We examined the influence of endogenous GM-CSF on the course of primary and secondary pulmonary histoplasmosis. A high proportion (≥75%) of C57BL/6 mice given mAb to GM-CSF did not survive primary infection, whereas 88–94% of infected controls survived. Analysis of leukocytes revealed significantly fewer CD4+ and CD8+ cells in lungs, but not airways, of anti-GM-CSF-treated mice as compared with infected controls. However, the histopathology was similar between the two groups. Lungs of mice given mAb to GM-CSF manifested depressed levels of TNF-α, IFN-γ, and reactive nitrogen intermediates and elevated levels of IL-4 and IL-10. Administration of mAb to IL-4, to IL-10, or both restored protective immunity in GM-CSF-neutralized mice. In secondary infection, administration of mAb to GM-CSF exacerbated infection but did not alter survival over 30 days. The character of the inflammatory response was similar, and no differences were detected in Th1 or Th2 cytokine production between the two groups. Thus, endogenous GM-CSF is essential for survival in primary but not secondary infection, and blockade perturbs protective immunity. These findings reveal a new mechanism whereby GM-CSF contributes to host protection and demonstrate differences in control of primary and secondary histoplasmosis.

Human infection with the pathogenic fungus, Histoplasma capsulatum (Hc),3 is highly prevalent in North and South America, but cases have been reported worldwide. Infection is acquired via inhalation of microconidia from the soil. Upon entry into the terminal bronchioles and alveoli, these forms convert into the yeast phase that is ingested by resident phagocytes. Subsequently, the organism disseminates to organs rich in mononuclear phagocytes; replication is limited once cell-mediated immunity is activated. In the vast majority of cases, the infection is self-resolving, but the fungus establishes a dormant state within visceral organs. Reactivation of infection, which can be life-threatening, is more frequently observed in those who are receiving immunosuppressive agents or who have AIDS (1, 2).

Host control of infection with Hc requires the production of cytokines that can directly or indirectly stimulate phagocytes to limit intracellular growth. The generation of protective immunity to this microbe is associated with a dominant Th1-type response (3, 4, 5, 6). Among the endogenous cytokines in mice that unequivocally contribute to elimination of Hc yeast cells from host tissues are IL-12, IFN-γ, and TNF-α (3, 4, 5, 6, 7, 8, 9, 10). In primary infection, neutralization of any one of these three during the acute stages of infection impairs host defenses, and mice succumb to i.v. or intranasal (i.n.) challenges with yeast cells (3, 4, 5, 7, 8, 9, 10). In secondary infection, IFN-γ and TNF-α are requisite for survival of mice given yeasts i.n., but the former is dispensable in animals injected i.v. (9, 10).

The pleiotropic cytokine, GM-CSF, expresses a number of beneficial effects on host defenses. Among its many biological activities, it can enhance hemopoiesis of myeloid lineage cells, thereby causing an influx into the circulation and into inflamed tissues (11, 12). GM-CSF also promotes the microbicidal and tumoricidal capacity of macrophages (Mφ) and increases their class II MHC expression (13, 14, 15). One previous publication had reported that recombinant GM-CSF stimulated human monocyte-derived Mφ to express fungistatic activity against Hc, although the mechanism by which it exerts a propitious effect in vitro has not been identified (16). This finding raises the intriguing possibility that GM-CSF may be important in the protective immune response to this fungus. Nevertheless, virtually nothing is known regarding the influence of endogenous GM-CSF to host defenses in either primary or secondary histoplasmosis. To determine its contribution, we have analyzed the effect of mAb to GM-CSF on the protective immune response to pulmonary histoplasmosis using an established murine model induced by i.n. injection of yeast cells.

Male C57BL/6 mice, 6 wk old, were purchased from The Jackson Laboratory (Bar Harbor, ME). Athymic nude mice, 6 wk old, were purchased from the National Cancer Institute (Frederick, MD) and used to produce ascites. All animal experiments were done in accordance with the Animal Welfare Act guidelines of the National Institutes of Health.

Hc yeasts (strain G217B) were prepared as described (5). This strain is a prototypical virulent strain of this fungus (5). To produce infection in naive mice, animals were infected i.n. with 2.5 × 106 Hc yeasts in a 50 μl volume. For secondary histoplasmosis, mice were initially inoculated with 104 yeasts i.n. in a volume of 50 μl. Six to 8 wk later, previously exposed animals were rechallenged i.n. with 2.5 × 106 yeasts.

Recovery of Hc was performed as described (17). Fungal burden was expressed as mean CFU per whole organ ± SEM. The limit of detection is 102 CFU.

Ascites-derived rat anti-mouse GM-CSF mAb and rat anti-mouse IL-10 mAb were produced from the hybridoma MP1-22E9 (rat IgG2a) and JES 2A5 (rat IgG2b), respectively. The cell lines were obtained from Dr. J. Abrams (DNAX Research Institute, Palo Alto, CA). The concentration of rat IgG in ascites was assessed by ELISA and calculated by linear regression from a rat IgG (Organon Teknika, Durham, NC) standard curve. Rat anti-mouse IL-4 mAb (11B.11, rat IgG1) was obtained from the Biological Response Modifiers Program (National Cancer Institute). All Abs contained <5 pg/ml of endotoxin as determined by Limulus amebocyte lysate test (BioWhittaker, Walkersville, MD).

Mice were injected i.p. with 500 μg of mAb to GM-CSF 24 h before challenge with Hc and an equal amount 24 h postinfection. This dose was selected based on preliminary experiments demonstrating that the lungs of Hc-infected mice administered mAb to GM-CSF contained <20 pg/ml of this cytokine. Mice were given 500 μg of mAb each week thereafter. In studies with anti-IL-4 and anti-IL-10 mAb, mice received 2 mg and 1 mg, respectively, on day 0 of infection and were given 1 mg each week. Control animals received an equal amount of rat IgG concomitantly.

