Release of tumour necrosis factor alpha into bronchial alveolar lavage fluid following antigen challenge in passively sensitized guinea-pigs

Five groups of ten female guinea-pigs were passively sensitized against ovalbumin (OA) (n = 9) or control guinea-pig serum (n = 1). 24 h later, they received mepyramine (0.5 mg/kg, i.p.) and 30 min later inhaled aerosols of: (A) OA (2 in 0.9% saline, 8 min, n = 4/9); (B) saline (40 min, n = 4/9); (C) LPS (40 min, Escherichia coli 0111:B4, 150 ng/kg in PBS, n = 1/9); and (D) the control animal was treated as in (C) (n = 1). Their tracheas were cannulated under pentobarbital anaesthesia and bronchial alveolar lavage (BAL) was performed with 2 × 5 ml PBS containing BSA (1%) (n = 1 group), or BSA (1%) and aprotinin (1000 KIU/ml) (n = 4 groups), at 30, 60, 90 or 120 min post-inhalations. BAL fluids recovered were centrifuged, the supernatants recovered and frozen until assayed for tumour necrosis factor-α (TNF-α), interleukin-1 (IL-1) and interleukin-6 (IL-6). No TNF-α could be detected unless aprotinin was present in the lavaging solution. BAL fluid from OA-sensitized and control animals that had inhaled LPS contained high levels of TNF-α that peaked at 90 min. BAL fluid from OA sensitized animals that inhaled OA aerosols contained no detectable TNF-α at 30 min, but it was found in increasing amounts at 60, 90 and 120 min; TNF-α was not detected in fluid from any of the animals that inhaled saline. As BAL fluids were toxic to the cells used in the assays, neither IL-1 nor IL-6 could be measured. We conclude that the monokine TNF-α is released into BAL fluid following anaphylactic challenge of passively sensitized guinea-pigs. The presence of the antiprotease, aprotinin, in the lavaging solution is essential for the detection and measurement of TNF-α in BAL fluid.


Introduction
Cytokine mediated interactions among macrophages, lymphocytes and eosinophils appear to be involved in the pathogenesis of the eosinophilia and airways' inflammation that characterize asthma. >4 Thus, tumour necrosis factor alpha (TNF-0) and granulocyte-macrophage colony-stimulating factor (GM-CSF), given parenterally, induce the accumulation of eosinophils in the airways of normal guinea-pigs. Also, TNF-cz is released from sensitized lung tissue following IgE receptor triggering, 6 and it contributes to mast cell dependent recruitment of leukocytes during IgE dependent cutaneous late phase reactions. Alveolar macrophages isolated from bronchial alveolar lavage (BAL) fluid from allergen challenged patients undergoing a late asthmatic response secrete greater amounts of TNFand interleukin-6 (IL-6) than macrophages isolated from BAL fluid of patients who develop no response or only an early response to allergen challenge. However, Gosset et al. were unable to detect TNF-0 in BAL fluid from asthmatic patients challenged with antigen inhalation. By contrast, Broide et al. 9 showed that levels of several cytokines, including TNF-, IL-1 and IL-6, were significantly elevated in BAL fluid from patients with symptomatic asthma compared to asymptomatic asthmatic controls. We used guinea-pigs passively sensitized to ovalbumin (OA), and determined whether the cytokines TNF-z, interleukin-1 (IL-1) and interleukin-6 (IL-6) could be detected in BAL fluid immediately after challenge with an aerosol of OA. Guinea-pigs passively sensitized to OA and to inert serum that inhaled LPS served as positive controls for TNF-cz release.

