cytokine profile for BAL cell cultures resembled that of the untreated WT group of mice. Production of IL-13 was detected in both WT and KO mice, although the amount was nearly 4-fold greater in the latter group (Figure 4d). Treatment of WT mice with rIL-12 resulted in a marginal decrease in the production of IL-13 from BAL cell cultures, but in KO mice treatment with rIL-12 reduced the production of IL-13 to the levels detected in WT mice.

RT-PCR Analysis of Cytokine mRNA from Whole Lung Tissue

The level of IFN- mRNA transcripts was extremely low in KO compared with WT mice (P < 0.05) and was similar to that recorded in naive mice (Figure 5a). In KO mice treated with rIL-12, the levels of IFN- mRNA increased significantly (P < 0.05), although this level was still well below that in untreated WT mice (P < 0.05). In contrast, IL-4 mRNA was significantly higher in KO than in WT mice (P < 0.05), and the administration of rIL-12 resulted in a dramatic decrease in the levels of IL-4 mRNA in both WT and KO mice (Figure 5b). A similar pattern of results was obtained for IL-5 mRNA, which was slightly higher in KO than in WT mice and was downregulated following rIL-12 treatment, although none of these changes were statistically significant (all P > 0.05; Figure 5c).

View larger version (29K): [in this window] [in a new window]   Figure 5.   Levels of IFN- (a), IL-4 (b), IL-5 (c), TNF- (d), IL-6 (e), and IL-13 ( f ) mRNA transcripts in the lung tissues of WT, KO, WT + IL-12, and KO + IL-12 mice 14 d after challenge; levels of mRNA in naive WT and KO mice are also shown. Bars are means of data for individual mice (n = 3) ± SEM. Counts represent cDNA product amplified by PCR, hybridized with 32P end-labeled oligonucleotides, and adjusted relative to the corresponding level of HPRT.

The profile of TNF- mRNA transcripts mirrored those of IFN-, with lower levels in KO than WT mice (P < 0.05), barely above their naive KO cohorts (Figure 5d). Administration of rIL-12 boosted the levels of TNF- mRNA in both WT and KO groups compared with their untreated cohorts (both P < 0.05). The patterns of IL-6 and IL-13 mRNA were similar, with both being detected more strongly in KO than in WT mice, but due to high variance in the KO group this was not significant (P > 0.05; Figures 5e and 5f). Levels of mRNA for both cytokines were significantly upregulated in both KO and WT mice compared with their respective naive control groups (IL-6, P < 0.05; IL-13, P < 0.01). The effect of rIL-12 administration on the levels of mRNA for these cytokines in WT mice was minimal, but in KO mice transcripts for both IL-6 and IL-13 were decreased compared with their untreated cohorts (P < 0.05).

	Discussion

In this study we have analyzed the pulmonary inflammatory response in mice exposed to the helminth S. mansoni as an in vivo model in which to investigate the factors that affect MGC formation. Previous studies in WT mice have demonstrated that the cellular composition of these inflammatory foci was quite heterogeneous, comprising macrophages and CD4⁺ T cells with some eosinophils (4, 21, 28). In addition, MGCs were often noted on the periphery of the infiltrates (4). However, it is unclear whether MGCs were a product of the predominantly Th1 cell environment that occurs in these mice with the secretion of abundant IFN- (20, 21), or whether they were linked to the development of a Th2 cell population when IFN-- mediated signaling was absent (23, 29). To determine which cytokines are important for the formation of MGCs in this model of pulmonary inflammation, we set out to polarize the Th cell profile of the immune response by manipulation of IL-12, which is one of the most important factors in Th subset differentiation. Hence, a dominant Th1 cell population was ensured by the administration of rIL-12 to WT mice, whereas a polarized Th2 cell population was achieved by using mice with a disruption of the IL-12 p40 gene.

We report here that numerous and very large MGCs were detected only in the lung tissue of IL-12 p40 KO mice, whereas none were detectable in WT or KO mice receiving rIL-12 treatment. The MGCs in IL-12 KO mice were distributed throughout the inflammatory foci in the lungs; they were easily detected, even in sections at low power. In addition to being numerous, some of the MGCs were enormous (at least 50 µm), with many tens of nuclei detected in serial sections of the same cell. It was also recorded that the average size of alveolar macrophages recovered from the airways was significantly greater in KO mice than in their WT cohorts, although the physiologic significance of this observation is unclear. Since the MGCs are so big, it is unlikely that they would be recoverable by BAL and so would not contribute to the increased mean size of macrophages as recorded by flow cytometry. Indeed, the fusion of several macrophages to form MGCs could explain why the total number of macrophages in the BAL is lower in KO than in WT mice. The cellular foci in the lungs of vaccinated mice form within a few days of the arrival of challenge larvae (21), and MGCs were detectable soon after, by Day 11 (4). We first observed a limited number of small MGCs by Day 11 in the lungs of KO mice following the influx of eosinophils. At later time points MGCs were still detectable, but on Day 28 they appeared more vacuolated and may have been in the process of disintegration.