Lungs from infected mice (n = 5–6) were removed on days 3, 5, 7, and 14 of primary infection or days 7, 14, and 21 of secondary infection. Tissue was homogenized in 10 ml of RPMI 1640, centrifuged at 1500 × g, filter sterilized, and stored at −70°C until assayed. The protein concentration of homogenates ranged from 2.7 to 5.7 mg/ml. There were no significant differences (p > 0.05) in protein content between GM-CSF-neutralized mice and those given rat IgG at each time point assayed. Commercially available ELISA kits were used to measure IFN-γ, IL-4, IL-10, and TNF-α (Endogen, Cambridge, MA). IL-12 was assayed by sandwich ELISA (PharMingen, San Diego, CA) specific for mouse IL-12 p70 protein. The sensitivity was >100 pg/ml.

Lungs were removed and tissues were fixed in 10% formalin and embedded in paraffin blocks. Sections (5 μm) were stained with hematoxylin and eosin or with silver (Grocott) for fungal elements. Analysis of the sections was performed in a “blinded” fashion.

BAL was performed on day 7 of primary and days 14 and 21 of secondary infection to obtain inflammatory cells. The trachea was exposed and intubated using a 1.7-mm OD polyethylene catheter. Cells were harvested by instilling PBS free of Ca2+ and Mg2+ in 1-ml aliquots. Approximately 5 ml of lavage fluid were retrieved per mouse. To isolate mononuclear cells from lungs, mice were sacrificed and lungs flushed with 20 ml of HBSS by inserting a catheter into the right heart. The lungs were excised and teased apart with forceps and homogenized by sequential passage through 16-, 18-, and 20-gauge needles. Mononuclear cells were isolated by separation on a 40–70% Percoll (Pharmacia, Piscataway, NJ) gradient (6).

BAL and lung cells were adjusted to 5 × 105/200 μl in HBSS containing 10% FBS and 0.02% sodium azide and stained with 0.5 μg of one of the following FITC-labeled mAbs (PharMingen): anti-CD4 (clone RM4-5), anti-CD8 (clone 53-6.7), anti-Ly-6G (Gr-1), (clone RB6–8C5, which recognizes polymorphonuclear cells), anti-I-Ab (clone 25-9-17), Mac-3 (clone M3/84, detects tissue Mφ), or isotype-matched rat IgG mAb. The samples were washed and fixed in 2% paraformaldehyde until analyzed on a flow cytometer.

BAL cells from rat IgG-treated and mAb to anti-GM-CSF-treated mice (n = 5/group) were obtained on days 3, 5, and 7 of primary and secondary infection. After washing twice, cells were plated at 1.5 × 105 cells per well in 48-well plates in DMEM supplemented with 10% FBS. Nonadherent cells were removed after 2 h and monolayers were stimulated with LPS (Sigma, St. Louis, MO) (1 μg/ml) plus IFN-γ (100 ng/ml) or IFN-γ (100 ng/ml) plus TNF-α (100 ng/ml). Supernatants were collected 48 h after seeding, and nitrite was measured using Cayman’s nitrate/nitrite assay kit (Alexis, San Diego, CA). Data are presented as micromolars of NO2 .

Student’s t test was used to compare groups if the data achieved normality, otherwise the Wilcoxon rank sum test was used. Survival data was analyzed using the log rank test.

Mice inoculated i.n. with Hc were treated with mAb to GM-CSF or rat IgG and observed over a 30-day period. The former began to appear ill on days 10–11 postinfection. They were noted to have decreased mobility, ruffled fur, and huddling. These symptoms progressed over the next several days until animals became moribund. In two experiments, survival was strikingly altered by blockade of endogenous GM-CSF as compared with infected controls (p < 0.0001) (Fig. 1).

FIGURE 1.

Survival curve of H. capsulatum-infected mice given mAb to GM-CSF or control Ab. Groups of naive mice (n = 8–10) were administered rat IgG or mAb to GM-CSF and infected with 2.5 × 106H. capsulatum yeasts and monitored for survival for 30 days.

FIGURE 1.

Survival curve of H. capsulatum-infected mice given mAb to GM-CSF or control Ab. Groups of naive mice (n = 8–10) were administered rat IgG or mAb to GM-CSF and infected with 2.5 × 106H. capsulatum yeasts and monitored for survival for 30 days.

Close modal

In subsequent experiments, we determined the fungal burden in lungs and spleens of infected mice given mAb to GM-CSF or an equal amount of rat IgG. The spleens were examined because they are a site of dissemination for this fungus, and we sought to determine whether neutralization of endogenous GM-CSF had an impact upon both visceral and lymphoid organs. In the first experiment, mice were sacrificed on day 7 of infection, and their lungs and spleens were cultured for Hc (Fig. 2). Lungs of GM-CSF-neutralized mice contained 1.1 log10 more CFU (p < 0.001) than control lungs, but the burden of Hc in spleens was similar (p > 0.05). In a second experiment, the difference between the mAb to GM-CSF-treated mice and controls in recovery of Hc at day 7 of infection was ∼0.5 log10 (p < 0.05); the number of Hc CFU in spleens was similar between the two groups. In this second experiment, we sacrificed mice at day 14 of infection and analyzed the number of Hc CFU in lungs and spleens. The CFU in lungs of recipients of mAb to GM-CSF were 2.5 log10 higher than in infected controls (p < 0.001), whereas the quantity of Hc CFU in spleens did not differ significantly between groups (p > 0.05). Thus, neutralization of endogenous GM-CSF was associated with impaired clearance of Hc and failure to control the infectious process.

FIGURE 2.

Fungal burden in lungs and spleens of naive mice given mAb to GM-CSF or rat IgG. Groups of mice (n = 5–6) were administered mAb to GM-CSF or rat IgG and infected with yeast cells. At days 7 (Expts. 1 and 2) and day 14 (Expt. 2), the lungs and spleens of mice were assayed for H. capsulatum CFU. The data represent the mean ± SEM. ∗, p < 0.05; ∗∗, p < 0.01.

