Abstract
Invasive pulmonary aspergillosis is a common and devastating complication of immunosuppression, whose incidence has increased dramatically in tandem with the increase in the number of immunocompromised patients. Given the role of TNF-α in other pulmonary infections, we hypothesized that TNF-α is an important proximal signal in murine invasive pulmonary aspergillosis. Intratracheal challenge with Aspergillus fumigatus conidia in both neutropenic (cyclophosphamide-treated) and nonneutropenic BALB/c mice resulted in the time-dependent increase in lung TNF-α levels, which correlated with the histologic development of a patchy, peribronchial infiltration of mononuclear and polymorphonuclear cells. Ab-mediated neutralization of TNF-α resulted in an increase in mortality in both normal and cyclophosphamide-treated animals, which was associated with increased lung fungal burden as determined by histology and as quantified by chitin content. Depletion of TNF-α resulted in a reduced lung neutrophil influx in both normal and cyclophosphamide-treated animals, which occurred in association with a decrease in lung levels of the C-X-C chemokine, macrophage inflammatory protein-2 and the C-C chemokines macrophage inflammatory protein-1α and JE. In cyclophosphamide-treated animals, intratracheal administration of a TNF-α agonist peptide (TNF70–80) 3 days before, but not concomitant with, the administration of Aspergillus conidia resulted in improved survival from 9% in control mice to 55% in TNF70–80-treated animals. These studies indicate that TNF-α is a critical component of innate immunity in both immunocompromised and immunocompetent hosts, and that pretreatment with a TNF-α agonist peptide in a compartmentalized fashion can significantly enhance resistance to A. fumigatus in neutropenic animals.
Invasive aspergillosis is a feared complication of immunodeficiency. Its incidence has increased dramatically over the past 2 decades, in tandem with an increase in the number of immunocompromised hosts (1, 2). Invasive pulmonary aspergillosis, the commonest form of the disease (3), carries a crude mortality rate of >80% with current therapy (4). Aspergillus species are the second commonest fungal pathogen in immunocompromised hosts, and A. fumigatus is responsible for 90% of cases of invasive aspergillosis.
The respiratory tract is the portal of entry in the majority of cases of human invasive aspergillosis (3). It is postulated that innate immunity is the principal pathway by which A. fumigatus is cleared from the lung. This innate response consists of alveolar macrophages, which represent the first line of defense against conidia entering the alveolus, and recruited neutrophils. Neutrophils are believed to kill conidia that have survived to form hyphae and cause invasive infection (5). Various cells of macrophages lineage can engulf and kill Aspergillus conidia in vitro (6, 7, 8) and in vivo (9, 10). Both macrophages (6, 11) and polymorphonuclear cells (5, 12) have been shown to damage hyphae in vitro. When exposed to conidia in vitro, cells of macrophage lineage secrete inflammatory mediators, including the proximal cytokines TNF-α and IL-1 (13, 14), as well as cytokines that promote Th-1 phenotype immune responses (15).
TNF-α is a 17-kDa cytokine secreted predominantly by various macrophage populations, including alveolar macrophages. It has shown to be a critical proximal signal in the initiation and maintenance of innate pulmonary immunity in animal models of pneumonia (16, 17) and human pneumonia (18). In the single study examining the role of TNF-α in invasive aspergillosis in vivo, immunocompetent animals administered A. fumigatus conidia i.v. had improved survival when treated with systemic murine rTNF-α (19). The role of TNF-α in pulmonary host defense against A. fumigatus has not been defined.
Given that TNF-α has been shown to be a critical mediator of innate immunity against several respiratory pathogens, attempts have been made to administer TNF-α to augment innate and acquired host responses. However, the therapeutic administration of TNF-α is limited by substantial dose-related toxicity, particularly when this cytokine is administered systemically. A TNF-α agonist peptide composed of the 11 amino acids that constitute the binding site of native human TNF-α to its receptors (referred to as TNF70–80)3 has recently been characterized (20, 21). Binding of TNF70–80 to TNF-α receptors (both p55 and p75) has been shown to mediate many leukocyte-activating effects of native TNF-α, including stimulation of neutrophils for enhanced protease release and respiratory burst, and enhancement of neutrophil phagocytic activity and killing. When administered systemically, TNF70–80 was associated with less toxicity than that observed with native TNF-α, due in part to the fact that this peptide did not alter adhesive properties of the endothelium (20). This peptide has been successfully administered directly into the lungs of mice with Gram-negative bacterial pneumonia, although the beneficial effects of TNF70–80 on bacterial clearance and survival were only observed when the peptide was administered 3 or 7 days before, but not concomitant with, bacterial challenge (22).
