Serum Stimulates Growth of and Proteinase Secretion by Aspergillus fumigatus

Abstract

Serum contains iron-binding proteins, which inhibit the growth of most pathogenic microorganisms, including fungi. The purpose of this research was to investigate the effect of serum on growth of the opportunistic fungal pathogen Aspergillus fumigatus. Supplementing minimal essential medium (MEM) with up to 80% human serum or up to 80% fetal bovine serum (FBS) stimulated growth and increased the amount of A. fumigatus dry biomass approximately fourfold. In addition, a 100-fold increase in proteinase secretion, as measured by azocasein hydrolysis, was observed when 10% human serum or 10% FBS was added to MEM. The fungal proteinases secreted in serum-containing media were shown to degrade ³H-labeled basal lamina proteins. The factor in serum that stimulated proteinase secretion was larger than 10 kDa and was 85% inactivated when the serum was heated for 30 min at 66°C. The proportions of proteinases of each catalytic class secreted by A. fumigatus in the presence of serum were different from the proportions secreted in media containing single proteins. Proteinase secretion did not result from increased protein concentration in the medium per se because bovine serum albumin (BSA) at a concentration equivalent to the concentration of serum produced only 20% of the proteinase activity per milligram (dry weight) that was produced by FBS. Addition of BSA plus 100 μM FeCl₃ to MEM resulted in the same level of growth as addition of serum, indicating that a combination of nutritional factors in serum may stimulate growth. However, the level of proteinase secretion was still only 30% of the level observed with FBS. These data indicate that serum does not inhibit the growth of A. fumigatus and that the nutrients in serum result in high levels of proteinase secretion, potentially increasing the invasiveness of this species.

Aspergillus fumigatus is an opportunistic fungal pathogen responsible for an increasing number of serious infections in immunocompromised individuals. A. fumigatus reproduces by producing large numbers of airborne conidia which, once inhaled by susceptible patients, can lead to life-threatening invasive pulmonary aspergillosis (22). Bone marrow and solid organ transplant recipients (46), cancer patients (4), AIDS patients (25), and patients with chronic granulomatous disease (39) are particularly at risk for developing invasive aspergillosis. Invasive aspergillosis can be treated with the antifungal drugs amphotericin B and itraconazole; however, these drugs have low rates of success. Even with prophylaxis and treatment with amphotericin B, the mortality rates average 65% for pulmonary aspergillosis and approach 100% if the disease spreads to the central nervous system (9). The basic virulence factors that allow some Aspergillus species to establish invasive infections remain unclear (11, 22).

A. fumigatus is a saprophytic organism and an abundant producer of proteinases. Proteinase secretion is inducible in A. fumigatus and occurs in response to the presence of proteins or protein hydrolysate (6). The role of proteinases in the virulence of A. fumigatus is unclear, but secretion of proteinases is thought to be necessary to break down protein barriers in the host. Infections due to Aspergillus are characterized by degradation of lung parenchyma (27), and some studies have shown that there is a correlation between elastase production and virulence in a mouse model of aspergillosis (21).

Mammalian serum inhibits the growth of many microbes, including some of the most common fungal pathogens. Fetal bovine serum (FBS) has been shown to inhibit the growth of Penicillium marneffei (43), while the growth of Candida albicans (26) and the growth of Cryptococcus neoformans (14, 32) are inhibited by human serum. Other fungal pathogens whose growth is inhibited by normal human serum include Histoplasma capsulatum (41), members of the genera Cunninghamella and Absidia (13), Rhizopus oryzae (3), and Rhizopus microsporus var. rhizopodiformis (5, 45).

The inhibitory action of serum is thought to be due to its ability to chelate iron, which deprives the invading pathogens of this essential nutrient (35). Iron is required by almost all organisms for catalysis during DNA synthesis and for enzymes involved in electron transport and energy metabolism (49). The role of iron in fungal pathogenesis has been reviewed recently (16, 48). Transferrin in serum binds iron with high affinity, and apotransferrin inhibits the growth of C. neoformans (40), C. albicans (16), H. capsulatum (41), and Rhizopus species (3). Consistent with the hypothesis that serum sequesters free iron is the finding that increased iron burden is a risk factor for infection by many different fungal pathogens. Supplementing serum with iron reverses its fungistatic effect on C. neoformans (40), P. marneffei (43), H. capsulatum (41), and C. albicans (26, 34). Moreover, treatment of dialysis patients with the iron-chelating drug desferrioxamine B is a risk factor for zygomycoses because iron bound to desferrioxamine is efficiently used as an iron source by some species of zygomycetes (1, 5, 10).

Little is known about the effects of serum on the growth of A. fumigatus. In this study, we examined the effects of serum on the growth of A. fumigatus and on the pattern of proteinase secretion in vitro. In contrast to what has been shown for most other common fungal pathogens, we found that neither FBS nor human serum inhibits the growth of A. fumigatus. Rather, both of these sera in fact stimulate growth when they are present at concentrations up to 80%. In addition, the levels of proteinase secretion in A. fumigatus are very high when either human serum or FBS is present.

