Aspergillus terreus Species Complex

SUMMARY

Infections due to Aspergillus species are an acute threat to human health; members of the Aspergillus section Fumigati are the most frequently occurring agents, but depending on the local epidemiology, representatives of section Terrei or section Flavi are the second or third most important. Aspergillus terreus species complex is of great interest, as it is usually amphotericin B resistant and displays notable differences in immune interactions in comparison to Aspergillus fumigatus. The latest epidemiological surveys show an increased incidence of A. terreus as well as an expanding clinical spectrum (chronic infections) and new groups of at-risk patients being affected. Hallmarks of these non-Aspergillus fumigatus invasive mold infections are high potential for tissue invasion, dissemination, and possible morbidity due to mycotoxin production. We seek to review the microbiology, epidemiology, and pathogenesis of A. terreus species complex, address clinical characteristics, and highlight the underlying mechanisms of amphotericin B resistance. Selected topics will contrast key elements of A. terreus with A. fumigatus. We provide a comprehensive resource for clinicians dealing with fungal infections and researchers working on A. terreus pathogenesis, aiming to bridge the emerging translational knowledge and future therapeutic challenges on this opportunistic pathogen.

INTRODUCTION

Aspergillus species are saprotrophic fungi found in a wide variety of habitats, such as soil, compost, and dust (1). Aspergillosis is an infection caused by Aspergillus species, and the most common representatives are A. fumigatus species complex, A. flavus species complex, and A. terreus species complex (1). Aspergillus species cause a wide spectrum of fungal diseases depending on the underlying immune status. The main syndromes include allergic bronchial pulmonary aspergillosis, chronic necrotizing Aspergillus pneumonia, aspergilloma, and invasive diseases. The illness resulting from an infection usually affects the respiratory system, with various degrees of severity. Invasive aspergillosis is the most severe form. Depending on the type of aspergillosis, treatment involves antifungal medication, surgery, or observation. Within the aspergilli, the A. terreus species complex takes a special position, as most representatives are amphotericin B resistant. In addition, A. terreus is an increasingly recognized opportunistic fungus that accounts for nearly 4% of all cases of invasive aspergillosis (2) and is frequently associated with dissemination and poor outcome. Herein, we review the basic aspects of microbiology, epidemiology, genomics, and pathogenesis of A. terreus species complex, highlight the underlying mechanisms of amphotericin B resistance, and discuss the associated clinical challenges. In addition, we contrast key elements of pathophysiology and clinical scenarios of aspergillosis caused by A. terreus versus A. fumigatus, which is the dominant Aspergillus species causing invasive disease.

TAXONOMY

The genus Aspergillus, consisting of approximately 350 accepted species, is divided into six subgenera (Aspergillus, Circumdati, Cremei, Fumigati, Nidulantes, Polypaecilum) (3–5) (Fig. 1). The Aspergillus section Terrei, belonging to the subgenus Circumdati, was first described in 1965 (6), and A. terreus was considered to be the only species within this section. Over time, molecular and phylogenetic studies (7–9) assigned additional species from the section Flavipedes (A. niveus, A. carneus, and A. niveus var. indicus) and Versicolores (A. allahbadii, A. ambiguus, and A. microcysticus) (10, 11) to this section. A. alabamensis was introduced in 2009 (12), and A. aureoterreus, A. floccosus, A. hortai, A. neoafricanus, and A. neoindicus were ranked to the species level in 2011 (11). A. neoniveus was transferred from the teleomorph genus Fennellia, and the newly described species A. pseudoterreus (11) and A. citrinoterreus (13) were introduced into the section Terrei. The recent discovery of A. iranicus in hypersaline soil in Iran and of A. bicephalus (14, 15) resulted in a total of 17 accepted species distributed over three series (Ambigui, Nivei, Terrei) (5). Most species of the section Terrei produce characteristic yellow- to cinnamon brown-colored asexual phialidic conidia, whereas a few members, such as A. allahabadii and A. neoindicus, produce white conidia. The cinnamon brown color of these conidia discriminates species from section Terrei from other pathogenic Aspergillus species, such as those from sections Fumigati (16), Flavi (17), and Nigri (18).

FIG 1.

Overview of Aspergillus subgeneric and sectional classification of Aspergillus section Terrei.