The lymphocyte population of the BAL cells from KO mice was also significantly greater than from WT mice, with the majority being CD4⁺ (data not shown). The elevated levels of transcripts for IL-4, IL-5, IL-6, and IL-13 mRNA in lung tissue, and the secretion of significant quantities of IL-4, IL-5, and IL-13 by antigen-specific BAL cells, indicated that many of these lymphocytes were of the Th2 type. However, quantitation of Th2-type versus Th1-type cells would be required before formal dominance of the former subset could be established. It was also noted that transcripts for IFN- mRNA were downregulated, along with those for another known macrophage activator, TNF-, and negligible quantities of IFN- were secreted in BAL cell cultures. Therefore, the enlarged macrophages and numerous MGCs detected in the lungs of KO mice in our study clearly cannot be due to the effects of IFN- and TNF-. Indeed, the formation of MGCs in response to P. carinii infection was also related to the absence of IFN- (2). Consequently, the formation of MGCs in the lungs appears to be more closely associated with the presence of several Th2-type cytokines.

Our observations linking the production of IL-4 with MGC formation accord well with the conclusions made from several in vitro studies. For example, human blood monocytes (12) and murine bone-marrow macrophages (13) can be induced to form into MGCs in the presence of IL-4, although the addition of IL-3 and granulocyte macrophage colony-stimulating factor clearly accelerated fusion. IL-4 is a highly pleiotropic cytokine with marked inhibitory effects on the expression and release of proinflammatory cytokines such as IFN-, IL-1, and TNF-. In this respect, the related cytokine IL-13 is also a potent downmodulator of these inflammatory mediators (30) and other macrophage effector functions, such as nitric oxide release (31). Moreover, Doherty and colleagues (14) observed a positive correlation between decreasing IL-13 concentrations and MGC numbers in cultures of murine bone-marrow monocytes, and De Fife and coworkers demonstrated that administration of IL-13 or IL-4 to in vitro macrophage cultures directly induced cell fusion (32). In addition, there are reports of a role for another Th2 cytokine, IL-6, in the formation of osteoclast-like MGCs from bone-marrow cultures (33). Together, these studies and our own data indicate a probable role for Th2-associated cytokines (particularly IL-4 but also IL-13 and IL-6) in MGC formation.

However, our observations in vivo do not agree with a number of in vitro studies, which show that IFN- is the major inducer of MGC formation (10, 34, 35) via the upregulation of lymphocyte function-associated antigen and intercellular adhesion molecule-1 interactions that enhance macrophage fusion (10). In fact, we observed very few MGCs in WT mice with a Th1 cell population in the lungs as judged by the detection of IFN- and TNF- mRNA, or by the secretion of copious IFN- by antigen-stimulated BAL cells. Moreover, when rIL-12 was administered to WT mice, transcripts for IFN- and TNF- were enhanced, as were the levels of secreted IFN-. Nevertheless, we failed to detect a single MGC in the lungs of mice with such a strongly polarized Th1 immune response.

In an attempt to resolve the contradictory evidence available on MGC formation from the in vitro studies, McNally and Anderson (12) suggested that IL-4 and IFN- mediated alternative pathways of macrophage fusion and classified two discrete types of MGC. Accordingly, foreign body giant cells (FBGC) form in the presence of IL-4 and contain many tens of nuclei, whereas Langerhans giant cells (LGC) form under the influence of IFN- and are much smaller, containing less than 12 nuclei. It is tempting to speculate that the MGCs detected in the lungs of IL-12 KO mice in the current in vivo study and in mice treated with anti-IFN- antibody (29), or with a targeted disruption to their IFN- receptor gene (23), are of the FBGC type. In contrast, in a different in vivo model of pulmonary inflammation induced by schistosome eggs, where the numbers of MGCs were reduced following anti-IFN- antibody treatment (36), the cells may be of the LGC type.

With regard to a function for MGCs of either classification, there is little hard evidence in the literature. It is possible that these cells form when macrophages try to engulf inert (e.g., asbestos fibers) or very large objects. Indeed, Fais and associates (7) suggest that MGCs represent an alternative, nonphagocytic method of antigen handling because they are both major histocompatibility complex class II and B7 positive, yet they are poorly phagocytic. Another explanation lies in the host's response to the huge influx of eosinophils in the pulmonary tissues. Indeed, eosinophils were the most frequent type of infiltrating cell in the lungs of KO mice. Moreover, many macrophages and MGCs contain clearly discernible eosinophils within vacuoles in their cytoplasm, which indicates that recent phagocytosis of these cells has occurred. In addition, many of the macrophages and MGCs have striations in their cytoplasm. These may be equivalent to the electron-dense inclusions, thought to be the breakdown products of phagocytosed eosinophils, reported previously in the pulmonary macrophages of schistosome-infected mice (4). Clearly in our KO mice, enormous and abundant MGCs are strongly associated with extreme eosinophilia.

In summary, we have shown that in this model of inflammation, MGC formation is associated with a biased Th2 immune response and the secretion of cytokines such as IL-4 and IL-13. The role of IL-12 in the regulation of polarized Th1 versus Th2 cell responses is emphasized, and this clearly affects the subsequent generation of MGCs. As such, KO mice deficient for IL-12 provide a valuable tool with which to study MGC formation in vivo. Although the role of MGCs in the immune response remains unclear, we suggest that they result from attempts by the host to clear and destroy the vast influxes of eosinophils, as detected in IL-12 p 40 KO mice. The role of endogenous IL-12 in the induction of protective immunity to S. mansoni is described elsewhere (37).