MATERIALS AND METHODS

Strains and growth conditions.

A. fumigatus ATCC 13073 was obtained from the American Type Culture Collection and was maintained on YM slants (0.3% malt extract, 0.3% yeast extract, 0.5% peptone, 0.5% glucose) at 4°C. A. fumigatus was cultured on YM plates at 28°C for 5 to 10 days, until it was fully conidiated. Conidia were harvested by flooding a culture plate with phosphate-buffered saline (PBS) containing 0.05% Tween 20 and swabbing with a sterile cotton swab. The conidia were then vortexed, filtered through a plug of sterile glass wool to remove hyphae, and resuspended in PBS. Concentrations of conidia were determined by counting with a hemacytometer.

For most experiments, A. fumigatus was cultured in 5 ml (total volume) of minimal essential medium (MEM) (Life Technologies, Burlington, Ontario, Canada) containing serum or other medium components in 25-ml culture flasks. For iron depletion experiments, glassware was treated overnight with 1 mM EDTA and then for 2 h with 0.5 M HCl, and then it was rinsed six times with deionized water. MEM contains 1 mg of glucose per ml, amino acids, vitamins, and salts but no iron (12). Trace levels of iron were removed from MEM by stirring the medium overnight with 6% (wt/vol) Chelex 100 (Sigma, Oakville, Ontario, Canada). After the Chelex 100 treatment, 2 g of CaCl₂ per liter and 0.98 g of MgSO₄ per liter were added to MEM. Media containing high serum concentrations were prepared by using a 10× stock preparation of MEM to ensure that the MEM concentration remained constant. Conidia were added to media at a final concentration of 1 × 10⁶ conidia/ml, and the flasks were incubated at 36°C and 150 rpm. Dry weights were measured by filtering the entire contents of each flask through Miracloth (Calbiochem, La Jolla, Calif.) and rinsing the preparation thoroughly with distilled H₂O to remove all traces of culture medium. Mycelia were then transferred to preweighed microcentrifuge tubes, lyophilized overnight, and weighed.

Type II pneumocyte human cell line A549 was obtained from the American Type Culture Collection (catalog no. CCL-185) and was maintained in RPMI 1640 (Life Technologies) containing 10% FBS (Life Technologies), 2 g of sodium bicarbonate per liter, 0.1 mg of streptomycin per ml, and 25 U of penicillin G per ml. To prepare a culture of A. fumigatus on A549 cells, the medium was aspirated from a confluent plate, cells were washed three times with PBS, and then 10 ml of MEM containing 10⁶ conidia per ml was added.

Sera and other reagents.

FBS was obtained from Life Technologies. Human serum (male) was obtained from Sigma. All sera were stored in 10- or 100-ml aliquots at −20°C until they were used. Sera were thawed in a 37°C water bath and processed as described below.

A serum was heat treated in a Braun Thermomix 1441 11-liter water bath for 30 min at 56 or 66°C or was incubated at 100°C in a boiling water bath for 30 min. The serum was separated into high- and low-molecular-weight fractions by centrifugation at 3,000 × g for 1 h at 0°C through an Ultrafree filter with a 10,000-molecular-weight cutoff (Millipore, Bedford, Mass.). The retained material was dialyzed against three changes of water at 4°C through a membrane with a 10,000- to 12,000-molecular-weight cutoff to thoroughly remove low-molecular-weight components.

Iron was removed from serum as described by Wilson et al. (50). The serum was filtered through an Ultrafree filter with a 10,000-molecular-weight cutoff (Millipore), and the filtrate was stored at 4°C. The pH of the protein-containing fraction was adjusted to 4.0 with 1 M HCl, and EDTA was added to a final concentration of 10 mM. This fraction was dialyzed overnight against PBS containing 6% (wt/vol) Chelex 100 (Sigma) before it was recombined with the low-molecular-weight components. The iron content of the treated serum was determined by digesting samples in boiling nitric acid, preparing appropriate dilutions in iron-free water, and measuring the free iron content by atomic absorption spectrometry with a Perkin-Elmer AAnalyst 100.

Matrigel was obtained from Becton Dickinson Labware, Bedford, Mass.

Azocasein assay.