In the current study we hypothesized that TNF-α is an important proximal signal in murine invasive pulmonary aspergillosis. We tested this hypothesis by assessing the production of TNF-α in the lungs of normal and immunocompromised mice challenged with A. fumigatus conidia. We next examined the effect of TNF-α depletion on survival, the burden of fungal organisms, and the host inflammatory response. Finally, the effects of intrapulmonary delivery of native TNF-α or TNF70–80 on survival in cyclophosphamide-treated animals were determined.
Materials and Methods
Reagents
A. fumigatus administration. The i.p. injections of 0.25 ml of antiserum or control rabbit serum were repeated every 48 h for two additional doses. TNF70–80 and control peptides were synthesized at the University of Michigan Protein and Carbohydrate Structure Facility and were composed of the amino acid sequences H-Pro-Ser-Thr-His-Val-Leu-Ileu-Thr-His-Thr-Ileu-OH and H-Gly-Gly-Asp-Pro-Gly-Ileu-Val-Thr-His-Ser-OH, respectively (20). Both peptides were purified by HPLC and mass spectrometry, and contained no detectable LPS, as determined by Limulus lysate assay (ICN Biomedicals, Costa Mesa, CA).
Animals
Specific pathogen-free BALB/c mice (8-wk-old females; Charles River Breeding Laboratories, Wilmington, MA) were used in all experiments. All mice were housed in specific pathogen-free conditions within the animal care facility at the University of Michigan until the day of sacrifice. Immunocompromised animals were treated with 200 mg/kg of cyclophosphamide (Sigma, St. Louis, MO), administered i.p., 4 days before intratracheal (i.t.) inoculation. Treatment with cyclophosphamide resulted in peripheral blood neutropenia (absolute neutrophil count, <200 cells/μl) by days 4 and 6 after treatment, with a return of peripheral counts to pretreatment levels by day 10.
Preparation and administration of A. fumigatus conidia
We chose to use A. fumigatus strain 13073 (American Type Culture Collection, Manassas, VA) in our studies, as this strain has previously been shown to induce invasive aspergillosis in immunocompromised mice (23). The organism was grown on Sabouraud dextrose agar plates (Becton Dickinson, Cockeysville, MD) for 7–10 days at 37°C. The surface of each plate was then washed with 100 ml of sterile 0.1% Tween-80 (SigmaUltra, St. Louis, MO) in normal saline. The resulting suspension of conidia was filtered through sterile gauze to remove clumps and hyphal debris, and then washed once and resuspended in 4 ml of 0.1% Tween-80. The concentration of Aspergillus conidia in the suspension was determined by a particle counter (Z2 particle analyzer, Coulter, Hialeah, FL). The suspension was then diluted to the desired concentration, and the concentration was again measured before administration. In preliminary experiments, the number of particles determined by the particle counter was in close agreement with the number of viable colony-forming units found by serial dilution and plating of the suspension. On the day of inoculation, each animal was anesthetized with 1.8–2 mg of pentobarbital i.p. Using standard aseptic technique, the trachea was exposed and a 30-μl inoculum (A. fumigatus suspension or 0.1% Tween-80) was administered via a sterile 26-gauge needle. The skin incision was closed with surgical staples. In all experiments, immunocompetent animals were challenged with 1–2 × 107 conidia, and cyclophosphamide-treated animals were challenged with 1–2 × 106 conidia.