Methods
Animals: Five groups of ten female Hartley strain guinea-pigs, SPF quality, weight range 320-400 g, obtained from Charles River, St Constant, Qu4bec were housed in laminar flow units (Bioclean TM, Hazleton, MD) on grids in cages suspended over trays of rock salt. They were fed normal guinea-pig chow supplemented with apples and allowed water ad libitum. They were monitored for at least one week after being shipped to ensure that they were in good health. They weighed 350--450 g at experiment.
Bronchial alveolar lavage: One group of guinea-pigs' lungs was lavaged with 2 x 5 ml PBS containing BSA (1%); the other four groups were lavaged with 2 x 5 ml PBS containing BSA (1%, to maintain viability of any cells recovered) and aprotinin (1000 KIU/ml, to inhibit protease activity8). BAL fluids recovered were centrifuged (2000 rpm, 5 rain) and the supernatant removed and immediately frozen at -20C until assay.
IL-6 was measured as described recently. 11 Briefly, B9 hybridoma (105 cells/well) was incubated in fortified RPMI 1640 medium, in the presence of serial dilutions of BAL fluid. Growth stimulation was assessed by the MTT procedure described above. IL-6 levels were derived from a standard curve prepared with human rIL-6. IL-1 was measured via the ability of samples of BAL fluid to support the concanavalin A driven growth of D10.G4.1 cells. 12 Cells (4 x 105) were suspended in complete medium containing concanavalin A (2.5 #g/ml) with serial dilutions of BAL fluid. After incubation (72h, 37C, 5% COg), growth was assessed by the MTT procedure. A standard curve was prepared using human recombinant IL-1.
Guinea-pig sera" On day 0, guinea-pigs were given OA (20 mg/kg, i.p.); 21 days later, they inhaled OA aerosols (2% in saline, mepyramine, 0.5 mg/kg, i.p., 30 min beforehand) for up to 8 min, on 8 consecutive days. On day 35, they were anaesthetized (pentobarbital, 50mg/kg, i.p.) and blood collected into plain glass tubes via cardiac puncture. Tubes were centrifuged and stored at 4C overnight to allow the clot to form and retract. Serum was collected, pooled and stored at -20C until use.
Anti-OA antibody titres, as total IgG, were estimated via ELISA. Control sera were prepared similarly except animals received only vehicle on day 0, and inhaled only saline aerosols. No anti-OA antibodies could be detected in control sera.

Results
TNF-: In the absence of aprotinin, TNFcould not be detected in BAL fluids obtained at 60 min post-challenge whether guinea-pigs had inhaled OA saline or LPS (positive controls) aerosols. However, if animals' lungs were lavaged with fluid containing the antiprotease, aprotinin (1000 KIU/ml), significant amounts of TNF-0 were detected (Table 1, Fig. 1). In the group challenged with OA aerosol, TNFwas not detectable in BAL TNF-o release after antigen challenge Table 1. Amounts of TNFdetected in bronchial alveolar lavage (BAL) fluids, aprotinin (1000 KIU/ml) present in the lavaging solution. Four groups of ten guinea-pigs were used. In each group, nine were passively sensitized with anti-ovalbumin (OA) serum, and one received control serum (CON). 24 h later, they received mepyramine (0.5 mg/kg, i.p.) and 30min later they inhaled aerosols of OA (2% in saline) (OA/OA, n 4), saline (OA/SAL, n 4), or LPS (OA/LPS, n 1) and control serum treated animals inhaled LPS (CON/LPS, n 1). 30 fluid collected 30 min post-challenge, but increasing amounts of TNFwere detected at 60, 90 and 120 min post-challenge and were correlated positively with time BAL fluid was collected postchallenge. By contrast, TNF-0 was not detected in BAL fluid from animals challenged with only saline. However, TNFwas detected in BAL fluid from both positive controls, the passively sensitized and control serum treated animals that had inhaled LPS aerosol ( Table 1, Fig. 1). The cytotoxicity of this TNF-0 was prevented if the BAL fluid, rabbit anti-human TNF-0 serum (5/il) and the L929 cells Acute effects of inhalation of OA, safine and LPS aerosols in guinea-pigs: In all guinea-pigs that were passively sensitized to OA and inhaled OA aerosols, some signs of anaphylaxis were noted despite the presence of mepyramine. Animals were carefully observed and those that showed signs of respiratory distress were immediately removed from the exposure chamber to prevent progression to severe respiratory distress and death. Exposure times in the OA/OA groups ranged from 0.5 to 8 min. Mean exposure time was 4.4-+-0.7 min (n 20). Duration of exposure to OA aerosol did not correlate with the amount of TNF-0 detected in BAL fluid.