FIGURE 2.

Fungal burden in lungs and spleens of naive mice given mAb to GM-CSF or rat IgG. Groups of mice (n = 5–6) were administered mAb to GM-CSF or rat IgG and infected with yeast cells. At days 7 (Expts. 1 and 2) and day 14 (Expt. 2), the lungs and spleens of mice were assayed for H. capsulatum CFU. The data represent the mean ± SEM. ∗, p < 0.05; ∗∗, p < 0.01.

Close modal

A concern in these experiments was that blockade of endogenous GM-CSF might impair mobilization of myeloid cells into inflamed tissues and thus render mice more susceptible to invasion by endogenous flora or airborne pathogens such as Aspergillus sp. If this event transpired, mice may have succumbed to bacterial rather than HC infection. To exclude this possibility, a portion of the homogenates was cultured on nonselective medium and observed for bacterial or saprobic fungal growth. All samples were devoid of aerobic bacterial or fungal contamination. In addition, three normal C57BL/6 mice were treated with mAb to GM-CSF for 1 mo using the identical regimen described in Materials and Methods. These mice remained healthy for the observation period and cultures of lungs and spleens did not reveal any bacterial or fungal growth.

In a separate set of studies, the course of secondary infection in mice administered mAb to GM-CSF was examined. Mice were inoculated with 104 yeasts and 6–8 wk later were injected with mAb to GM-CSF or rat IgG and rechallenged with 2.5 × 106 yeasts. All mice survived a 30-day observation period (data not shown). The fungal burden in the lungs of GM-CSF-neutralized mice was similar to that of infected controls on day 7 (p > 0.05), but exceeded that of controls on days 14 (p < 0.01) and 21 (p < 0.05) (Fig. 3). In spleens, the only significant difference was found on day 14 of experiment 1 in which organs from infected controls contained far less Hc yeasts (p < 0.01) than mAb to GM-CSF recipients (Fig. 3). Thus, neutralization of endogenous GM-CSF in secondary infection blunted the clearance of Hc yeasts, especially in the lungs, but did not lead to uncontrolled growth of the organism.

FIGURE 3.

Fungal recovery of H. capsulatum from lungs and spleens of mice with secondary infection and treated with mAb to GM-CSF. Groups of immunized mice (n = 5–6) were given either mAb to GM-CSF or rat IgG and infected with yeast cells. At days 7, 14, and 21, lungs and spleens were assayed for H. capsulatum CFU. The data represent the mean ± SEM. ∗, p < 0.05; ∗∗∗, p < 0.01; ∗∗∗, p < 0.001.

FIGURE 3.

Fungal recovery of H. capsulatum from lungs and spleens of mice with secondary infection and treated with mAb to GM-CSF. Groups of immunized mice (n = 5–6) were given either mAb to GM-CSF or rat IgG and infected with yeast cells. At days 7, 14, and 21, lungs and spleens were assayed for H. capsulatum CFU. The data represent the mean ± SEM. ∗, p < 0.05; ∗∗∗, p < 0.01; ∗∗∗, p < 0.001.

Close modal

We sought to determine whether administration of mAb to GM-CSF was associated with perturbations in the production of other cytokines known to be influential in primary and/or secondary infection. IFN-γ, TNF-α, IL-4, IL-10, and IL-12 were analyzed because they augment (IFN-γ, TNF-α, and IL-12) or impair (IL-4 and IL-10) host defenses to this fungus (3, 4, 5, 6, 7, 8, 9, 10). Mice were treated with mAb to GM-CSF or rat IgG and infected with 2.5 × 106 yeasts; at days 3, 5, 7, and 14 postinfection, cytokine levels were measured in lung homogenates. When compared with controls, the levels of TNF-α in lungs of treated animals was significantly less than controls at days 3 and 5 of infection (p < 0.01), but was higher than controls on days 7 and 14 postinfection (Fig. 4). IFN-γ levels in GM-CSF-neutralized mice were similar on day 3, but exceedingly less than controls on days 5 and 7 (p < 0.01, day 5; and p < 0.04, day 7) (Fig. 4). By day 14, they were similar between the groups. The quantity of IL-12 was comparable between controls and mAb to GM-CSF recipients, except on day 7 in which IL-12 was higher (p < 0.01) in GM-CSF-neutralized mice. IL-4 levels in lungs of these animals exceeded controls (p < 0.01) on days 5, 7, and 14, and IL-10 levels were greater than controls on days 7 and 14 (p < 0.05, day 7; and p < 0.01, day 14).

FIGURE 4.

Cytokine profiles in lungs of infected mice given anti-GM-CSF or rat IgG. During primary and secondary infection, lungs of mice (n = 5–6) were assayed for cytokine levels by ELISA of tissue homogenates. Results for primary infection are depicted on the left and those for secondary infection on the right. One of two experiments is shown. The data represent the mean ± SEM. ∗, p < 0.05; ∗∗, p < 0.01.

FIGURE 4.

Cytokine profiles in lungs of infected mice given anti-GM-CSF or rat IgG. During primary and secondary infection, lungs of mice (n = 5–6) were assayed for cytokine levels by ELISA of tissue homogenates. Results for primary infection are depicted on the left and those for secondary infection on the right. One of two experiments is shown. The data represent the mean ± SEM. ∗, p < 0.05; ∗∗, p < 0.01.

Close modal

In secondary infection, cytokine levels in the lungs were examined on days 7, 14, and 21 postinfection. IFN-γ and TNF-α levels did not differ (p > 0.05) between the two groups. IL-4 and IL-10 were similar on days 7 and 21, but significantly lower (p < 0.009) in mAb to GM-CSF recipients on day 14 (Fig. 4).

