INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Aspergillus fumigatus is the causative agent of a variety of diseases in humans, such as allergic bronchopulmonary aspergillosis (ABPA), asthma, sinusitis, aspergilloma (AO), invasive pulmonary aspergillosis (IPA), and invasive disseminated aspergillosis (IDA). Deleterious invasive aspergillosis (IA) is an increasing problem in the immunocompromised host, and the rate of lethality of IA is estimated to be as high as 95% (8, 27).

An efficient antifungal therapy for IA depends on an early diagnosis, but this is limited due to the absence of specific clinical symptoms at the early stage of disease. Specific diagnostic tools that lead to an early and sufficient therapy could substantially improve the clinical outcome (1, 4, 6), but conventional serological tests face some major obstacles. Most immunoassays for detection of circulating antibodies are based on crude extracts of A. fumigatus. These extracts contain complex and undefined mixtures of polysaccharide and protein components and are derived from various sources such as conidial, mycelial, cytoplasmic, metabolic, or cell wall fractions of the fungus (13-15, 28). Consequently, there is little knowledge about the role of the humoral immune response, the cross-reactivities of different epitopes in these crude antigenic preparations, the diagnostic value of class-specific antibody detection, and antibody kinetics in human aspergillosis (7).

The utilization of well-characterized immunodominant fungal antigens could substantially improve the serodiagnosis in Aspergillus-related diseases. Galactomannan (GM), the immunodominant polysaccharide cell wall antigen of Aspergillus spp., has been studied extensively in the past (9, 23, 24, 33), and recently, monoclonal antibody EB-A2 has successfully been used in a sensitive direct double-sandwich enzyme-linked immunosorbent assay (ELISA) to detect circulating GM during IA (9, 29, 30, 33). A drawback of this method, however, is the high frequency of false-positive results (34, 35, 37), which may be due to cross-reactivity with other fungi or other unidentified serum components (35, 36).

Few data are available about A. fumigatus protein antigens that are preferentially produced in vivo. Mitogillin is a small basic protein of approximately 18 kDa with cytotoxic activity released by A. fumigatus (16, 17, 31). Together with the related toxins A. fumigatus allergen I (AspfI), restrictocin from Aspergillus restrictus (20), and -sarcin from Aspergillus giganteus (26), mitogillin is a member of a family of conserved RNases that cleave a single phosphodiester bond of the 29S rRNA of eukaryotic ribosomes (11, 16, 17).

An indication that mitogillin is produced in vivo during infection was provided by Lamy et al. (21), who detected mitogillin within kidney cells of mice infected with A. fumigatus in regions of necrosis surrounding fungal colonies. Furthermore, Arruda et al. (3) demonstrated that 85% of the patients with immunoglobulin E (IgE) antibodies to A. fumigatus also had IgE antibodies to AspfI, which they defined as a major allergen of the fungus. Interestingly, mitogillin was found to be one of the major A. fumigatus antigens detectable in human urine (12, 21, 23, 25), and among the water-soluble ethanol-precipitated proteins (WSEPs) from broth cultures of A. fumigatus, mitogillin was recognized by pooled sera of aspergilloma patients (25), suggesting that the ribotoxin can successfully be used as a specific diagnostic marker for A. fumigatus-associated diseases.

To study the diagnostic potential of mitogillin in human aspergillosis, we expressed mitogillin from A. fumigatus in Escherichia coli and purified the recombinant protein to homogeneity. After protein analysis, the recombinant mitogillin was used for specific IgG, IgM, and IgA antibody detection in patients suffering from different forms of aspergillosis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Fungal strains and plasmids. A. fumigatus strain M2045 was isolated from the tracheal aspirate of a patient suffering from cystic fibrosis. A. fumigatus strain M5299 was obtained from the bronchoalveolar lavage fluid of an AIDS patient suffering from disseminated IA. Strain M5299 was deposited at the Centraalbureau voor Schimmelcultures (CBS) culture collection (CBS 109032). The isolates were maintained on Sabouraud dextrose agar slants. Plasmid pMIT+ carrying the gene encoding mitogillin (17) was a gift from Jacqui Shea (Royal Postgraduate Medical School, Hammersmith Hospital, London, United Kingdom).