Lung harvest
At designated time points, the mice were sacrificed by CO2 asphyxiation. The chest cavity was opened aseptically, and the pulmonary vasculature was perfused with PBS via the right ventricle. For histologic examination, lungs were perfused with 1 ml of 4% paraformaldehyde in PBS, inflated with 1 ml of 4% paraformaldehyde in PBS via the trachea, and then excised en bloc. Lungs for various assays were perfused with 1 ml of PBS containing 5 mM EDTA, removed, frozen in liquid nitrogen, and stored at −20°C until the day of the assay. Lungs for cytokine and myeloperoxidase (MPO) assays were homogenized in 1 ml of 2× complete protease inhibitor mixture buffer (Boehringer Mannheim, Mannheim, Germany) in PBS using a tissue homogenizer (Biospec Products, Bartlesville, OK). A 900-μl aliquot of PBS was added to 900 μl from each sample, sonicated for 10 s, and centrifuged at 500 × g for 10 min. Supernatants were passed through a 0.45-μm pore size filter (Gelman, Ann Arbor, MI) and stored at 4°C for cytokine ELISA.
Lung chitin assay
Given that molds (including the Aspergillus species) do not reliably form reproductive units in tissue, we employed an assay for chitin to measure the burden of organisms in lungs. Chitin, a component of the hyphal wall, is absent from mammalian cells and conidia. The assay was adapted from a previously described method (24). Lungs were homogenized in 5 ml of distilled water and centrifuged (1500 × g, 5 min, 20°C). The supernatants were discarded, and pellets were resuspended in sodium lauryl sulfate (3%, w/v) and heated at 100°C for 15 min. Samples were then centrifuged (1500 × g, 5 min, 20°C), and pellets were washed with distilled water and resuspended in 3 ml of KOH (120%, w/v). Samples were heated at 130°C for 60 min. After cooling, 8 ml of ice-cold ethanol (75%, v/v) was added to each sample, and tubes were shaken until ethanol and KOH made one phase. Samples were incubated on ice for 15 min, and 0.3 ml of Celite suspension (supernatant of 1 g of Celite 545 (Fisher Scientific, Pittsburgh, PA) added to 75% ethanol and allowed to stand for 2 min) was added to each. Samples were centrifuged (1500 × g, 5 min, 4°C), and supernatants were discarded. Pellets were washed once with ethanol (40%, v/v) and twice with distilled water, and resuspended in 0.5 ml of distilled water. Standards, consisting of 0.2 ml of distilled water and 0.2 ml of glucosamine (10 μg/ml), were made up. NaNO2 (0.2 ml; 5%, w/v) and KHSO4 (0.2 ml; 5%, w/v) were added to each standard, and NaNO2 (0.5 ml; 5%, w/v) and KHSO4 (0.5 ml; 5%, w/v) were added to each tissue prep; all samples were mixed gently for 15 min and then centrifuged (1500 × g, 2 min, 4°C). Two 0.6-ml aliquots of supernatant from each tissue prep were transferred to separate tubes. Ammonium sulfamate (0.2 ml) was added to each tube, and all tubes were shaken vigorously for 5 min. A fresh solution of 3-methyl-2-thiazolone hydrazone HCl monohydrate (50 mg in 10 ml of distilled water) was made, and 0.2 ml was added to each tube. Samples were then heated to 100°C for 3 min and cooled. FeCl3·6H2O (0.2 ml; 0.83%, w/v) was added to each, and OD was measured at 650 nm after 25 min. Chitin content, measured in glucosamine equivalents, was measured by the following formula: chitin content = [5 × (OD of organ − OD of control organ)]/(OD of glucosamine − OD of water).
Lung MPO activity
Lung MPO activity was measured as a marker of neutrophil sequestration, as described previously (25). Briefly, a 100-μl aliquot of each lung homogenate was added to 100 μl of a buffer containing 50 mM potassium phosphate (pH 6.0), 5% hexadecyltrimthylammonium bromide, and 5 mM EDTA. Samples were sonicated for 10 s and centrifuged at 3000 ×g for 15 min. The supernatant was mixed 1/15 with assay buffer and read at 490 nm. MPO units were calculated as the change in absorbance over time.