Lung tissue of infected controls and mice given mAb to GM-CSF was examined at days 7 and 14 after primary infection to determine whether there were differences in the inflammatory response to Hc in these respective groups. At day 7, mild to moderate perivascular lymphoid cuffing was observed in lung parenchyma of infected controls. An admixture of neutrophils and granulomatous inflammation was present involving between 30 and 90% of the lung tissue. In anti-GM-CSF-treated animals, the inflammatory response was quite similar. There was mild perivascular lymphoid cuffing and a pyogranulomatous response that involved 20–80% of lung parenchyma. At day 14, the inflammatory response in infected controls had shifted to a strictly granulomatous reaction that affected between 40 and 90% of the lung tissue. In addition, diffuse lymphocytic infiltration of tissue was present. In mice infected for 14 days and treated with mAb to GM-CSF, the tissue reaction remained pyogranulomatous and occluded between 20 and 40% of the pulmonary parenchyma. No lymphocytic infiltration was observed (data not shown).

Lung tissue from rechallenged mice was examined on days 7, 14, and 21 postinfection. In infected controls, there was severe peribronchial and perivascular infiltration of an equal number of lymphocytes and mononuclear phagocytes on day 7. Neutrophils were sparse and scattered throughout the inflamed tissue. The histological picture was similar at this time in GM-CSF-neutralized mice. On days 14 and 21, there was a range of inflammation in both groups extending from mild infiltration of lymphocytes and mononuclear phagocytes to massive infiltration of these cells in the perivascular and peribronchial areas. The extent and severity of inflammation did not differ between the two groups (data not shown).

Histopathological examination reveals the architecture of inflammation, but does not provide information concerning the precise composition of cells involved in this process. Therefore, we determined if administration of mAb to GM-CSF altered the numbers of cells in lungs and lavage fluid in primary infection. The proportion of CD4+ and CD8+ cells, Mφ (Mac-3+), neutrophils (Gr-1+), and I-A+ cells was determined by FACS, and the absolute numbers calculated. These cell populations were selected because of their know influence on host defenses to Hc (10, 18, 19). The number of CD4+ and CD8+ cells was significantly less (p = 0.03) in lungs of mAb to GM-CSF recipients than in infected controls (Table I). Numbers of Mφ, neutrophils, and I-A+ cells did not differ (p > 0.05) between the two groups.

Table I.

Enumeration of cell populations in lungs and BAL of Hc-infected mice given rat IgG or anti-GM-CSFa

Cell PopulationCell No. in Lungs (×106)Cell No. in BAL (×104)
Rat IgG (n = 6)Anti-GM-CSF (n = 6)Rat IgG (n = 4)Anti-GM-CSF (n = 6)
Total cells 8.0 ± 0.9 10.7 ± 2.1 26.5 ± 1.6 48.4 ± 1.2 
CD4+ 0.9 ± 0.2 0.3 ± 0.1 1.7 ± 0.7 2.1 ± 0.5 
CD8+ 1.0 ± 0.3 0.3 ± 0.1 2.9 ± 1.0 3.6 ± 0.7 
GR1+ 1.8 ± 0.4 6.0 ± 1.9 13.5 ± 0.3 24.7 ± 0.7 
Mac-3+ 1.1 ± 0.3 3.1 ± 1.3 2.1 ± 0.6 6.5 ± 0.1 
I-A+ 3.7 ± 0.7 3.7 ± 1.2 2.8 ± 0.7 9.9 ± 0.3 
Cell PopulationCell No. in Lungs (×106)Cell No. in BAL (×104)
Rat IgG (n = 6)Anti-GM-CSF (n = 6)Rat IgG (n = 4)Anti-GM-CSF (n = 6)
Total cells 8.0 ± 0.9 10.7 ± 2.1 26.5 ± 1.6 48.4 ± 1.2 
CD4+ 0.9 ± 0.2 0.3 ± 0.1 1.7 ± 0.7 2.1 ± 0.5 
CD8+ 1.0 ± 0.3 0.3 ± 0.1 2.9 ± 1.0 3.6 ± 0.7 
GR1+ 1.8 ± 0.4 6.0 ± 1.9 13.5 ± 0.3 24.7 ± 0.7 
Mac-3+ 1.1 ± 0.3 3.1 ± 1.3 2.1 ± 0.6 6.5 ± 0.1 
I-A+ 3.7 ± 0.7 3.7 ± 1.2 2.8 ± 0.7 9.9 ± 0.3 
a

Mice were lavaged on day 7 of infection. Total cells were enumerated and the percent expressing the surface phenotype was determined by flow cytometric analysis. The number of cells was calculated by multiplying total cell number by percent of cell expressing a phenotype. The data represent the mean ± SEM.

In separate experiments, we enumerated the cell populations in lavage fluid. The only significant difference between the two groups was in the absolute number of Mac-3+ cells in which more were found (p = 0.03) in the lavage of mice given mAb to GM-CSF as compared with controls (Table I).

We also analyzed the phenotype of cells in lungs and BAL from mice rechallenged with Hc. Mice infected for 14 and 21 days were examined because those time periods were associated with alterations in CFU. As shown in Table II, no differences in the numbers of CD4+, CD8+, Gr-1+, Mac-3+, and I-A+ cells were detected in either lung parenchyma or BAL.

Table II.