General Characteristics

A. terreus, the most important species within the section Terrei, has a worldwide distribution and is found in a wide variety of habitats, such as soil, compost heaps, dust, and food, particularly in plant products like stored corn, rice, peanuts, and barley (19), but is not categorized as a common spoilage mold (20). On Czapek or Sabouraud dextrose agar, colonies grow from beige to buff to cinnamon (Fig. 2A) and are able to become floccose, the reverse is yellow, and yellow soluble pigments are frequently present (6). Conidial heads are compact, columnar (reaching up to 500 μm long by 30 to 50 μm in diameter), and biseriate (Fig. 2B). Conidiophores are hyaline and smooth walled; conidia are globose shaped, smooth walled, 1.5 to 2.5 μm in diameter, and vary in color from hyaline to light yellow (21–23). In contrast to other Aspergillus species, A. terreus produces accessory conidia—also called aleurioconidia—under both in vitro and in vivo conditions (24, 25). These accessory conidia are formed directly from vegetative hyphae (Fig. 2C), are nonpigmented, and present large amounts of immunogenic β-glucans on their surface (26). They play a potential role in pathogenesis and dissemination (25–27). For more information, see Biology and Pathogenesis.

FIG 2.

(A) Aspergillus terreus sensu stricto presenting with typical cinnamon brown colony culture morphology on Sabouraud dextrose agar (72 h, 37°C). (B) Wet mount preparation with lactophenol cotton blue showing biseriate and densely columnar conidial heads. (C) Calcofluor white staining showing A. terreus accessory conidia in vivo (bronchoalveolar lavage fluid); magnification, ×1,000. (D) A. terreus producing sectors on Sabouraud dextrose agar (72 h, 37°C) displaying white (less-pigmented), fluffy, and conidium-sparse colonies.

A. terreus species produce a large number of secondary metabolites and mycotoxins in vitro, but the in vivo production of such substances during invasive growth has so far been poorly investigated. The cholesterol-lowering drugs lovastatin (28), terrein, quadrone, and asterriquinone (29–32), which have antitumor activities, the acetylcholinesterase inhibitors territrem B and cyclosporine (33, 34), and the antiviral compound acetylaranotin are the most important secondary metabolites (35, 36) (for more information, see Biology and Pathogenesis). In addition, like other molds, A. terreus species play an important role in the fermentation industry, as they produce the primary metabolite itaconic acid, which is mainly used in the plastic and paint industry (37). A variety of mycotoxins, such as citreoviridin, patulin, citrinin, terretonin, and gliotoxin, were found to be produced by A. terreus species complex (38–42). It appears that different members of the Aspergillus section Terrei produce different types of secondary metabolite and mycotoxin in vitro (11).