Cytokine ELISA
Murine TNF-α, MIP-1α, JE, MIP-2, and KC were quantified using a modification of a double ligand method, as described previously (26). Briefly, flat-bottom 96-well microtiter plates (Immuno-Plate I 96-F, Nunc, Copenhagen, Denmark) were coated with 50 μl/well of rabbit anti-cytokine Ab (1 μg/ml in 0.6 M NaCl, 0.26 M H3BO4, and 0.08 N NaOH, pH 9.6) for 16 h at 4°C and then washed with PBS (pH 7.5) and 0.05% Tween-20 (wash buffer). Microtiter plate nonspecific binding sites were blocked with 2% BSA in PBS and incubated for 90 min at 37°C. Plates were rinsed four times with wash buffer, and undiluted and diluted (1/10) cell-free supernatants (50 μl) in duplicate were added followed by incubation for 1 h at 37°C. Plates were washed four times, followed by the addition of 50 μl/well biotinylated rabbit anti-cytokine Abs (3.5 μg/ml in PBS (pH 7.5), 0.05% Tween-20, and 2% FCS), and plates were incubated for 30 min at 37°C. Plates were washed four times, streptavidin-peroxidase conjugate (Bio-Rad, Richmond, CA) was added, and plates were incubated for 30 min at 37°C. Plates were washed again four times, and chromogen substrate (Bio-Rad, Richmond, CA) was added. The plates were incubated at room temperature to the desired extinction, and the reaction was terminated with 50 μl/well of 3 M H2SO4 solution. Plates were read at 490 nm in an ELISA reader. Standards were 1/2 log dilutions of murine rTNF, MIP-1α, JE, MIP-2, or KC, from 1 pg/ml to 100 ng/ml. This ELISA method consistently detected the relevant cytokine at concentrations >25 pg/ml. The ELISAs did not cross-react with IL-1, IL-2, IL-4, IL-6, IL-10, or IFN-γ. In addition, each ELISA did not cross-react with other members of the murine chemokine family.
Statistical analysis
Data were analyzed by a Power Macintosh 8600/300 computer using the InStat version 2.01 statistical package (GraphPad, San Diego, CA). Survival data were compared using the Fisher’s exact test. All other data were expressed as the mean ± SEM and compared using unpaired two-tailed Mann-Whitney (nonparametric) test. The p values were considered statistically significant if they were <0.05.
Results
Development of pulmonary inflammation in normal and cyclophosphamide-treated mice after A. fumigatus administration
To characterize the host response to Aspergillus in immunocompetent and immunocompromised animals, normal and cyclophosphamide-treated BALB/c mice were challenged i.t. with 1–2 × 107 and 1–2 × 106 A. fumigatus conidia, respectively. Lungs were examined histologically at various time points after challenge. By 2 days after A. fumigatus administration in both normal and cyclophosphamide-treated mice, infiltration of inflammatory cells was noted within the alveolar and interstitial compartments in a patchy, peribronchial pattern (Fig. 1⇓). The inflammatory cell infiltrate consisted predominantly of neutrophils and, to a lesser extent, mononuclear cells. By 4 days after A. fumigatus administration, cellular infiltrates were primarily mononuclear, with resolution occurring in immunocompetent mice, whereas a continued inflammatory response was noted in cyclophosphamide-treated animals up to 7 days after challenge (data not shown). Conidia were visible in normal mice 2 days after fungal inoculation and were nearly completely cleared by 5 days. In contrast, hyphae (the invasive form of A. fumigatus) were present only in cyclophosphamide-treated mice and were most numerous by 2–3 days after A. fumigatus administration. The inflammatory infiltrates observed corresponded to areas of conidial deposition in normal mice and to hyphae in cyclophosphamide-treated mice.
Effect of i.t. administration of A. fumigatus conidia on histopathology. A and B depict representative lung hematoxylin and eosin (H&E) and Gomori methanamine silver (GMS) stains in immunocompetent mice 2 days after inoculation with 1–2 × 107 A. fumigatus conidia (magnification, ×400). Inflammatory cellular infiltrate occurred in peribronchial areas, corresponding to sites of conidial deposition. Hyphal forms were not present. C and D depict representative lung H&E and GMS stains in cyclophosphamide-treated 2 days after inoculation with 1–2 × 106 A. fumigatus conidia (magnification, ×100). Inflammatory cellular infiltrate occurred peribronchially, in areas where branching hyphae were present. The data shown are representative of three experiments.