Cell populations in mice rechallenged with Hc and given mAb to GM-CSF or rat IgGa

Day of InfectionCell PopulationCell No. in Lungs (×105)Cell no. in BAL (×105)
Rat IgG (n = 4)Anti-GM-CSF (n = 4)RatIgG (n = 5)Anti-GMCSF (n = 5)
14 Total cells 81.9 ± 31.3 110.0 ± 20.7 17.2 ± 0.44 18.0 ± 3.4 
 CD4+ 43.7 ± 13.5 48.0 ± 11.3 11.7 ± 3.1 10.5 ± 2.4 
 CD8+ 12.1 ± 7.4 18.9 ± 8.2 1.7 ± 0.4 2.8 ± 0.9 
 GR1+ 8.8 ± 3.2 9.0 ± 2.8 2.1 ± 0.9 2.4 ± 1.0 
 Mac-3+ 13.6 ± 6.9 12.5 ± 4.2 0.3 ± 0.01 0.3 ± 0.01 
 I-A+ 40.9 ± 21.0 56.0 ± 2.9 5.0 ± 0.2 4.9 ± 0.6 
      
21 Total cells 18.8 ± 3.7 22.2 ± 6.5 9.7 ± 0.2 7.4 ± 0.8 
 CD4+ 7.7 ± 2.0 9.3 ± 2.7 6.1 ± 0.8 5.1 ± 0.1 
 CD8+ 1.7 ± 0.6 3.4 ± 1.3 0.8 ± 0.03 0.6 ± 0.02 
 GR1+ 1.8 ± 0.6 1.8 ± 0.4 0.3 ± 0.1 0.31 ± 0.01 
 Mac-3+ 1.9 ± 0.3 1.8 ± 0.6 0.2 ± 0.04 0.1 ± 0.02 
 I-A+ 10.6 ± 2.5 15.7 ± 5.2 2.7 ± 0.4 1.8 ± 0.5 
Day of InfectionCell PopulationCell No. in Lungs (×105)Cell no. in BAL (×105)
Rat IgG (n = 4)Anti-GM-CSF (n = 4)RatIgG (n = 5)Anti-GMCSF (n = 5)
14 Total cells 81.9 ± 31.3 110.0 ± 20.7 17.2 ± 0.44 18.0 ± 3.4 
 CD4+ 43.7 ± 13.5 48.0 ± 11.3 11.7 ± 3.1 10.5 ± 2.4 
 CD8+ 12.1 ± 7.4 18.9 ± 8.2 1.7 ± 0.4 2.8 ± 0.9 
 GR1+ 8.8 ± 3.2 9.0 ± 2.8 2.1 ± 0.9 2.4 ± 1.0 
 Mac-3+ 13.6 ± 6.9 12.5 ± 4.2 0.3 ± 0.01 0.3 ± 0.01 
 I-A+ 40.9 ± 21.0 56.0 ± 2.9 5.0 ± 0.2 4.9 ± 0.6 
      
21 Total cells 18.8 ± 3.7 22.2 ± 6.5 9.7 ± 0.2 7.4 ± 0.8 
 CD4+ 7.7 ± 2.0 9.3 ± 2.7 6.1 ± 0.8 5.1 ± 0.1 
 CD8+ 1.7 ± 0.6 3.4 ± 1.3 0.8 ± 0.03 0.6 ± 0.02 
 GR1+ 1.8 ± 0.6 1.8 ± 0.4 0.3 ± 0.1 0.31 ± 0.01 
 Mac-3+ 1.9 ± 0.3 1.8 ± 0.6 0.2 ± 0.04 0.1 ± 0.02 
 I-A+ 10.6 ± 2.5 15.7 ± 5.2 2.7 ± 0.4 1.8 ± 0.5 
a

Mice were lavaged on day 14 and 21 of infection. Total cells were enumerated, and the percent expressing the surface phenotype was determined as described in Table I.

Although the number of I-A-bearing cells was similar, it was possible that surface expression differed between the groups. Therefore, we analyzed the mean fluorescence intensity of I-A expression on cells from lungs or lavage of infected controls and mAb to GM-CSF-treated animals. I-A expression did not differ significantly (p > 0.05) on day 7 of primary infection and days 14 and 21 of secondary infection between the two groups of mice (data not shown).

NO is an important mediator of host resistance to primary, but not secondary, Hc infection (9, 10). Therefore, we examined RNI production by alveolar Mφ from infected controls and mice given mAb to GM-CSF. Groups of mice were treated with rat IgG or with mAb to GM-CSF and at 3, 5, and 7 days postinfection, and alveolar Mφ were harvested, stimulated with LPS plus IFN-γ or TNF-α plus IFN-γ, and assayed for NO2. Alveolar Mφ from mice administered rat IgG released elevated quantities of RNI (p < 0.01) in response to LPS plus IFN-γ at days 3, 5, and 7 as compared with unstimulated cells (Fig. 5). In contrast, the release of RNI by cells exposed to LPS and IFN-γ in mice treated with mAb to GM-CSF did not differ significantly (p > 0.05) from that of unstimulated controls on days 3 and 5 postinfection. By day 7, RNI production by alveolar Mφ incubated with LPS plus IFN-γ was greater than that (p < 0.05) of unstimulated cells (Fig. 5). Similar results were obtained when TNF-α plus IFN-γ was used as a stimulus (data not shown).

FIGURE 5.

NO2 levels in from infected mice given mAb to GM-CSF or rat IgG. Supernatants from BAL cells obtained from infected mice (n = 5–6) given mAb to GM-CSF or rat IgG were incubated in the presence of absence of LPS plus IFN-γ and assayed for NO2. Results are depicted as the mean ± SEM at each time point. One of two experiments is depicted. Similar results were obtained if TNF-α plus IFN-γ was used as a stimulus (data not shown). ∗, p < 0.05; ∗∗, p < 0.01.

FIGURE 5.

NO2 levels in from infected mice given mAb to GM-CSF or rat IgG. Supernatants from BAL cells obtained from infected mice (n = 5–6) given mAb to GM-CSF or rat IgG were incubated in the presence of absence of LPS plus IFN-γ and assayed for NO2. Results are depicted as the mean ± SEM at each time point. One of two experiments is depicted. Similar results were obtained if TNF-α plus IFN-γ was used as a stimulus (data not shown). ∗, p < 0.05; ∗∗, p < 0.01.