Culture and WSEPs. Culture, fungal antigen extraction, and the isolation of WSEPs from the culture filtrate were performed as described by Latgé et al. (25). Briefly, conidia of A. fumigatus strains M2045 and M5299 were inoculated in 50 ml of Sabouraud liquid medium (2% [wt/vol] glucose, 1% [wt/vol] Mycopeptone). The cultures were incubated for 4 to 5 days at 25°C and 100 rpm. The shaken liquid cultures were transferred to 1 liter of Sabouraud medium and cultured at 25°C and 600 rpm. Maximal fungal growth was observed after 48 to 72 h. The culture filtrates from the strains were precipitated with 4 volumes of ethanol overnight at 4°C. The precipitate was washed two times with ethanol and resuspended in water (approximately 20 mg [dry weight] per ml). The undissolved precipitate was removed by centrifugation, and the WSEPs were aliquoted and stored at 80°C for further analysis.

Construction of the mitogillin expression vector. The coding region without the leader sequence of the mitogillin gene was amplified by PCR from plasmid pMIT+ with primer RES1-f (5'-AGGGAGCTCATGGCGACCTGGACATGCATC-3'), which contains a SacI restriction site, and primer RES1-r (5'-AACTGCAGCTAATGAGAACACAGTCTCAAGT-3'), which contains a PstI restriction site. PCR was performed in a volume of 50 µl (500 ng of plasmid pMIT+ DNA; 20 mM [each] dATP, dCTP, dGTP, and dTTP; 25 pmol of each primer; 2.5 U of Taq DNA polymerase [Amersham, Braunschweig, Germany]; 5 µl of 10× Taq buffer [Amersham]; 3 µl of MgCl₂ [25 mM]; 38.5 µl of sterile water). Amplification was performed in an Omni-Gene temperature cycler (Hybaid; MWG-Biotech, Ebersberg, Germany). For a hot start, the PCR reagent mixture was preheated to 94°C before DNA was added. Twenty-five cycles were carried out, with each cycle consisting of (i) denaturation for 1 min at 94°C, (ii) annealing for 1 min at 59°C, and (iii) extension for 90 s at 72°C. Final extension was performed for 10 min at 72°C after the last cycle.

Recombinant expression and purification of mitogillin. The expression plasmid pQEMW1 was transformed by electroporation into E. coli host strain M15 containing a repressor plasmid (pREP4). Successful transformation was confirmed by PCR analysis of the colonies with primers RES1-f and RES1-r. Twenty milliliters of Luria-Bertani (LB) broth containing 100 µg of ampicillin per ml and 25 µg of kanamycin per ml was inoculated with a single colony of the transformed strain (37°C, vigorous shaking, overnight). For large-scale expression, 1 liter of LB broth (with 100 µg of ampicillin per ml and 25 µg of kanamycin per ml) was inoculated 1:50 with the overnight culture. Growth of the bacteria was continued at 37°C with vigorous shaking until the A₆₀₀ reached 0.7 to 0.9. The culture was induced with 2 mM isopropyl--D-thiogalactopyranoside and continued to grow at 37°C for another 5 h. Bacterial cells were harvested by centrifugation at 4,000 × g for 10 min. To avoid protein degradation, purification of the recombinant mitogillin was done immediately after growth, induction, and cell lysis. Protein purification was performed with an Ni²⁺ HiTrap chelating affinity column and the GradiFrac system (Pharmacia, Freiburg, Germany) according to the instructions of the manufacturer. The human immunoglobulin preparations Pentaglobin and Intraglobin F (Biotest Pharma GmbH, Dreieich, Germany) were used in Western blot analysis to detect contaminating antigens from E. coli in the recombinant antigen preparation. Protein refolding was done by dialysis by the protocol for the Qiaexpressionist kit (Qiagen). The protein concentration was determined by the bicinchoninic acid protein assay (Pierce, Rockford, Ill.).