Lung TNF-α expression after administration of A. fumigatus conidia
We next correlated changes in lung TNF-α levels with the development of pulmonary inflammation in normal and cyclophosphamide-treated mice after administration of A. fumigatus conidia. As shown in Fig. 2⇓A, a rapid increase in lung TNF-α levels was noted in immunocompetent mice when they were challenged with 1–2 × 107 conidia, which reached a plateau by 24 h. Administration of 1–2 × 106 A. fumigatus conidia to cyclophosphamide-treated mice resulted in a marked elevation in lung TNF-α levels by 1 day after administration, with maximal TNF-α levels at 2 days, and a return to baseline levels by 4 days after A. fumigatus administration (Fig. 2⇓B). No significant induction of TNF-α was observed in the lungs of immunocompetent animals challenged with 1–2 × 106 A. fumigatus conidia or in cyclophosphamide-treated animals challenged with 0.1% Tween-80 (vehicle) i.t.
Time-dependent production of TNF-α protein in lungs from BALB/c mice after the i.t. administration of vehicle or A. fumigatus conidia. A depicts lung TNF-α levels in immunocompetent mice after A. fumigatus (1–2 × 107 conidia) or vehicle challenge. B depicts lung TNF-α levels in cyclophosphamide-treated mice after A. fumigatus challenge (1–2 × 106 conidia), immunocompetent mice after A. fumigatus challenge (1–2 × 106 conidia), and cyclophosphamide-treated mice after vehicle challenge. Cy, cyclophosphamide. Experimental n = 6. ∗, p < 0.05 compared with animals receiving vehicle i.t.
Effect of TNF-α neutralization on survival in normal and immunocompromised mice inoculated with A. fumigatus conidia
Additional studies were undertaken to evaluate the contribution of TNF-α to survival after inoculation of conidia. To ascertain the effectiveness of the polyclonal anti-murine TNF-α Ab in depleting TNF-α in vivo, we measured lung TNF-α levels in normal and cyclophosphamide-treated animals passively immunized with anti-TNF-α Ab or control serum 1 day before inoculation with A. fumigatus conidia. On the basis of earlier studies showing time of maximum expression of TNF, lungs were harvested 1 day after inoculation in normal mice and 2 days after inoculation in cyclophosphamide-treated mice. Lung homogenate TNF-α levels in normal mice were 3.53 ± 0.54 in the control group and 0.13 ± 0.051 in the anti-TNF-α group (mean ± SEM; n = 6 mice/group; p < 0.01). In cyclophosphamide-treated animals, levels were 6.16 ± 0.70 in the control group and 0.080 ± 0.007 in anti-TNF-α group (mean ± SEM; n = 6 mice/group; p < 0.01).
Having demonstrated significant in vivo neutralization of TNF, normal and cyclophosphamide-treated animals were treated with anti-murine TNF-α or control rabbit serum 1 day before challenge with conidia. Normal and cyclophosphamide-treated mice were then inoculated with 1.0 × 107 and 9.3 × 105 conidia, respectively. As shown in Fig. 3⇓, treatment with anti-murine TNF-α Ab resulted in 14% mortality in normal mice and 47% mortality in cyclophosphamide-treated mice, compared with no deaths in normal and cyclophosphamide-treated mice receiving control serum. All animals that survived beyond 7 days were considered long term survivors (>4 wk; data not shown).
Effect of TNF-α neutralization on survival after A. fumigatus challenge in immunocompetent and cyclophosphamide-treated mice. A depicts survival in immunocompetent mice after A. fumigatus conidia administration (1.0 × 107 conidia). B depicts survival in cyclophosphamide-treated mice after A. fumigatus challenge (9.3 × 105 conidia). Cy, cyclophosphamide. Experimental n = 15. ∗, p < 0.01 compared with animals receiving control rabbit serum.