Close modal

Because the levels of IL-4 and IL-10 were raised compared with controls during the course of primary infection, we determined if the elevated levels contributed to the impairment in the protective immune response. Groups of mice received mAb to GM-CSF and either rat IgG, mAb to IL-4, mAb to IL-10, or both and were evaluated for burden of infection at 1 wk and survival over 30 days. Hc CFU in lungs of GM-CSF-neutralized mice given mAb to IL-4, IL-10, or IL-4 plus IL-10 was significantly less (p ≤ 0.04) than those given mAb to GM-CSF alone (Fig. 6,A). Fungal recovery in lungs of GM-CSF-neutralized mice administered mAb to IL-4, IL-10, or both was similar to that of the infected controls (p > 0.05). In spleens, mice given both mAb to GM-CSF and IL-4 contained less CFU (p < 0.03) than recipients of mAb to GM-CSF alone or infected controls (Fig. 6 B).

FIGURE 6.

Effect of treatment with anti-IL-4 mAb, anti-IL-10 mAb, or both on the course of Hc infection in mice given mAb to GM-CSF. Groups of mice were infected with Hc and treated with either rat IgG, mAb to GM-CSF, or mAb to GM-CSF plus mAb to IL-4, mAb to IL-10, or both. At 1 wk, the burden of infection was assessed in lungs (A) and spleens (B) from a portion of the mice (n = 6). The data are presented and mean ± SEM. ∗, p ≤ 0.04 from mice given mAb to GM-CSF alone. ∗∗, p < 0.03 vs anti-GM-CSF-treated mice and infected controls. The remainder (n = 8) were followed for 30 days to monitor survival (C). One of two experiments is shown.

FIGURE 6.

Effect of treatment with anti-IL-4 mAb, anti-IL-10 mAb, or both on the course of Hc infection in mice given mAb to GM-CSF. Groups of mice were infected with Hc and treated with either rat IgG, mAb to GM-CSF, or mAb to GM-CSF plus mAb to IL-4, mAb to IL-10, or both. At 1 wk, the burden of infection was assessed in lungs (A) and spleens (B) from a portion of the mice (n = 6). The data are presented and mean ± SEM. ∗, p ≤ 0.04 from mice given mAb to GM-CSF alone. ∗∗, p < 0.03 vs anti-GM-CSF-treated mice and infected controls. The remainder (n = 8) were followed for 30 days to monitor survival (C). One of two experiments is shown.

Close modal

The remaining mice were observed for ability to survive infection. All controls survived for 30 days; 87.5% of the mice given mAb to GM-CSF expired (Fig. 6 C). All mice in the groups given either mAb to IL-4 or to IL-10 or both survived. At 30 days, all the surviving mice were sacrificed, and the quantity of Hc CFU was assessed in lungs and spleens. The organs from all the groups contained <103 CFU. Thus, survival was associated with clearance of the fungus.

We have demonstrated that endogenous GM-CSF is a critical cytokine in the generation of an optimal protective immune response to pulmonary challenge with Hc. In primary infection, neutralization of it was associated with an increased fungal burden in lungs, and a high proportion of animals failed to control infection. In secondary infection, neutralization of GM-CSF impaired elimination of the fungus only on days 14 and 21. Nevertheless, there was a sharp decline in CFU between these time points, and mice survived for 30 days postinfection. Thus, the impact of neutralizing endogenous GM-CSF was much more pronounced on the nascent immune response rather than on memory.

Because GM-CSF is pivotal in generation of myeloid lineage cells (11, 12), a likely explanation for the observed effects of mAb to GM-CSF on protective immunity was that blockade reduced recruitment of phagocytes. This assumption was not supported by the data. In either primary or secondary infection, the numbers of neutrophils and Mφ in lungs and lavage fluid were similar or increased in mice given mAb to GM-CSF as compared with controls. Furthermore, the organization of the inflammatory response, as manifested by histopathology, was not altered by administration of mAb to GM-CSF.

Although ingress of phagocytes was not decreased, the number of CD4+ and CD8+ cells in lungs was sharply lower in GM-CSF-neutralized mice as compared with controls during primary infection. In Hc infection, CD4+ cells are the major generators of IFN-γ, and this cytokine is necessary for survival in primary histoplasmosis (17, 18). A reduction in CD4+ cells can account, in part, for the diminished IFN-γ production in lungs of mice given mAb to GM-CSF. CD8+ cells also contribute to protective immunity but in an IFN-γ-independent manner (17). The decrease in both populations likely contributed to weakened host defenses. Because alteration in numbers of T cells was detected only in lung, the influence of endogenous GM-CSF on the influx of inflammatory cells in infected parenchyma differs from that of airways.

Neutralization of GM-CSF was accompanied by aberrant cytokine generation in the lung. Levels of either TNF-α or IFN-γ or both were reduced during the early stages of infection (≤7 days), but returned to control levels by days 7 and 14, respectively. Endogenous IFN-γ and TNF-α are unequivocally necessary for controlling primary infection and both must be present during the acute phase of infection in order that the host eliminate Hc (3, 4, 9, 10, 17). Decreases in the levels of these two cytokines following neutralization of GM-CSF contributed to the exacerbation of infection. Although the low levels of IFN-γ could be attributed partially to fewer CD4+ and CD8+ cells, it was not associated with depressed amounts of IL-12, which is pivotal for generation of IFN-γ in Hc infection (3, 4).

Impaired generation of RNI by alveolar Mφ incubated with either of two potent stimuli, IFN-γ plus LPS or with TNF-α plus IFN-γ, was detected in mice given mAb to GM-CSF. The failure to release RNI during the acute phase most likely added to the progressive nature of the infection because NO is required for resolution of primary, but not secondary, disease (10). Either alone or in concert, TNF-α and IFN-γ are important stimuli for release of RNI (20). In Hc infection, neutralization of TNF-α is associated with poor RNI generation and progressive infection (9). Impaired RNI production in GM-CSF-neutralized animals can be a result of the diminished levels of IFN-γ and/or TNF-α. However, GM-CSF also can independently stimulate the production of NO (21, 22). Thus, neutralization of GM-CSF in conjunction with decreases in TNF-α and IFN-γ account for the blunted release of RNI.