Infection and immunization procedures. Four New Zealand White rabbits (weight, 3.0 to 3.5 kg) were used. Two rabbits received 10⁸ viable conidia via subcutaneous injection. The first was infected with strain M2045 and the second was infected with strain M5299. The procedure was repeated every second week five times before the animals were killed. Two additional rabbits were used for immunization with native or recombinant mitogillin. For these immunization procedures Antibody-Multiplier normal and Antibody-Multiplier special adjuvants (Linaris, Bettingen, Germany) were used according to the suggestions of the manufacturer. Briefly, a polyclonal rabbit antiserum against native A. fumigatus mitogillin was raised by immunizing a third animal with the protein that was eluted from the gels on which sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of WSEPs from the culture filtrate of strain M2045 was performed. A polyclonal rabbit antiserum against recombinant mitogillin was raised by immunizing a fourth animal with the purified recombinant mitogillin. In total, 3.0 mg of protein was given to each animal. Serum samples were obtained from each animal before the immunization and at 2, 4, 6, 8, 10, and 12 weeks after the first immunization or infection procedure.

Western blotting and ELISA. (i) Antigenic characterization of recombinant mitogillin. After SDS-PAGE under reducing conditions, recombinant mitogillin and the WSEPs of A. fumigatus strain M2045 were transferred onto nitrocellulose (cellulose nitrate; pore size; 0.45 µm; Schleicher & Schuell, Dassel, Germany) and used as antigens in Western blot analysis. The blots were probed with polyclonal rabbit sera derived either from rabbits infected with A. fumigatus M2045 or M5299 or from rabbits immunized with either recombinant or native mitogillin. In addition, Western blots were carried out with polyclonal antiserum obtained by immunizing mice with putative degradation products of the recombinant protein.

(ii) Recombinant mitogillin ELISA. A recombinant ELISA was developed to detect specific IgG, IgM, and IgA antibodies to mitogillin from A. fumigatus. A total of 1 µg of the recombinant protein per ml was diluted in coating buffer (0.015 M Na₂CO₃, 0.035 M NaHCO₃ [pH 8.4]), and 100 µl of the antigen solution was coated (4°C, overnight) onto each well of ELISA microtiter plates (high binding, U shape; Greiner, Frickenhausen, Germany) before 150 µl of ELISA blocking reagent (Boehringer, Mannheim, Germany) was added for 30 min. Subsequently, 150 µl of diluted serum (1:100 in phosphate-buffered saline-0.05% [vol/vol] Tween 20) was added to each well for 2 h at 37°C. Thereafter, 50 µl of conjugates was added to each well, and the plates were incubated for 1 h at 37°C. The conjugates were diluted in phosphate-buffered saline-0.05% (vol/vol) Tween 20 as follows: anti-human IgG-horseradish peroxidase (HRP) conjugate (DAKO, Hamburg, Germany), 1:6,000; anti-human IgM-HRP conjugate (DAKO), 1:2,000; anti-human IgA-HRP conjugate (DAKO), 1:4,000; anti-rabbit IgG-peroxidase conjugate (Dianova, Hamburg, Germany), 1:4,000. Five washings with washing buffer (0.35 M NaCl, 0.05 M Tris-HCl [pH 7.4], 0.1% [vol/vol] Tween 20) was performed between each step of the ELISA. Substrate was prepared with 14 mg of 1,2-phenylenediamine dihydrochloride (DAKO), 12 ml of H₂O, and 5 µl of 30% (vol/vol) H₂O₂. One hundred microliters of the substrate was added to each well, and the plate was incubated for 15 min. The reaction was stopped with 100 µl of 0.5 N H₂SO₄, and the optical density (OD) was measured at a wavelength of 491 nm (reference filter, 620 nm). The OD values were adjusted for any background detected for assays in wells that had not been coated with mitogillin. All serum samples were tested in duplicate.