Proteinase secretion was quantified by the azocasein assay as outlined by Reichard et al. (36). Azocasein (Sigma) was dissolved at a concentration of 5 mg/ml in assay buffer containing 50 mM Tris (pH 7.5), 0.2 M NaCl, 5 mM CaCl₂, 0.05% Brij 35, and 0.01% sodium azide. Media were removed from Aspergillus cultures at various times and centrifuged to pellet the cells. The azocasein solution (400 μl) was mixed with 100-μl portions of supernatants from Aspergillus cultures and incubated in a 37°C water bath for 90 min. The reactions were stopped by adding 150 μl of 20% trichloroacetic acid, and the reaction mixtures were allowed to stand at the ambient temperature for 30 min. Tubes were then centrifuged for 3 min at 8,000 × g, and 500 μl of each supernatant was added to 500 μl of 1 M NaOH. The absorbance at 436 nm of released azo dye was determined with a spectrophotometer. For inhibitor studies, 400 mM stock solutions of phenylmethylsulfonyl fluoride (PMSF) and 1,10-phenanthroline (both obtained from Sigma) were prepared in ethanol. Culture supernatants were preincubated with 4 mM inhibitor for 15 min at 37°C, and then assay buffer containing 4 mM inhibitor was added.

Azocoll assay.

Azocoll substrate (>100 mesh; Calbiochem) is composed of insoluble particles of collagen to which a bright red azo dye has been attached. The Azocoll assay was based on the method of Chavira et al. (8). Azocoll was suspended in a buffer containing 50 mM Tris (pH 7.5), 1 mM CaCl₂, and 0.01% sodium azide. The Azocoll suspension (800 μl) was incubated with 5-μl portions of supernatants from Aspergillus cultures for 3 h at 37°C on a rotator. The tubes were centrifuged at 8,000 × g for 3 min, and the release of azo dye was determined by measuring the absorbance at 520 nm of each supernatant.

Preparation of radioactively labeled extracellular matrix.

The method used to prepare labeled extracellular matrix was modified from the method of Morschhauser et al. (31). A549 cells were seeded onto 24-well plates and incubated in RPMI 1640 containing 10% serum for 48 h in a 5% CO₂ atmosphere until they reached subconfluence. The medium was removed and replaced with glucose-free RPMI 1640 containing 1% FBS for 1 h. The medium was then replaced with fresh glucose-free RPMI 1640 containing 1% FBS and 2.5 μCi of D-[6-³H]mannose (American Radiolabeled Chemicals Inc., St. Louis, Mo.) per ml, and the preparation was incubated for 48 h. Cells were lysed by deoxycholate treatment (15) as follows. The confluent cells were rinsed three times with 1 ml of PBS and then incubated three times (10 min each) in 1 ml of 10 mM Tris-Cl (pH 8.0)-0.5% sodium deoxycholate-1 mM PMSF with slow shaking at 4°C. The wells were then washed four times (5 min each) at 4°C with 2 mM Tris-Cl (pH 8.0) with slow shaking. The absence of cells after this treatment was confirmed by microscopy. The amount of [³H]mannose incorporated into the extracellular matrix was estimated by subtracting the combined amount of radioactivity remaining in the supernatant, cell lysate, and washes from the total amount of radioactivity initially added.

Extracellular matrix degradation assays.

Samples of culture medium were withdrawn from a 48-h culture of A. fumigatus in MEM containing 10% serum or MEM alone. Control samples that contained fresh MEM or fresh MEM containing 10% serum were also prepared. Accumax (Innovative Cell Technologies, San Diego, Calif.) containing trypsin, collagenases, and other proteinases was used as a positive control for basal lamina hydrolysis. One-milliliter portions of culture media and control media were added to the wells containing the labeled extracellular matrix, and the plates were incubated at 37°C. Samples (50 μl) were withdrawn from each well at zero time and after 0.5, 1.0, and 3.5 h, and the radioactivity released was measured by liquid scintillation counting in Biodegradable Count scintillant (Amersham, Piscataway, N.J.) with a Beckman LS6500 scintillation counter.

Statistics.

Data analysis was performed by using the Student t test or analysis of variance followed by Tukey’s multiple-comparison procedure.

RESULTS

FBS and human serum stimulate growth and proteinase secretion in A. fumigatus.

Incubation of A. fumigatus in MEM containing 10% FBS resulted in a higher rate of growth and a larger total biomass than incubation in the same medium lacking serum (Fig. 1A). Growth was slightly delayed in media containing human serum, but the same maximum growth levels were reached. In serum-containing media, cultures reached the stationary phase after roughly 60 h of incubation, and the dry weight of fungi cultured in the presence of 10% serum was approximately fourfold greater than the dry weight of fungi cultured in serum-free MEM.

FIG. 1.

Growth of and proteinase secretion by A. fumigatus in 10 ml of MEM with or without 10% FBS or 10% human serum. A. fumigatus was cultured in MEM alone (▴), MEM containing 10% FBS (▪), or MEM containing 10% human serum (⧫) as described in Materials and Methods. Flasks were removed from incubation at several times, the contents of the whole culture flasks were filtered through Miracloth, and the mycelia were transferred to preweighed microcentrifuge tubes. (A) Mycelia were lyophilized overnight and weighed to determine dry weights. (B) Proteinase secretion measured by the azocasein assay as described in Materials and Methods. The data are means ± standard deviations based on three replicates and are representative of the data obtained in three independent experiments.