A possible explanation for the unrestrained growth of Hc in mice given mAb to GM-CSF is diminished generation of oxygen intermediates. GM-CSF, TNF-α, and IFN-γ induce an oxidative burst by phagocytes (12, 23, 24, 25). Neutralization of endogenous GM-CSF combined with depressed production of TNF-α and IFN-γ would blunt release of oxygen intermediates. This explanation is unlikely because Hc yeasts can replicate in the presence of a vigorous oxidative burst, and in murine Mφ, yeasts inhibit the respiratory burst (26, 27).

The lungs of mice given mAb to GM-CSF manifested markedly elevated IL-4 and IL-10 levels, two cytokines that exacerbate histoplasmosis (4, 9). Administration of mAb to IL-4 or to IL-10 to GM-CSF-neutralized mice resulted in a reduction of fungal burden in lungs at 1 wk and markedly improved survival. Thus, IL-4 and IL-10 independently exert an inimical effect on the course of infection in mice given mAb to GM-CSF.

One consideration for the perturbations in cytokine production in lungs of mice given mAb to GM-CSF is the increased burden of Hc. This contention is not supported by previous data in which large fungal burdens of Hc-infected animals are not consistently associated with modulation of IFN-γ, TNF-α, IL-4, or IL-10 levels in lungs (3, 9). Furthermore, on days 14 and 21 of secondary infection, the increased fungal burden of mice given mAb to GM-CSF was not accompanied by alterations in any of the four cytokines referenced above. Thus, no direct relationship exists between fungal burden and cytokine levels in the lungs.

GM-CSF is known to interact with IFN-γ and TNF-α. In GM-CSF-knockout mice exposed to LPS, levels of circulating IFN-γ are less than that of controls, but TNF-α levels were similar (28). In this model system, there is a GM-CSF-IFN-γ axis, but not an interaction between GM-CSF and TNF-α. Alternatively, exogenous GM-CSF enhances TNF-α transcription and TNF-α induces GM-CSF synthesis (29, 30). In experimental primary Hc infection, there exists a link between endogenous GM-CSF production and IFN-γ and TNF-α generation.

The findings also document a connection between endogenous GM-CSF and IL-4 and IL-10 in primary infection. Others have reported that IL-4 and IL-10 can inhibit monocyte production of GM-CSF in vitro (31), but there is virtually no information regarding the ability of GM-CSF to regulate IL-4 or IL-10 generation. The elevated IL-4 and IL-10 levels at days 5 and 7 may have been caused by a decrease in IFN-γ, but this possibility is unlikely because IFN-γ-deficient mice that were infected with Hc did not manifest absolute increases in IL-4 levels (3, 5).

Mice congenitally deficient in GM-CSF by homologous recombination might complement the studies performed with mAb neutralization (32). These mice accumulate abundant amounts of surfactant lipids and protein in the alveolar spaces by an early age and develop a disease that mimics alveolar proteinosis. This condition can alter normal lung physiology and air exchange. Accordingly, the preexisting lung pathology and compromised lung function in these mice presents a barrier to their use for Hc infection.

In secondary infection, treatment with mAb to GM-CSF only delayed clearance of the fungus, but did not alter survival of mice. The impaired clearance was not associated with marked disturbances in IFN-γ, TNF-α, IL-4, or IL-10 generation. No differences in either the histopathology or the cellular composition of inflammation were detected between controls and mAb to GM-CSF recipients. Neutralization of GM-CSF in secondary infection does not produce the striking perturbations in immunity as observed in primary histoplasmosis. Hence, compensatory mechanisms must exist to control secondary infection.

In summary, endogenous GM-CSF is important in the formation of the protective immune response to Hc. These data indicate that there are complex networks involved in Hc infection and that blockade of GM-CSF in primary infection produces a plethora of perturbations of the immunity that culminate in the death of the animal. The findings do establish that GM-CSF production is closely intertwined in the signaling events that lead to up-regulation of Th1 and dampening of Th2 cytokines. Furthermore, the results highlight differences in the contribution of endogenous GM-CSF to primary and secondary infection.

1

This work was supported by Grants AI-42747 and AI-34361 from the National Institutes of Health and by the Veterans Affairs Hospital.

3

Abbreviations used in this paper: Hc, Histoplasma capsulatum; BAL, bronchoalveolar lavage; i.n., intranasal; Mφ, macrophage; RNI, reactive nitrogen intermediates.