Effect of TNF-α neutralization on lung fungal burden in normal and cyclophosphamide-treated mice
To determine the mechanism of increased lethality in mice passively immunized with anti-TNF-α Ab, we examined fungal burden in lungs of infected animals by histology and assessment of lung chitin content. Immunocompetent BALB/c mice or mice treated with cyclophosphamide were given either anti-TNF-α Ab or rabbit serum 1 day before challenge with 1–2 × 107 (immunocompetent) or 1–2 × 106 (cyclophosphamide-treated) A. fumigatus conidia. Animals were then sacrificed after 3 days for histology and lung chitin measurement. In immunocompetent mice, treatment with anti-TNF-α Ab resulted in the appearance of both conidial and hyphal forms at 3 days after the administration of conidia, whereas only conidial forms of A. fumigatus were seen in animals receiving preimmune serum (data not shown). In cyclophosphamide-treated animals, a marked increase in the number of hyphae was noted in animal treated with anti-TNF-α Ab, as evident on histology, compared with that in control animals (Fig. 4⇓, B and D). To quantitate differences in fungal burden, lung chitin levels were determined in cyclophosphamide-treated animals. Lungs of TNF-depleted animals had a greater than twofold increase in chitin content compared with those of control animals (Fig. 5⇓).
Effect of TNF-α neutralization on lung histopathology after A. fumigatus challenge in cyclophosphamide-treated mice. Cyclophosphamide-treated animals were given rabbit anti-TNF-α serum or rabbit control serum 1 day before challenge with A. fumigatus. A and B depict representative lung hematoxylin and eosin (H&E) and Gomori methanamine silver (GMS) stains in cyclophosphamide-treated mice given control serum 3 days after inoculation of 1–2 × 106A. fumigatus conidia (magnification, ×100). C and D depict representative lung H&E and GMS stains in cyclophosphamide-treated mice given anti-TNF-α serum 3 days after inoculation of 1–2 × 106 A. fumigatus conidia (magnification, ×100). Greater number of branching hyphae are visible in animals treated with anti-TNF. The data shown are representative of three experiments.
Effect of TNF-α neutralization on lung chitin after A. fumigatus challenge in cyclophosphamide-treated mice. Lung chitin levels in cyclophosphamide-treated mice given rabbit anti-murine TNF-α serum or control serum were determined 3 days after A. fumigatus challenge (1–2 × 106). Asp, A. fumigatus. Experimental n = 6. ∗, p < 0.05 compared with animals receiving control rabbit serum.
Effect of TNF-α neutralization on recruitment of inflammatory cells
To determine whether impaired fungal clearance in TNF-depleted animals occurred as a result of changes in lung neutrophil influx, lungs from cyclophosphamide-treated and normal mice given anti-TNF-α Ab or control serum were harvested for estimation of MPO activity, a surrogate measure of neutrophil presence. On the basis of studies showing the time of maximum expression of MPO in normal and cyclophosphamide-treated animals (not shown), lung MPO levels were evaluated 1 day after inoculation in normal animals and 2 days after inoculation in cyclophosphamide-treated animals. In normal animals treated with anti-TNF, there was a 79% decrease in lung MPO levels compared with control values at 1 day, indicating reduced neutrophil recruitment in the absence of TNF. In cyclophosphamide-treated mice, the lung MPO level was reduced by 45% in TNF-depleted animals compared with that in controls at 2 days (Fig. 6⇓).
Effect of TNF-α neutralization on lung MPO activity after A. fumigatus challenge in immunocompetent and cyclophosphamide-treated mice. Lung MPO activity was measured 1 day after A. fumigatus challenge in normal mice (1–2 × 107 conidia) and 2 days after A. fumigatus challenge in cyclophosphamide-treated mice (1–2 × 106 conidia). Asp, A. fumigatus; Cy, cyclophosphamide. Experimental n = 6. ∗, p < 0.05 compared with animals receiving control rabbit serum.
Effect of TNF-α neutralization on expression of C-X-C and C-C chemokines
Given that TNF-α has been shown to be a potent inducer of both C-X-C and C-C chemokines, levels of the C-X-C chemokines MIP-2 and KC, and of the C-C chemokines, MIP-1α and JE, were measured in lung homogenates from normal and cyclophosphamide-treated mice. On the basis of preliminary studies showing the time of maximum expression of these chemokines, levels were measured 1 day after Aspergillus inoculation in normal mice, and 3 days after Aspergillus inoculation in cyclophosphamide-treated mice. In both normal and cyclophosphamide-treated animals, there was a significant decrease in MIP-2 levels in anti-TNF-treated animals challenged with conidia compared with that in control animals receiving serum. No significant change in KC levels was detected in either group. Treatment of normal or cyclophosphamide-treated animals with anti-TNF-α serum also resulted in significant reduction in the C-C chemokines MIP-1α and JE/monocyte chemoattractant protein-1 (Fig. 7⇓ and Fig. 8⇓).