Genomes and Phylogeny

The genus Aspergillus forms a broad monophyletic group, showing substantial taxonomic divergences in morphology and phylogeny (43–45). Whole-genome studies of relevant Aspergillus species, such as A. fumigatus, A. nidulans, and A. oryzae, were among the first to be reported for filamentous fungi (44, 46, 47). These data provided the basis for comparative and functional genome studies in the Aspergillus research community. The annotation of the full genome sequence of A. terreus isolate NIH2624 was published in 2005 by The Broad Institute of MIT and Harvard within the Aspergillus Comparative Sequence Project (https://www.broadinstitute.org/fungal-genome-initiative/aspergillus-genome-projects). A recent study described a high-quality whole-genome assembly of the A. terreus clinical isolate M6925 derived by single-molecule real-time sequencing with short-read polishing (48). The genome sequence of the A. terreus NIH2624 strain is 29.36 Mb and has a GC content of 52.8% and 10,401 predicted genes (Table 1). The A. terreus clinical isolate M6915 revealed genomic features similar to the NIH2624 strain and shared an average nucleotide identity of over 99% with previously deposited short-read assemblies of A. terreus (strain NIH2624 and strain KM017963 isolated from soil) (43, 48, 49). However, the limited number of A. terreus whole-genome sequences restricts further statements on common features. Comparative genome studies of 18 industrially and medically important Aspergillus species, including A. terreus, revealed both strong conservation for biological functions and high diversity for biological traits (50). Aspergillus species genomes are similar in size (29 to 36 Mb) and GC content (48% to 53%), and the numbers of predicted genes range from 9,113 to 13,553. Eighty percent of identified Aspergillus genus genes show homology with other fungal lineages, indicating that about 20% are specific to Aspergillaceae (50). A comparison of A. fumigatus proteins to their orthologs in A. terreus showed a median percentage identity of 71% (51). Fedorova et al. point out that on a molecular basis, the two species are as distant from each other as human and fish (51). Overall, 6,460 A. fumigatus (strain Af293) genes have orthologs in the A. terreus genome, 728 A. fumigatus (Af293) genes have no orthologs in the A. terreus genome; the Aspergillus core group (A. terreus, A. niger, A. oryzae, A. nidulans) has 5,424 proteins with orthologs in the three most closely related aspergilli (A. fumigatus, Neosartorya fischeri, and A. clavatus [Affc-core group]). A total of 2,090 proteins of the Affc-core group have no orthologs in the Aspergillus core group (51). Allergen coding genes of A. fumigatus were compared to the A. terreus clinical isolate M6915 genome, which indicated a strong homology between several A. fumigatus allergens (>85% sequence identity for Asp f 3, Asp f 12, Asp f 18, Asp f 22, and Asp f 23). This raises the possibility of cross-reactivity between A. fumigatus and A. terreus allergens (48). The core set of known and predicted allergens is conserved in all sequenced Aspergillus spp. (51). However, some allergens show divergent regions in the center or at the C-terminal end of the predicted protein. Some of these putative cell wall allergens (e.g., Asp f 2, Asp f 4, Asp f 7, Asp f 9, and Asp f 17) are members of the adhesion family. Differences in structure may reflect differences in adhesion substrate between the various aspergilli. Also, the variable domains of these proteins may reduce the likelihood of cross-reactivity to IgE from non-A. fumigatus species (52). A. fumigatus (Af293) allergens that are found in A. terreus are Asp f 3, Asp f 8, Asp f 9, Asp f 10, Asp f 11, Asp f 12, Asp f 13, Asp f 18, Asp f 22, Asp f 23, and Asp f 28, and 19 predicted A. fumigatus (Af293) allergens (proteins that display significant sequence homology with known fungal allergens, some of which might contribute to pathogenicity) are also found in A. terreus (51).

TABLE 1.

Genomic features of Aspergillus terreus in comparison to clinically relevant Aspergillus species

Species	Strain	Section	Genome size (Mb)	GC (%)	No. of predicted genes	Reference or source
A. terreus	NIH2624	Terrei	29.33	52.8	10,401	https://www.broadinstitute.org (accessed 23 March 2021)
A. terreus	M6915	Terrei	31.84	52.15	11,102	48
A. fumigatus	Af293	Fumigati	29.40	49.9	9,926	47
A. niger	CBS513.88	Nigri	33.98	50.4	14,165	205
A. flavus	NRRL 3357	Flavi	36.89	48.3	13485	206

On a genomic level, the diversity of secondary metabolite production highlights another important characteristic among A. terreus and related species. Notably, species of the Aspergillus section Terrei produce a diverse array of secondary metabolites and also differ in the set of metabolites produced (11). Polyketide synthases (PKSs) and nonribosomal synthetases (NRPSs), frequently clustered together, are involved in the biosynthesis of the vast majority of secondary metabolites (53–55). A comparison of 19 Aspergillus genomes showed a large variation in both the number of individual PKS and NRPS genes (from 18 in A. glaucus to 58 in A. tubingensis) and the number of associated gene clusters (from 18 in A. glaucus to 57 in A. niger). Comparative genomic studies displayed 28 PKS genes, 20 NRPS genes, and one 1 PKS-NRPS hybrid gene in A. terreus to be involved in the production of secondary metabolites (50). It must be noted that these numbers are subject to fluctuations as a result of progressive analytical methods. In addition, the discovery of new functional units involved in secondary metabolite production creates new subgroups that lead to further variation. Ongoing updates on secondary metabolite clusters are found in the MycoCosm database (https://mycocosm.jgi.doe.gov/mycocosm/home). Interestingly, these large variations do not correlate with genome size, suggesting that the genes and their associated clusters rely on the lifestyle and respective requirements of each species (50). For more information, see Biology and Pathogenesis.