Reproducibility of the assay. To ensure the reproducibility of the recombinant mitogillin assay, we have tested (i) control serum samples 1 (negative), 2 (borderline), and 3 (positive) on 10 runs on different days and (ii) we have compared the index values of control serum samples 4 to 8 (samples 4 and 5 were negative, sample 6 was borderline, and samples 7 and 8 were positive) with four different affinity chromatography preparations of purified recombinant mitogillin. All tests were done in duplicate.

Statistical analysis: cutoff values. The mean values of the ODs for specific IgG, IgM, and IgA antibodies against mitogillin were determined for the serum samples of 307 blood donors. The cutoff value for the mitogillin IgG, IgM, or IgA ELISA was defined as the mean value for 307 donor serum samples plus 3 standard deviations. A test result smaller than the cutoff value but larger than the mean value for the blood donor sera plus 2 times the standard deviation was considered borderline. The results of the mitogillin ELISAs are given as index values, which were calculated as the ratio of a given OD compared to the cutoff value. An index value of 1 indicates a positive test result.

A. fumigatus antigen detection. A. fumigatus GM antigen detection in serum samples was done with the Platelia kit and the Pastorex kit (Sanofi Diagnostics Pasteur, Marnes la Coquette, France) according to the instructions of the manufacturer.

Detection of A. fumigatus antibodies to crude antigens. Anti-A. fumigatus antibody detection was performed by two conventional assays based on crude Aspergillus extracts. Anti-A. fumigatus complement-fixing antibodies were tested with a commercially available preparation of A. fumigatus metabolic antigen suitable for complement fixation (CF; Immuno-Mycologics, Norman, Okla.), according to the instructions of the manufacturer. Counterimmunoelectrophoresis (IE) was done with a commercially available A. fumigatus antigen for the detection of immunoprecipitating antibodies (HAL GmbH, Düsseldorf, Germany).

Serum samples and patients. For the determination of the cutoff values and the specificity of the recombinant mitogillin ELISA we examined serum samples of 307 healthy blood donors (mean age, 40.0 years; age range, 19 to 65 years) for specific IgG, IgM, and IgA antibodies against mitogillin. In addition, we tested serum samples of a patient who was suffering from mucormycosis, of a patient who was suffering from histoplasmosis of the lung and who had antibodies against Histoplasma capsulatum, and of a patient who had candidemia and antibodies against Candida albicans. Furthermore, we screened four serum samples of two healthy immunocompetent individuals who showed positive antibody titers in a commercially available anti-A. fumigatus IgG antibody detection kit (HALISA; HAL Allergie GmbH).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of recombinant mitogillin and antigen characterization. The identity of the PCR product that was cloned into expression vector pQE30 with mitogillin (17) was proved by sequencing. Following affinity chromatography the recombinant mitogillin was analyzed by SDS-PAGE; its molecular mass was slightly larger than that of the native mitogillin due to the 6× HIS affinity tag (data not shown). In Western blot experiments, no contamination with E. coli proteins was detected in the recombinant protein preparation with human Pentaglobin and Intraglobin F. These polyclonal human immunoglobulin preparations reacted well with multiple proteins from E. coli M15, which was used for expression of recombinant mitogillin (data not shown). In Western blots, serum IgG from polyclonal rabbit serum samples against native or recombinant mitogillin as well as IgG from polyclonal rabbit serum samples that were raised by infecting animals with A. fumigatus M2045 or M5299 reacted with recombinant mitogillin. Likewise, serum IgG of polyclonal rabbit serum raised against recombinant mitogillin reacted with native mitogillin of the WSEPs from A. fumigatus strain M2045. In addition, the natural mitogillin of the WSEPs was recognized by IgG from polyclonal rabbit sera generated against native mitogillin and those polyclonal sera raised by infecting the animals with conidia of A. fumigatus clinical strains M2045 and M5299 (Fig. 1). Using follow-up serum samples, we observed that mitogillin is an immunodominant A. fumigatus antigen that elicits an early antibody response compared to the times of the responses elicited by the other A. fumigatus antigens of the WSEPs when rabbits were infected with viable A. fumigatus conidia (Fig. 2). In the Western blots we observed a potential degradation product of mitogillin that reacted with serum IgG of polyclonal rabbit serum samples raised against native and recombinant mitogillin (Fig. 1, lanes B of parts I and II). To test this hypothesis, mice were immunized with the purified degradation product. In Western blots mitogillin was well recognized by these polyclonal mouse serum samples (data not shown).