Discussion
We report the detection of TNF--like activity in BAL fluid obtained from guinea-pigs passively sensitized to OA and challenged with OA aerosols.
TNF-0-like activity could not be detected at 30 min post-challenge, but it was measurable (8-26 U/ml) in fluid obtained at 60, 90 and 120 min post challenge. As rabbit anti-human TNF-0 antibodies prevented the cytotoxicity of the lavage fluids, it is highly likely that it is TNF-0. TNFlevels in BAL fluids from animals that had inhaled LPS appeared to peak about 90 min post-aerosol inhalation. It is worth noting that these guinea-pigs inhaled LPS aerosol for 40 min compared to OA aerosol for up to 8 min in the other group. Thus, as we have not defined the peak of TNFsecretion into BAL fluid for OA challenged animals, we cannot say whether 120 rain represents the maximal concentration that may be achieved or whether its pattern of secretion follows that seen after LPS inhalation. The levels of TNF-0 detected in these experiments are similar to those found by others in BAL fluid from guinea-pigs after inhalation of LPS 13 or cotton dust. 14 Unlike Broide et al., 9 who measured 'resting' TNF-0 levels in BAL fluid t'rom symptomatic and asymptomatic asthmatics, but like Gosset et al., 8 we could not detect TNFunless a protease inhibitor was present in the lavaging solution. This is noteworthy. All the passively sensitized guinea-pigs D. E. Kelly, M. Denis and D. F. Biggs showed signs of anaphylaxis upon inhalation of OA aerosols, but the duration of exposure to the OA aerosol did not appear to influence the amount of TNF-0 detected in BAL fluid. Others 14'1s have shown that antigen challenge of sensitized animals leads not only to the release of a variety of mediators, but also to the release of proteases that could metabolize cytokines such as TNF-0. Thus, as BAL was performed 30-120 min after challenge, lack of aprotinin in the lavaging solution could have allowed metabolism of any TNF-0 released and thus prevented its detection. This finding emphasizes the importance of adding antiproteases to lavaging solutions to help prevent breakdown of cytokines during and after BAL. In these experiments, the sensitizing antibody is most likely to be a subtype of IgG, probably IgG1 or IgG2, as passive cutaneous anaphylaxis tests (data not shown) revealed little or no IgE in the serum used to passively sensitize animals. These findings show that passive sensitization with anti-OA antibodies of the IgG class followed by antigen challenge can induce directly or indirectly the synthesis and secretion of TNF-0. The production of TNF-0 is not unique to a particular cell type, but a major source of TNF-0 is macrophages. 2 Gosset et a/. 8'17 showed that alveolar macrophages collected from antigen challenged asthmatics undergoing a late reaction generated significantly more TNFthan similar patients who underwent only an early reaction or who had experienced no reaction. We speculate that, in our experiments in guinea-pigs, most of the TNF-0 that is detectable is secreted from alveolar macrophages. Neither IL-1 nor IL-6 could be detected as the lavage fluid was toxic to the cell types used in the bioassay of these cytokines. Preliminary experiments (data not shown) showed that ELISAs for murine IL-1 and IL-6 could not demonstrate the presence of these cytokines in BAL fluid. Thus, whether these cytokines are present in physiologically significant amounts in BAL fluid remains to be determined.
The physiological significance of TNF-0 release following passive sensitization and antigen challenge is unclear. Our data suggest that the release of proteases and other mast cell associated mediators during anaphylaxis may act to reduce TNF-'s pro-inflammatory effects. Gosset et al.18 reported that human alveolar macrophages from asthmatics have an enhanced capacity to produce TNF-, compared with cells from normal control subjects. Thus, TNF-0 may have a greater role in the chronic inflammatory processes in asthma than was previously supposed.
We conclude that TNF-0 is released in passively sensitized guinea-pigs' lungs soon after antigen inhalation. A protease inhibitor must be present in the lavaging solution to enable TNF-'s detection.