EPIDEMIOLOGY

A. terreus accounts for a minority of invasive aspergillosis (27), although higher infection rates have been found in Houston, Texas (USA), and Innsbruck, Tyrol (Austria) (2, 56–58). The frequency of A. terreus infections in these specific regions cannot be explained by only geoclimatic factors and reflects the complex interactions of host, fungal pathogen, and environment. In addition, previous administration of antifungal drugs to which A. terreus is resistant, such as amphotericin B, might have accounted for selection events (57, 59). In contrast, the widespread use of triazoles in the last 2 decades has resulted in a decreased incidence of A. terreus in our center (60). Lackner et al. (61) aimed to analyze whether epidemic genotypes are responsible for the accumulation of invasive aspergillosis cases due to A. terreus in Innsbruck, Austria. The study identified three persisting genotypes as responsible for 70% of A. terreus infections. These genotypes were found predominantly in the Tyrolian environment (80%). We conclude that the high frequency of environmental A. terreus may be at least one reason for increased human infection rates in specific regions (62).

Genotypes and Transmission

A variety of molecular typing methods is useful for studying the epidemiology of A. terreus species complex infections in humans. Random amplification of polymorphic DNA-PCR (RAPD-PCR) was developed in the 1990s and has been one of the most widely used genotyping techniques (63–65). This method is rapid and straightforward but suffers from poor reproducibility and does not allow data exchange and storage in global databases. However, the technique has been successfully applied in A. terreus disease outbreak analyses and surveillance studies (66–68). Other approaches include single-strand conformation polymorphism analysis (SSCP) of PCR-amplified intergenic spacer regions (ITS) and amplified fragment length polymorphism analysis (AFLP). SSCP of amplified ITS regions successfully differentiated medically important Aspergillus species, including A. terreus (69). AFLP has been used for Aspergillus species identification and differentiation (70–72) and to study the population structure of A. terreus obtained from clinical and environmental sources (73). Currently, the PCR-based microsatellite length polymorphism (MLP) method, which amplifies short tandem DNA repeats (STRs) located at numerous loci in eukaryotic genomes, is far more often in use. This method is well established for Aspergillus species, is highly species specific, displays reliable discriminatory power and excellent reproducibility, and last but not least allows objective data interpretation (74).

The primary means of acquisition of Aspergillus is the inhalation of airborne conidia, which are released from environmental sources such as soil, decomposing organic material, and contaminated food. A study from Lass-Flörl and colleagues demonstrated A. terreus-contaminated potted plants in the hospital to be the most probable source of invasive infection in patients with hematological disease (75). Hospital-acquired (nosocomial) infections with A. terreus have been mainly associated with immunocompromised hosts and environmental disturbances, such as construction sites and/or renovation work in units for high-risk patients (75–77). Environmental surveys have focused mainly on A. fumigatus, and little is known about A. terreus and other filamentous fungi. Air sampling, both outside and inside, revealed low concentrations of A. terreus (also A. flavus and A. niger) in comparison to A. fumigatus (66, 68, 78).

The majority of genotype studies on A. terreus did not reveal clonal relationships, indicating that patients are infected with unique strains (57, 66, 68, 79). A high degree of genetic diversity was also obtained using both STR (microsatellite) and AFLP for A. terreus isolates collected from India, North America, and Europe (73). On the other hand, Balajee et al. explored 94 worldwide-collected clinical and environmental strains by using multilocus comparative sequence analysis of three genes and reported the existence of a single, globally distributed A. terreus population (12). Two years later, inter-simple sequence repeat (ISSR) PCR of 117 A. terreus isolates revealed one clade that was exclusively from Europe and another from the United States, indicating an association between geography and genotype (80). STR genotyping supports the latter data, as A. terreus isolates taken from clinical and environmental sources in Tyrol (Austria) presented three endemic major genotypes (MGs) (61). The three MGs were prevalent in both human infections and the environment. No major differences in virulence were observed using Galleria mellonella as a model. Arabatzis and Velegraki (81) proved that A. terreus is able to reproduce sexually, the reason why we studied the mating type distributions. Mating type analyses revealed that all three MGs carry the mat1-2 allele, resulting in a mating barrier between these well-adapted strains. No recombination was observed within the MGs present in Innsbruck (61).