View larger version (63K): [in this window] [in a new window]   FIG. 1.   Western blots with WSEPs of A. fumigatus strain M2045 (lanes A) and recombinant mitogillin (lanes B) as antigens. The blots were probed with serum samples of rabbits that were immunized with native mitogillin (I) and recombinant mitogillin (II) and of an animal that was infected with A. fumigatus M2045 (III). Molecular mass markers (in kilodaltons) are shown in the right margin. Every combination results in the reaction of the polyclonal antibody with mitogillin (arrows).

View larger version (84K): [in this window] [in a new window]   FIG. 2.   Kinetics of antibody production to WSEP antigens of A. fumigatus. The Western blots were probed with consecutively obtained polyclonal serum samples of a rabbit that was infected with viable A. fumigatus conidia of strain M2045. Molecular mass markers (in kilodaltons) are shown in the right margin. Antibodies to mitogillin appear early after infection (arrow).

Control sera. The reproducibility of the anti-mitogillin antibody assay was shown by testing control serum samples 1 to 3 (negative, borderline, and positive, respectively) in 10 runs on different days. The standard deviations within the runs were 0.10 for control serum sample 1 (mean index value, 0.51), 0.12 for control serum sample 2 (mean index value, 0.92), and 0.14 for control serum sample 3 (mean index value, 2.17). When four different preparations of recombinant mitogillin were compared, the standard deviations for control serum samples 4 to 8 were as follows: 0.03 for control serum sample 4 (mean index value, 0.33), 0.08 for control serum sample 5 (mean index value, 0.57), 0.06 for control serum sample 6 (mean index value, 0.76), 0.07 for control serum sample 7 (mean index value, 1.22), and 0.12 for control serum sample 8 (mean index value, 1.76).

Patients with A. fumigatus-associated diseases. Table 1 presents the results of anti-A. fumigatus antibody and Aspergillus antigen detection in serum samples of patients suffering from AO, IPA, IDA, or type III allergy to A. fumigatus. In all serum samples (n = 32) obtained from 13 episodes of AO, specific IgG antibodies against mitogillin were detected (sensitivity = 100%). In nine episodes (17 serum samples) we detected specific IgM against mitogillin, and in eight episodes (14 serum samples) specific IgA antibodies were present. Precipitating anti-A. fumigatus antibodies were found in all serum samples tested in the AO group (n = 32) by IE. Complement-fixing antibodies were detected in four episodes (six serum samples). GM was detected in nine episodes (16 serum samples) by the Platelia sandwich ELISA. GM antigen was not detected in any of the serum samples obtained from patients 5, 9, 10, and 12 (all tested by the Platelia sandwich ELISA).

View larger version (13K): [in this window] [in a new window]   FIG. 3.   Index values for specific IgG antibodies to mitogillin obtained with serum samples of patient 22 suffering from IPA. The cutoff is indicated by a solid line. Borderline test results are indicated by a dotted line.