1
Deepe, G. S., Jr, W. E. Bullock.
1992
. Histoplasmosis: a granulomatous inflammatory response. J. I. Gallin, Jr, and I. M. Goldstein, Jr, and R. Snyderman, Jr, eds.
Inflammation: Basic Principles and Clinical Correlates
943
Raven Press, New York.
2
Graybill, J. R..
1988
. Histoplasmosis and AIDS.
J. Infect. Dis.
158
:
623
3
Allendoerfer, R., G. P. Boivin, G. S. Deepe, Jr.
1997
. Modulation of immune responses in murine pulmonary histoplasmosis.
J. Infect. Dis.
175
:
905
4
Zhou, P., M. C. Sieve, J. Bennett, K. J. Kwon-Chung, R. P. Tewari, R. A. Seder.
1995
. IL-12 prevents mortality in mice infected with Histoplasma capsulatum through induction of IFN-γ.
J. Immunol.
155
:
785
5
Allendoerfer, R., G. S. Deepe, Jr.
1997
. Intrapulmonary response to Histoplasma capsulatum in γ interferon knockout mice.
Infect. Immun.
65
:
2564
6
Cain, J. A., G. S. Deepe, Jr.
1998
. Evolution of the primary immune response to Histoplasma capsulatum in murine lung.
Infect. Immun.
66
:
1473
7
Smith, J. G., D. M. Magee, D. M. Williams, J. R. Graybill.
1990
. Tumor necrosis factor-α plays a role in host defense against Histoplasma capsulatum.
J. Infect. Dis.
162
:
1349
8
Wu-Hsieh, B. A., G. S. Lee, M. Franco, F. M. Hofman.
1992
. Early activation of splenic macrophages by tumor necrosis factor α is important in determining the outcome of experimental histoplasmosis in mice.
Infect. Immun.
60
:
4230
9
Allendoerfer, R., G. S. Deepe, Jr.
1998
. Blockade of endogenous TNF-α exacerbates primary and secondary pulmonary histoplasmosis by differential mechanisms.
J. Immunol.
160
:
6072
10
Zhou, P., G. Miller, R. A. Seder.
1998
. Factors involved in regulating primary and secondary immunity to infection with Histoplasma capsulatum: TNF-α plays a critical role in maintaining secondary immunity in the absence of IFN-γ.
J. Immunol.
160
:
1359
11
Metcalf, D..
1991
. Control of granulocytes and macrophages: molecular, cellular, and clinical aspects.
Science
254
:
529
12
Lieschke, G. J., A. W. Burgess.
1992
. Granulocyte colony-stimulating factor and granulocyte-macrophage colony-stimulating factor.
N. Engl. J. Med.
327
:
28
13
Grabstein, K. H., D. L. Urdal, R. J. Tushinski, D. Y. Mochizuki, V. L. Price, M. A. Cantrell, S. Gillis, P. J. Conlon.
1986
. Induction of macrophage tumoricidal activity by granulocyte-macrophage colony-stimulating factor.
Science
232
:
506
14
Reed, S. G., C. F. Nathan, D. L. Pihl, P. Rodricks, K. Shanebeck, P. J. Conlon, K. H. Grabstein.
1987
. Recombinant granulocyte/macrophage colony-stimulating factor activates macrophages to inhibit Trypansoma cruzi and release hydrogen peroxide: comparison with interferon-γ.
J. Exp. Med.
166
:
1734
15
Fischer, H.-G., S. Frosch, K. Reske, A. B. Reske-Kunz.
1988
. Granulocyte-macrophage colony-stimulating factor activates macrophages derived from bone marrow cultures to synthesis of MHC class II molecules and to augmented antigen presentation function.
J. Immunol.
141
:
3882
16
Newman, S. L., L. Gootee.
1992
. Colony-stimulating factors activate human macrophages to inhibit intracellular growth of Histoplasma capsulatum yeasts.
Infect. Immun.
60
:
4593
17
Allendörfer, R. G. D. Brunner, G. S. Deepe, Jr.
1999
. Complex requirements for nascent and memory immunity in pulmonary histoplasmosis.
J. Immunol.
162
:
7389
18
Gomez, A. M., W. E. Bullock, C. L. Taylor, G. S. Deepe, Jr.
1988
. Role of L3T4+ T cells in host defense against Histoplasma capsulatum.
Infect. Immun.
56
:
1685
19
Deepe, G. S., Jr.
1994
. Role of CD8+ T cells in host resistance to systemic infection with Histoplasma capsulatum in mice.
J. Immunol.
152
:
3491
20
MacMicking, J., Q.-W. Xie, C. Nathan.
1997
. Nitric oxide and macrophage function.
Annu. Rev. Immunol.
15
:
323
21
Hill, A. D., H. Naarma, J. Shou, S. E. Calvano, J. M. Daly.
1995
. Antimicrobial effects of granulocyte-macrophage colony-stimulating factor in protein-energy malnutrition.
Arch. Surg.
130
:
1273
22
Blau, H., S. Riklis, J. F. Van Iwaarden, F. X. McCormack, M. Kalina.
1997
. Nitric oxide production by rat alveolar macrophages can be modulated in vitro by surfactant protein A.
Am. J. Physiol.
272
:
L1198
23
Sarmento, A., R. Appelberg.
1996
. Involvement of reactive oxygen intermediates in tumor necrosis factor α-dependent bacteriostasis of Mycobacterium avium.
Infect. Immun.
64
:
3224
24
Li, Y., A. Ferrante, A. Poulos, D. P. Harvey.
1996
. Neutrophil oxygen radical generation: synergistic responses to tumor necrosis factor and mono/polyunsaturated fatty acids.
J. Clin. Invest.
97
:
1605
25
Billiau, A..
1996
. Interferon-γ: biology and role in pathogenesis.
Adv. Immunol.
62
:
61
26
Bullock, W. E., S. D. Wright.
1987
. Role of adherence-promoting receptors, CR3, LFA-1, and p150,95 in binding of Histoplasma capsulatum by human macrophages.
J. Exp. Med.
165
:
195
27
Eissenberg, L. G., W. E. Goldman.
1987
.
Histoplasma capsulatum fails to trigger release of superoxide from macrophages. Infect. Immun.
55
:
29
28
Basu, S., A. R. Dunn, M. W. Marino, H. Savoia, G. Hodgson, G. J. Lieschke, J. Cebon.
1997
. Increased tolerance to endotoxin by granulocyte-macrophage colony-stimulating factor-deficient mice.
J. Immunol.
159
:
1412
29
Rhoades, K. L., S. Cai, S. H. Golub, J. S. Economou.
1995
. Granulocyte-macrophage colony-stimulating factor and interleukin-4 differentially regulate human tumor necrosis factor-α promoter region.
Cell. Immunol.
161
:
125
30
Munker, R., J. Gasson, M. Ogawa, H. P. Koeffler.
1986
. Recombinant human TNF induces production of granulocyte-monocyte colony-stimulating factor.
Nature
323
:
79
31
Lenhoff, S., B. Sallerfors, T. Olofsson.
1998
. IL-10 as an autocrine regulator of CSF secretion by monocytes: disparate effects on GM-CSF and G-CSF secretion.
Exp. Hematol.
26
:
299
32
Ikegami, M., T. Ueda, W. Hull, J. A. Whitsett, R. C. Mulligan, G. Dranoff, A. H. Jobe.
1996
. Surfactant metabolism in transgenic mice after granulocyte macrophage-colony stimulating factor ablation.
Am. J. Physiol.
270
:
L650