Proteinase secretion by A. fumigatus was monitored by azocasein hydrolysis. The presence of either 10% FBS or 10% human serum greatly stimulated proteinase secretion (Fig. 1B). In MEM alone, proteinase secretion by A. fumigatus was negligible during the 4-day incubation period. In MEM containing 10% serum, proteinase secretion peaked between 40 and 60 h, as late log phase was reached. The proteinase activity in the culture medium then decreased gradually as the culture aged. The levels of proteinase secretion were similar with human serum and with FBS, although the maximum levels of proteinase secretion were reached at an earlier stage of growth with human serum. The pH of the media was 7.2 at the start of the incubation and rose steadily during incubation to a maximum of 8.9. The pH was never observed to drop below the initial level, pH 7.2.

Serum contains many proteinases, including matrix metalloproteinases (MMPs). The serum MMPs are not able to hydrolyze casein (unpublished data); therefore, azocasein hydrolysis demonstrates that fungal proteinases are present. To exclude the possibility that serum MMPs simply activated latent fungal proteinases, proteinases secreted in media containing boiled serum were incubated with 10% fresh serum. Proteinase activity, as measured by the azocasein assay, decreased after 30 min of incubation of fungal proteinases with fresh serum (data not shown). Thus, it is unlikely that the increased fungal proteinase activity in the presence of serum was caused by activation of latent fungal proteinases by serum proteinases.

Serum concentrations up to 80% cause increased growth and proteinase secretion.

Previous work with Rhizopus rhizopodiformis var. rhizopodiformis showed that fungal growth was stimulated by human serum at concentrations up to 20% but was inhibited at all higher human serum concentrations (45). Therefore, the ability of A. fumigatus to grow in the presence of higher concentrations of FBS and human serum was evaluated. The growth of A. fumigatus was found to increase as the serum concentration was increased, and no evidence of growth inhibition by FBS was observed at concentrations up to 80% (Fig. 2A). Human serum also supported increased growth of A. fumigatus at concentrations up to 80% (Fig. 2B).

FIG. 2.

Growth of A. fumigatus in MEM containing high concentrations of FBS or human serum. A. fumigatus (10⁶ conidia/ml) was cultured in 5 ml of MEM containing FBS (A) or human serum (B) at concentrations ranging from 0 to 80%. The MEM concentration was constant. Flasks (25 ml) were incubated for 50 h at 37°C and 150 rpm. Dry weights were measured and azocasein assays were performed as described in Materials and Methods. The data are means ± standard deviations based on three replicates and are representative of the data obtained in three independent experiments (A) or two independent experiments (B). Symbols: ⧫, dry weight; ⋄, proteinase activity.

Proteinase secretion by A. fumigatus also increased steadily as the FBS concentration was increased. Human serum also stimulated proteinase secretion, but the effect reached a maximum with 40% serum and proteinase secretion declined slightly in the presence of higher concentrations.

Proteinases secreted by A. fumigatus are able to hydrolyze basal lamina proteins.

To determine if the proteinases secreted by A. fumigatus in serum-containing medium could hydrolyze proteins that might be encountered during infection, including those found in basal lamina, two assays were conducted: (i) Azocoll hydrolysis by supernatants from A. fumigatus cultured in MEM with or without serum (Azocoll is an insoluble collagen) and (ii) the release of proteins from radioactively labeled basal lamina. Azocoll was efficiently hydrolyzed by the proteinases secreted by A. fumigatus (Table 1), and the pattern of relative hydrolysis rates when the organism was grown in the presence of FBS or human serum was similar to the pattern observed with azocasein (Fig. 1B). A. fumigatus produced proteinases capable of Azocoll hydrolysis when it was grown in MEM alone, but the level of hydrolysis was only 1/10 the level observed when serum was present (Table 1).

TABLE 1.

Azocoll hydrolysis by A. fumigatus proteinases secreted in serum-containing media

Addition to MEM	A520a
	No conidia	46 hb	69 hb
FBS	0.00 ± 0.01	0.07 ± 0.07	0.50 ± 0.15*c
Human serum	0.01 ± 0.00	0.60 ± 0.28*	0.23 ± 0.08*
No serum	0.00 ± 0.00	0.06 ± 0.05	0.02 ± 0.00

^aThe values are means ± standard deviations based on triplicate A₅₂₀ values obtained after incubation of 5-μl portions of culture supernatants with 800 μl of a 5-mg/ml Azocoll suspension for 3 h at 37°C.

^bLength of A. fumigatus incubation.

^cAn asterisk indicates that a value is significantly greater than the value obtained in media without A. fumigatus (P < 0.05).