View larger version (17K): [in this window] [in a new window]   FIG. 4.   Index values for specific IgG antibodies to mitogillin obtained with serum samples of patient 28 suffering from IDA. The cutoff is indicated by a solid line. Borderline test results are indicated by a dotted line.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our knowledge about the role of the humoral immune response in human aspergillosis is still limited (7). Crude and undefined antigen preparations of A. fumigatus are commonly used for antibody detection, but standardized immunodominant antigens which indicate invasive disease are still not available for A. fumigatus. As a consequence, profound difficulties exist in the interpretation of serological test results (5, 7). We decided to investigate the usefulness of mitogillin for improving the serodiagnosis of aspergillosis. We identified this antigen as an 18-kDa protein in all WSEP preparations of 12 clinical A. fumigatus isolates (M. Weig, data not shown). These data are in accordance with the observations of Latgé et al. (25), who showed that a major immunodominant protein in the WSEPs of different A. fumigatus reference strains had a molecular mass of approximately 18 kDa.

In the present study we show that mitogillin can be used as a standardized marker for the early diagnosis of A. fumigatus-associated diseases. In all 13 episodes of AO (32 of 32 serum samples were positive; sensitivity = 100%) and in 18 of 20 episodes of IA (51 of 82 serum samples were positive; sensitivity = 62.2%), IgG antibodies to mitogillin were detected when repetitive sampling was performed. Two episodes of IA (in two AIDS patients) were anti-mitogillin IgG negative but IgA positive. In comparison, Aspergillus GM antigen was found in only 9 of 13 episodes of AO and in 10 of 20 episodes of IA. In 2 episodes of IA, anti-mitogillin antibodies were detected prior to GM antigen, GM antigen was detected earlier in 4 episodes, and in 10 episodes only anti-mitogillin antibodies were detected but GM antigen was absent. It is important that 8 of these 14 GM antigen-negative but anti-mitogillin antibody-positive episodes were tested by the sensitive Platelia sandwich ELISA. These findings underline the value of repetitive and combined testing for the diagnosis of IA. At this point we cannot draw a definite conclusion on the diagnostic value of serum IgM and IgA antibodies in the stage-specific diagnosis of A. fumigatus-associated diseases. However, the ubiquitous nature of the fungus and the high average daily level of exposure to A. fumigatus conidida might explain why IgG antibodies against mitogillin are superior to IgM antibodies as diagnostic markers in the diagnosis of A. fumigatus-associated diseases.

In our study we have compared the detection of antibodies to mitogillin by antibody tests based on crude Aspergillus extracts (CF, IE). In an evaluation of eight antibody tests for the diagnosis of IA, Kappe et al. (18) reported that CF (with the metabolic antigen of A. fumigatus) was superior to indirect hemagglutination tests and enzyme immunoassays. However, unsatisfactory sensitivity was observed for all antibody tests based on crude Aspergillus extracts (18). IE is one of the diagnostic procedures most commonly used to detect anti-Aspergillus antibodies in immunocompetent patients suffering from AO and ABPA (for a review, see reference 22). This method is sufficiently insensitive for the elimination of false-positive results that occur in Aspergillus-exposed healthy individuals. The disadvantages of this method are the inability to quantitate the immune response, the lack of standardization due to the use of crude extracts, and the fact that a minor contamination during the production of large quantities of antigen leads to erroneous results. These disadvantages are reflected in the results of our study. As expected, IE was useful in the diagnosis of AO. However, both conventional methods (IE, CF) failed to contribute sufficiently to the serodiagnosis of IA.