Effect of TNF-α neutralization on lung MIP-2, KC, MIP-1α, and JE levels after A. fumigatus administration in immunocompetent mice. A and B depict C-X-C and C-C chemokine levels in immunocompetent mice 1 day after A. fumigatus challenge (1–2 × 107), respectively. Experimental n = 6. ∗, p < 0.05 compared with animals receiving control rabbit serum.
Effect of TNF-α neutralization on lung MIP-2, KC, MIP-1α, and JE levels after A. fumigatus administration in cyclophosphamide-treated mice. A and B depict C-X-C and C-C chemokine levels in cyclophosphamide-treated mice 3 days after A. fumigatus challenge (1–2 × 106), respectively. Experimental n = 6. ∗, p < 0.05 compared with animals receiving control rabbit serum.
Effect of intrapulmonary delivery of rmTNF-α or TNF70–80 on survival in invasive pulmonary aspergillosis
Having demonstrated that TNF-α depletion increases fungal burden and mortality, we next examined the effect of compartmentalized TNF-α augmentation on survival in cyclophosphamide-treated animals challenged with A. fumigatus. To avoid systemic effects, we administered either rmTNF-α (1 μg/animal) or TNF70–80 (10 μg/animal) i.t. concomitant with the administration of A. fumigatus conidia (1.8 × 106). Neither rmTNF-α nor TNF70–80 administration altered survival in cyclophosphamide-treated animals when given at the time of Aspergillus challenge compared with that in animals receiving a control peptide of vehicle control (data not shown). However, significantly enhanced survival was observed in cyclophosphamide-treated mice that were pretreated with TNF70–80 i.t. 3 days before the administration of A. fumigatus compared with that in animals pretreated with either control peptide or vehicle 3 days before inoculation with A. fumigatus conidia (Fig. 9⇓).
Effect of TNF70–80 on survival in cyclophosphamide-treated mice pretreated with TNF70–80. Mice were treated with vehicle, control peptide (10 μg), or TNF70–80 (10 μg) i.t. 3 days before inoculation with A. fumigatus (1.8 × 106 conidia). ∗, p < 0.05 compared with animals receiving vehicle or control peptide. p-control, control peptide; p-TNF, TNF70–80 peptide. Experimental n = 15.
Discussion
Invasive pulmonary aspergillosis is a common and devastating complication of immunosuppression. Given its poor outcome with current therapy, the precise mechanism of the immune response to A. fumigatus is of interest, since immunomodulation may be beneficial as adjunctive therapy.
We studied the pulmonary host response to A. fumigatus in normal and immunocompromised animal models. Despite the 10-fold greater concentration of conidia used in all experiments involving immunocompetent animals, normal hosts readily cleared these large inocula without developing invasive disease. Cyclophosphamide, an alkylating agent in common clinical use as a chemotherapeutic and immunosuppressive agent, was used to render animals susceptible to invasive disease. Cyclophosphamide induces a variety of effects on immunologically active cells; these include a dose-dependent depletion of circulating neutrophils as well as alterations in the number and function of circulating and pulmonary monocyte/macrophages and various lymphocyte populations (27, 28). Although peripheral blood neutropenia is an easily measurable clinical risk factor for the development of invasive aspergillosis in humans, this susceptibility may also be related to the numerous other effects of cyclophosphamide on other immune effector cells. In our model, animals pretreated with cyclophosphamide before inoculation with Aspergillus conidia developed an acute, severe pneumonia that clinically and histologically resembled invasive pulmonary aspergillosis complicating severe immunosuppression in humans. Therefore, this model has the advantage of notable clinical relevance, but is complicated by the diverse effects of cyclophosphamide on various immune effector cells.
In this study we demonstrated that in both normal and cyclophosphamide-treated mice, TNF-α is produced in the lungs in response to the i.t. administration of A. fumigatus. In immunocompetent animals, the TNF-α level is significantly elevated by 8 h after A. fumigatus inoculation. Resident alveolar macrophages are most likely to be the source of this early and localized release of TNF, which occurs before significant recruitment of other immune effector cells occurs (data not shown). This view is supported by the fact that TNF-α is released from alveolar macrophages coincubated with A. fumigatus conidia in vitro (13). Elevated TNF-α at later time points probably represents release from various recruited cells, including neutrophils, and recruited monocyte/macrophages. Notably, neutrophils were present in the lungs of cyclophosphamide-treated mice infected with A. fumigatus while the animals had peripheral blood neutropenia. This is presumably due to preferential accumulation of neutrophils at the site of inflammation despite a reduced pool of available neutrophils.