We next investigated whether serum-enhanced proteinase secretion might contribute to the invasiveness of A. fumigatus by aiding in the breakdown of human pulmonary basal lamina. This protein barrier must be breached by pathogens if they are to move from the lungs into the vasculature and other organs. Radiolabeled basal lamina was prepared by using human type II pneumocyte cell line A549, and the basal lamina glycoproteins were labeled with [³H]mannose. The proteinases secreted by A. fumigatus in a medium containing serum resulted in a rate of hydrolysis of basal lamina proteins that was higher than the rate in a medium lacking serum (Fig. 3). In agreement with the Azocoll assay results, low levels of basal lamina protein hydrolysis were observed for A. fumigatus cultured in MEM alone. Hydrolysis of basal lamina proteins by fungal proteinases was rapid, and significant release of radioactivity occurred mostly during the first 30 min of incubation. After 200 min, the fungal proteinases from media containing serum and Accumax, a mixture of trypsin and collagenases, had released equivalent amounts of radioactivity from basal lamina. When controls containing fresh media in which fungi had not been cultured were used, there was no increase in the release of [³H]mannose during the 200-min sampling period (Fig. 3).

FIG. 3.

Basal lamina hydrolysis by A. fumigatus proteinases. Basal lamina was metabolically labeled with [³H]mannose by using A549 cells as described in Materials and Methods. Media from A. fumigatus cultures grown in MEM with FBS or in MEM alone were withdrawn after 48 h and applied to the radiolabeled basal lamina at zero time. Hydrolysis of basal lamina proteins was detected by monitoring the release of ³H into solution by liquid scintillation counting. Control prepartions contained MEM with serum or MEM alone but no A. fumigatus. The data are means ± standard deviations for the total radioactivity released into solution based on three replicates. The total radioactivity released from basal lamina by Accumax after 200 min was 29,800 ± 6,400 dpm. The data are representative of the data obtained in three independent experiments. At each time point (30, 60, and 210 min), the radioactivity released by supernatant containing fungi and serum was significantly greater than the radioactivity released by supernatant containing fungi without serum (P ≤ 0.05).

Pattern of proteinase secretion depends upon the composition of the growth medium.

The molecular weights and catalytic classes of the proteinases induced by serum were compared to the molecular weights and catalytic classes of the proteinases induced by other proteins. A. fumigatus was cultured in MEM alone or in MEM containing either 8 mg of gelatin per ml, 10% FBS, 10% human serum, 8 mg of bovine serum albumin (BSA) per ml 8 mg of BSA per ml and 100 μM FeCl₃, 6% Matrigel, or A549 cells. Matrigel is isolated from the EHS mouse sarcoma and is composed of the basal lamina proteins laminin, collagen IV, entactin, and heparan sulfate proteoglycan (19). Culture supernatants were withdrawn and analyzed by the azocasein assay with proteinase inhibitors.

The proportions of proteinases in the serine proteinase and metalloproteinase catalytic classes depended on the protein substrate (Fig. 4). For example, in MEM containing gelatin, serine proteinases were secreted almost exclusively, whereas in media containing BSA, roughly equal amounts of serine proteinases and metalloproteinases were secreted. In media containing either human serum or FBS, the ratio of serine proteinase activity to metalloproteinase activity was roughly 2:1. No inhibition was detected with inhibitors of cysteine or aspartic acid proteinases in any of the samples (data not shown).

FIG. 4.

Catalytic classes of proteinases secreted by A. fumigatus. A. fumigatus was cultured in MEM containing 6% Matrigel (columns A), 8 mg of gelatin per ml (columns B), 10% human serum (columns C), 10% FBS (columns D), 8 mg of BSA per ml (columns E), 8 mg of BSA per ml and 100 μM FeCl₃ (columns F), or A549 cells (columns G). Supernatants from the cultures were incubated with 5 mg of azocasein per ml in buffer containing either a serine proteinase inhibitor, 4 mM PMSF, or an metalloproteinase inhibitor, 4 mM 1,10-phenanthroline. An equivalent volume (1%) of ethanol was used as a control. The azocasein activities without inhibitors were 0.34 ± 0.01 A₄₃₆ unit (columns A), 0.20 ± 0.02 A₄₃₆ unit (columns B), 0.46 ± 0.18 A₄₃₆ unit (columns C), 0.37 ± 0.07 A₄₃₆ unit (columns D), 0.08 ± 0.02 A₄₃₆ unit (columns E), 0.15 ± 0.06 A₄₃₆ unit (columns F), and 0.63 ± 0.12 A₄₃₆ unit (columns G). The data for each inhibitor are the percentages of activity in the presence of the inhibitor relative to the activity in the absence of the inhibitor and are means ± standard deviations based on three replicates. Open bars, percent inhibition by 4 mM PMSF (= percentage of serine proteinases); shaded bars, percent inhibition by 4 mM 1,10-phenanthroline (= percentage of metalloproteinases). No inhibition was detected with inhibitors of cysteine and aspartic proteinases in any of the samples (data not shown).

The factor in FBS that stimulates proteinase secretion in A. fumigatus is larger than 10 kDa and is partially heat labile.