In follow-up sera of rabbits that were infected with viable A. fumigatus conidia, we observed that anti-mitogillin antibodies appear early in the course of the infection in comparison with the time of appearance of antibodies against different antigens of the WSEPs of A. fumigatus. To evaluate the kinetics of antibody production in human aspergillosis, we examined follow-up serum samples of patients suffering from fatal IPA and IDA and correlated the test results with the clinical course of disease. The earliest time possible for infection in patient 22 was the day of the lung transplantation, as IPA was restricted to the transplanted lung. Five days after surgery the patient's serum showed borderline levels of IgG against mitogillin, and the IgG levels were highly positive on day 12. In patient 28, IgG antibodies against mitogillin were observed 10 days after the onset of respiratory symptoms. Circulating GM antigens were not detected in any of the serum samples of patients 22 and 28. These results stress the diagnostic value of detection of antibody against mitogillin for the early diagnosis of A. fumigatus-associated diseases. We observed a sharp decrease in the levels of IgG antibody against mitogillin in patient 28 prior to his death. It could be speculated that antibodies to mitogillin play a key role in the control and humoral immune defense of human IA. We are therefore evaluating the value of detection of antibody against mitogillin as a prognostic marker of IA.

False-negative test results and a high frequency of samples with false-positive results presumably due to cross-reactivity with other polysaccharide fungal antigens or other serum components are major drawbacks of solely GM antigen-based serodiagnosis of aspergillosis (34, 35, 37). Mitogillin shares a high degree of amino acid sequence homology only with restrictocin and, to lower extents, with -sarcin (17) and clavin (D. Parente et al., unpublished data). These proteins are expressed by species of the genus Aspergillus (A. restrictus, A. giganteus, and Aspergillus clavatus) (21). Antigens showing cross-reactivity with mitogillin have been identified in cultures of Aspergillus fischeri, Aspergillus ochraceus, Neosartoria aureola, and Neosartoria stramenia but were absent from other genera of medically important fungi (25).

A. fumigatus is ubiquitous in nature, and Smith et al. (32) calculated that the average number of inhaled conidia is approximately 30 per day. Individuals working in gardens, in animal quarters, or with compost heaps might be burdened with massive numbers of A. fumigatus conidia. High levels of exposure to conidia lead to the production of antibodies against several A. fumigatus antigens in healthy individuals (25). This makes it difficult to discriminate exposure from infection by conventional serological techniques (19) based on crude extracts. However, the array of A. fumigatus antigens expressed depends largely on environmental conditions such as the medium used, the origin of the crude extract, or the type of propagule (i.e., conidium versus mycelium) (25). Ribotoxins of the mitogillin family are scarcely found in resting conidia but are highly expressed during active growth of the fungus (2). For these reasons, we speculated that a mitogillin-based serological test system is highly specific for the diagnosis of Aspergillus-related diseases. Sera from patients who suffered from mucormycosis, candidemia, or histoplasmosis showed no mitogillin-reactive antibodies. In addition, only 4 (1.3%) of the 307 blood donor serum samples examined in our study were positive for IgG antibody against mitogillin, 1.3% were positive for IgM antibody, and 1.0% were positive for IgA antibody by the recombinant mitogillin ELISA. These data obtained for the blood donors confirm the high degree of specificity (95.4%) of the mitogillin-based test system. Similar ratios were identified by Heymann and colleagues (10), who detected AspfI-reactive IgG antibodies in 6% of control patients. We tested serum samples of blood donors with an age range of from 19 to 65 years. Although prolonged exposure to A. fumigatus conidia in older donors is likely, higher levels of antibody against mitogillin were not observed, resulting in a high degree of specificity of the test. The limited number of positive blood donor sera might reflect the prevalence of A. fumigatus allergy in the population. Our data indicate that even high levels of exposure to A. fumigatus conidia do not elicit production of antibody against mitogillin, as serum samples of healthy immunocompetent individuals positive for A. fumigatus antibody with a commercial kit based on crude antigens showed no significant levels of antibody against mitogillin.

The availability of recombinant mitogillin for specific antibody detection provides a new serological tool for the diagnosis of A. fumigatus-associated diseases. Our data indicate a high degree of correlation of production of antibody against mitogillin and the clinical course of disease. Finally, the standardized antigen will be used in further studies to define the role of the humoral immune response in human aspergillosis.