In vivo depletion of TNF-α resulted in increased fungal burden and mortality, which occurred in association with a reduction in lung neutrophil infiltration. TNF-α is likely to contribute to host defense against A. fumigatus by several mechanisms. Although not directly chemotactic for neutrophils, TNF-α induces leukocyte and endothelial cell expression of adhesion molecules, thus influencing neutrophil trafficking in the lungs (29). In addition, our studies indicate that the neutralization of TNF-α resulted in significant reductions in both C-X-C and especially C-C chemokines. Previous studies have suggested a role for chemokines in host defense against A. fumigatus: isolated rat alveolar macrophages produce MIP-1α, MIP-2, and KC, as well as TNF-α in response to A. fumigatus conidia in vitro (30), and knockout mice lacking CCR1, a receptor for MIP-1α and RANTES, develop disseminated infection when administered A. fumigatus i.v. (31). In this context, our study provides evidence that TNF-α drives the expression of both C-X-C and C-C chemokines, including MIP-1α, JE, and MIP-2, which may in part mediate the influx and activation of neutrophils and mononuclear cells in the lungs. Reduced amounts of MIP-2 and MIP-1α may explain the reduction in early neutrophil influx (and possibly activation), increased fungal burden, and increased mortality in TNF-depleted animals. In addition, MIP-1α and JE/monocyte chemoattractant protein-1 may play important roles in mononuclear cell recruitment and activation in the setting of invasive pulmonary aspergillosis. Studies establishing causal relationships are in progress.
The intratracheal administration of TNF70–80 provided survival benefits when given 3 days before the i.t. inoculation of A. fumigatus conidia, but not when given concomitantly with conidia. These results parallel our earlier observations in murine bacterial pneumonia (22). The lack of a beneficial effect in animals treated with TNF70–80 concomitant with conidial administration may be due to several possible mechanisms. Firstly, TNF70–80 is more potent as a priming agent than as a direct activator of leukocyte respiratory burst and degranulation in vitro (20). Therefore, pretreatment with TNF70–80 may provide sufficient macrophage priming, whereas concomitant treatment may result in insufficient leukocyte activation. Alternatively, leukocyte release of proteolytic enzymes in response to infectious organisms within the airspace may lead to cytokine degradation and loss of biologic effects. Lastly, the administration of native TNF-α or TNF70–80 given at the time of fungal challenge may induce lung injury in the setting of pneumonia, thereby negating any beneficial effect on mortality.
Given that immunocompromised patients at risk of developing pulmonary invasive aspergillosis are easily identifiable, early institution of prophylactic immunomodulatory treatment is a potential application of this study. Such clinical applications require further study.
Acknowledgments
We are indebted to Dr. Thomas Moore for numerous instructive discussions, to Dr. Steven Kunkel for generously providing us with anti-TNF-α serum, and to Mrs. Mary Glass for technical assistance in performing ELISAs.
Footnotes
↵1 This work was supported in part by National Institutes of Health Grants HL57243, HL58200, and P50HL60289.
↵2 Address correspondence and reprint requests to Dr. Theodore J. Standiford, Division of Pulmonary and Critical Care Medicine, University of Michigan Medical Center, 6301 MSRB III, 1150 West Medical Center Drive, Ann Arbor, MI 48109-0360. E-mail address: tstandif{at}uv1.im.med.umich.edu
↵3 Abbreviations used in this paper: TNF70–80, TNF-α agonist peptide; MIP-1α; macrophage inflammatory protein-1α; MIP-2, macrophage inflammatory protein-2; i.t., intratracheal; MPO, myeloperoxidase; CCR-1, C-C chemokine receptor-1.
- Received July 6, 1998.
- Accepted October 8, 1998.
- Copyright © 1999 by The American Association of Immunologists
References
In this issue