In an effort to determine whether a particular serum component stimulated proteinase secretion, FBS was passed through a filter with a 10,000-molecular-weight cutoff, and the retained material was dialyzed against distilled water, which separated the serum into high- and low-molecular-weight fractions. The volume of the high-molecular-weight fraction was adjusted to restore the original serum protein concentrations. When A. fumigatus was incubated with the two fractions, proteinase secretion was observed only in the presence of the high-molecular-weight fraction, and the proteinase levels were similar to those obtained with whole serum (Table 2). A. fumigatus cultured in media containing the low-molecular-weight fraction and in MEM alone secreted very low levels of proteinases (Table 2). Thus, proteinase secretion is induced by a serum component whose molecular mass is greater than 10 kDa.

TABLE 2.

Growth of and proteinase secretion by A. fumigatus in MEM containing serum, dialyzed serum, or low-molecular-weight serum components

Addition to MEM	Proteinase secretion (A436 unit)a	Dry wt (mg)b
10% FBS	0.70 ± 0.07	6.9 ± 1.0
10% dialyzed serum (>10 kDa)	0.51 ± 0.12	8.9 ± 0.5
10% serum filtrate (<10 kDa)	0.05 ± 0.04*c	3.7 ± 0.3*
MEM alone	0.01 ± 0.00*	1.9 ± 0.2*

^aProteinase secretion was measured by the azocasein hydrolysis method. The values are means ± standard deviations based on triplicate determinations of proteinase secretion by A. fumigatus after 72 h of incubation in MEM containing 10% serum or 10% high- or low-molecular-weight serum fractions, prepared as described in Materials and Methods.

^bThe values are mean ± standard deviations for total dry mycelial mass from 5-ml cultures after 72 h of incubation.

^cAn asterisk indicates that a value is significantly different from the value obtained with 10% FBS (P < 0.025).

FBS was also heat treated to determine if the serum factor that induces proteinase secretion in A. fumigatus is heat labile. It was found that heating serum to 56°C did little to change the growth of A. fumigatus or the level of proteinase secretion. However, when serum was heated to 66 or 100°C, proteinase secretion decreased by about 85% (Table 3). In contrast, growth of A. fumigatus was not affected by heat treatment of serum (Table 3).

TABLE 3.

Growth of and proteinase secretion by A. fumigatus in MEM containing FBS or heat-treated FBS

Serum treatment	Proteinase secretion (A436 unit)a	Dry wt (mg)b
Unheated	0.84 ± 0.21	8.5 ± 1.0
56°C	0.87 ± 0.05	9.1 ± 0.2
66°C	0.11 ± 0.02*c	8.5 ± 0.2
100°C	0.09 ± 0.02*	9.3 ± 0.5

^aProteinase secretion was measured by the azocasein assay. The values are means ± standard deviations based on triplicate determinations of proteinase secretion by A. fumigatus that was incubated for 72 h in MEM containing FBS that had been heated to 56, 66, or 100°C for 30 min.

^bThe values are means ± standard deviations for total dry mycelial mass from 5-ml cultures after 72 h of incubation.

^cAn asterisk indicates that a value is significantly different from the value obtained with unheated serum (P < 0.025).

Stimulation of growth by serum can be mimicked by the effect of BSA plus iron.

Proteinase secretion in A. fumigatus can be induced by the presence of protein or protein hydrolysate in the medium (6). Thus, we tested the ability of one serum protein, BSA, to stimulate growth and proteinase secretion. A stock solution containing BSA at a concentration of 80 mg/ml, which was roughly equivalent to the protein concentration in serum, was prepared and added to culture medium at a final concentration of 8 mg/ml. Addition of BSA to MEM significantly stimulated the growth of A. fumigatus, although the growth was still only 80% of the growth observed in the presence of FBS (Fig. 5). The amount of proteinase secreted per milligram (dry weight) increased significantly when BSA was added to MEM. Nevertheless, the levels of proteinase secreted in the presence of BSA were only 20% of those observed when cultures were grown in the presence of 10% FBS (Fig. 5).

FIG. 5.

Proteinase secretion by A. fumigatus in MEM containing BSA and iron. MEM (5 ml) containing 10% FBS, 8 mg of BSA per ml, 100 μM FeCl₃, or 8 mg of BSA per ml and 100 μM FeCl₃ and MEM alone were inoculated with 10⁶ A. fumigatus conidia per ml and incubated at 37°C and 150 rpm. Proteinase secretion per milligram of dry weight (bars) and dry weight (⧫) were measured as described in Materials and Methods. The data are means ± standard deviations based on five independent experiments. Statistical analysis of the growth data revealed that the effects of FBS and BSA plus iron were both significantly greater than the effects of BSA alone (P < 0.01), which in turn were significantly greater than the effects of iron plus MEM and MEM alone (P < 0.01). For proteinase secretion, the effects of FBS were significantly greater than the effects of BSA and BSA plus iron, and the effects of these two treatments were significantly different from the effects of MEM plus iron and MEM alone (P < 0.01).

Serum can be a source of iron if a microorganism possesses a mechanism to remove iron bound to transferrin or other iron-binding proteins (47). Using atomic absorption spectrometry, we found that the iron levels of the sera used in our experiments were as follows: FBS, 4.7 ppm (84 μM); and human serum, 1.1 ppm (20 μM). The level in MEM was below the detection limit (0.1 ppm). We therefore postulated that the presence of iron in serum might be responsible for its growth-enhancing effect on A. fumigatus. Addition of 100 μM FeCl₃ to MEM had no effect on either growth of or proteinase secretion by A. fumigatus, while addition of 100 μM FeCl₃ and 8 mg of BSA per ml together resulted in the same amount of growth that was observed with 10% FBS (Fig. 5). While the combination of FeCl₃ and BSA significantly improved growth, it had only a small effect on proteinase secretion compared to the effect of BSA alone; the levels of proteinase secretion in the presence of BSA plus iron remained 30% of the levels observed in the presence of FBS (Fig. 5). We also attempted to remove iron from serum by treating the serum with Chelex 100 beads as described by Wilson et al. (50). However, treatment of either FBS or human serum resulted in removal of only 50% of the total iron (data not shown). The remaining iron was sufficient to maintain the same level of growth of A. fumigatus that was observed in the presence of untreated sera; after 66 h of incubation, the biomasses of cultures containing Chelex 100-treated human serum were 6.84 ± 0.46 mg, whereas in the presence of untreated sera the biomasses were 5.81 ± 0.91 mg (n = 3). A small effect on proteinase secretion was observed; the amounts of proteinase secreted in the presence of Chelex 100-treated human serum were 0.05 ± 0.01 A₄₃₆ units/mg (dry weight), and the amounts secreted in the presence of untreated human serum were 0.07 ± 0.01 A₄₃₆ units/mg (dry weight). Supplementing samples containing Chelex 100-treated serum with 10 μM iron restored the level of proteinase secretion to 0.07 ± 0.01 A₄₃₆ units/mg (dry weight). These data indicate that removal of up to 50% of the iron present in human serum (final iron concentration, 10 μM) did not inhibit the growth of A. fumigatus but did have a small inhibitory effect on proteinase secretion by this fungus.

DISCUSSION

Serum plays an important role in defense against microbial growth. In extracellular fluids, iron-chelating proteins maintain the free iron concentration at about 10⁻¹⁸ M (7), a concentration too low to support the growth of many microbes, including most pathogenic fungi. Transferrin is considered to be a major antifungal agent in serum, while lactoferrin is present in other bodily fluids and is responsible for the antifungal properties of milk (2). Successful pathogens must evolve mechanisms to access iron if they wish to proliferate in a mammalian host. Pathogenic fungi employ several different strategies for iron acquisition, including reduction of ferric iron to ferrous iron and production of various siderophores (16).

The purpose of the present study was to determine if serum inhibits the growth of the fungal pathogen A. fumigatus. Two interesting observations were made while we studied the growth of A. fumigatus in serum. First, A. fumigatus, unlike other fungal pathogens, was able to thrive in an iron-limiting environment and to grow in the presence of up to 80% human serum and up to 80% FBS. Second, addition of serum resulted in high levels of proteinase secretion by A. fumigatus. We hypothesized that A. fumigatus is able to use iron in serum and that a combination of available iron and protein in serum permits abundant growth of A. fumigatus. Iron was not released by changes in pH in the batch cultures because the pH of the culture medium remained neutral to alkaline.

The basal medium used throughout this study was MEM, which contains glucose, amino acids, vitamins, and salts (12). Although MEM does not contain added iron, the trace levels of iron present in this medium were sufficient to permit some growth of A. fumigatus. Addition of FBS or human serum to MEM greatly increased the growth of A. fumigatus, even at serum concentrations up to 80%. To our knowledge, A. fumigatus is the only fungal pathogen able to tolerate and thrive in the presence of such high serum concentrations. Human serum at a concentration of 1% in RPMI 1640 inhibited the growth of C. albicans by 98% 2(26). The growth of C. neoformans was inhibited by 76% in RPMI 1640 containing 5% human serum but was not inhibited by FBS (14, 32). Human serum inhibited the growth of R. rhizopodiformis, but only at concentrations greater than 20% (44).

Although fetal and adult sera contain roughly equal transferrin concentrations, fetal transferrin is more highly iron saturated. The concentration of transferrin in FBS is 1.2 to 1.8 mg/ml, and this transferrin is 55 to 92% iron saturated (18), while transferrin in human serum is usually 25 to 35% saturated (47). FBS therefore has less potential to chelate free iron in the medium, which could explain its inability to inhibit growth of pathogens such as C. neoformans. Our study showed that human serum stimulated growth of and proteinase secretion by A. fumigatus to the same extent as FBS, despite its greater theoretical iron-chelating potential. Nevertheless, the observed delay in growth in the presence of human serum compared to growth in the presence of FBS may have been caused by more limited iron availability in human serum.

Supplementing MEM with 100 μM FeCl₃ alone did not increase growth, indicating that iron levels in MEM were not growth limiting for A. fumigatus. Furthermore, iron was not limiting for growth of A. fumigatus in MEM containing 10% FBS because addition of 100 μM FeCl₃ did not result in increased growth and proteinase secretion (data not shown). However, addition of BSA and 100 μM FeCl₃ to MEM allowed growth at levels equivalent to those observed in serum-containing MEM. These results show that growth of A. fumigatus in media containing 8 mg of BSA per ml was limited by iron availability. Since iron was not growth limiting in media containing serum, A. fumigatus must be able to access the iron in serum. Dialysis of serum against water did not result in a loss of growth-promoting activity; therefore, the source of iron accessed by A. fumigatus was likely molecules whose molecular masses are greater than 10 kDa, probably protein-bound iron.

The mechanism by which A. fumigatus obtains iron in serum-containing media is unknown. Unlike some bacterial pathogens, no fungus studied to date can remove iron directly from transferrin (16). It is thought that A. fumigatus can produce at least six siderophores, the most prominent of which have been tentatively identified as N,N′,N"-triacetylfusarinine C and ferricrocin (33). These siderophores have the potential to help scavenge iron in serum-containing media, but their role in virulence has not been examined.

A. fumigatus is a saprophytic organism and a prolific producer of proteinases. Proteinase secretion is inducible and is stimulated by protein and protein hydrolysate (6). The role of proteinases in virulence of A. fumigatus has been extensively studied, and at least three secreted proteinases have been described. These proteinases include an alkaline proteinase (Alp) (20, 30, 36), a metalloproteinase (Mep) (24, 28), and an aspartic proteinase (Pep) (37, 38). A. fumigatus strains lacking one or more of these proteinases have been shown to completely retain their virulence (17, 29, 38, 42). All three known A. fumigatus-secreted proteinases are secreted in vivo (23, 24, 36), but their role in vivo is unknown, nor is it known if other proteinases are also secreted in vivo. It is not known which, if any, of the proteinases that have been described are induced by serum. Serum also contains many inhibitors of proteinases, such as α₂-macroglobulin and α₁-antichymotrypsin. Our results show that the proteinases produced by A. fumigatus are able to evade or overwhelm the proteinase inhibitors in serum.

The results presented in this paper suggest that regulation of proteinase secretion by A. fumigatus is complex. The patterns of proteinase secretion were highly dependent on the composition of the medium. For example, gelatin and Matrigel stimulated production of serine proteinases exclusively, while serum, BSA, and A549 cells induced secretion of both serine proteinases and metalloproteinases.

Hydrolysis of transferrin is one potential means of iron acquisition in serum. High levels of proteinase secretion could allow A. fumigatus to degrade iron-binding proteins and release iron required for growth. However, hydrolysis of iron-binding proteins is not likely to be the only mechanism for iron acquisition by A. fumigatus, because proteinase secretion was not detected until cultures reached the late log phase. Thus, most of the growth and the need for iron occurred before high levels of proteinases were produced.

The proteinases secreted in response to the presence of serum have the potential to contribute to the virulence of A. fumigatus. They efficiently cleaved Azocoll, a general proteinase substrate derived from bovine collagen. The secreted proteinases also degraded human pulmonary basal lamina proteins in intact basal lamina. The basal lamina is a protein barrier underlying the epithelial cells and is composed of mainly laminin, collagen IV, and heparan sulfate proteoglycans (19). Previous studies have shown that A. fumigatus can produce proteinases capable of hydrolyzing basal lamina components. For example, Tronchin et al. (44) reported a loss of laminin epitopes from kidney and lung tissue slices incubated with a crude protease extract from A. fumigatus grown in nitrogen-restricted Czapek Dox medium (44). Increased secretion of basal lamina-degrading proteinases in vivo could result in increased invasiveness of this species.

The serum factor that induces proteinase secretion is not merely protein in the medium, because BSA alone did not induce high levels of proteinase secretion. Furthermore, while heat treatment of serum did not affect the growth yield, it did decrease the ability of serum to induce proteinase secretion by A. fumigatus. This suggests that serum stimulates growth by acting as a nutrient source, whereas stimulation of proteinase secretion requires nondenatured serum proteins. Proteinases are secreted earlier in the growth of A. fumigatus in media containing BSA and FeCl₃ than in serum-containing media. The delay in proteinase secretion in media containing serum may be related to the inaccessibility of iron as accessing iron bound to serum proteins would be expected to take longer than accessing free iron in solution. Work is under way to identify the mechanism by which A. fumigatus obtains iron from serum and to explore the possible role of siderophores in growth in serum-containing media.