Ecology and Global Distribution

The ability to utilize a wide variety of organic substances and adaptation to a broad range of environmental conditions enable Aspergillus species to survive and grow in diverse habitats, including the human body (82, 83). The asexually produced conidia are highly stress tolerant and easily become airborne due to their hydrophobicity and small size; these are excellent conditions in which to act as an opportunistic pathogen. A. terreus shares characteristics with A. fumigatus that cause pathogenicity, and these include the ability to colonize and damage the host, adaptation to nutritional and biophysiological challenges, flexibility to undergo asexual cycle, formation of cryptic species (taxonomically accepted distinct species that share a morphology), longevity (conidia remain viable for periods of decades), high tolerance to heat and O₂ pressure, and finally, the ability to germinate at a low water activity (a_W) (the lowest a_W supporting growth of A. terreus has been reported to be 0.78) (12, 84, 85). As mentioned before, A. terreus is a less common cause of invasive aspergillosis but has been reported to be geographically more restricted with a high incidence in Innsbruck (Tyrol, Austria) and Houston (Texas, USA). These two areas do not share a climate or environmental features that would explain this phenomenon. However, the presence of A. terreus in the local environment might be the key requisite for the development of invasive aspergillosis (56, 57, 61, 66, 79). The occurrence of airborne A. terreus (indoor and outdoor) and its possible clinical impact have been addressed in several studies. In general, A. fumigatus was the most frequently isolated species from environmental sources (ranging from 54% to 97%), independent of the geographical region. This correlates with the clinical incidence demonstrating that A. fumigatus is the leading cause of invasive aspergillosis. Among culture-positive Aspergillus isolates in Cologne, Germany, and Madrid, Spain, A. terreus accounted for only 0.2% and 0.5%, respectively (68, 78). Similar studies did not detect A. terreus at all (86–90). Based on this low environmental distribution, it is unlikely that A. terreus is a ubiquitous pathogen worldwide. In Innsbruck, A. terreus accounted for 30% of invasive aspergillosis cases between 1990 and 2000, but only 1% of collected air samples from inside and outside the hospital area were positive for A. terreus (66). The low number of aspergilli inside the hospital, particularly in high-efficiency particulate air-filtered areas, indicates that fungal uptake most likely occurred prior to hospitalization (56). A possible explanation may be the intracellular persistence of A. terreus conidia, as is discussed in Biology and Pathogenesis.

Prevalence and Incidence

A. terreus infections have been reported increasingly as the cause of acute and chronic aspergillosis (2). Invasive diseases are associated with a high rate of dissemination and show a poor outcome in comparison to A. fumigatus infections (56, 91). At the University of Alabama Hospital in Birmingham, A. terreus isolates rose from 1.5% of all Aspergillus isolates in 1996 to 15.4% in 2001 (79). In the 1990s and 2000s, the frequency of invasive aspergillosis caused by A. terreus varied from 3% to 12.5% (27, 44, 58, 83, 92–95). A more recent prospective international multicenter surveillance study, involving 21 countries and 38 centers, demonstrated an overall A. terreus prevalence of 5.2% (370/7116) among all Aspergillus infections (2). This study included cases from Europe, the Middle East, India, and South and North America. Spain, followed by Austria, Israel, and Germany (2), are countries with the highest density of A. terreus species complex isolates collected. It is remarkable that Aspergillus section Terrei was most commonly isolated from patients suffering from chronic lung diseases (39%). However, it should be noted that these epidemiological data may suffer from reporting bias, as participation was voluntary, not nationwide or obligatory. Another multicenter surveillance study in Italy investigated 292 Aspergillus isolates from patients with hematological malignancies and showed that 4.8% (14/292) of isolated species belonged to section Terrei (96). The number of cryptic species within the genus Aspergillus is increasingly recognized and of clinical relevance, since antifungal drug resistance is common in isolates of cryptic taxa. Several studies report an increase in cryptic Aspergillus species in clinical samples (97–99). A follow-up study of an international A. terreus survey evaluated the frequency of A. terreus and related (cryptic) species in a global set of 495 clinical isolates. A. terreus sensu stricto was the most frequently isolated species (86.7%). Cryptic species accounted for 13%, with A. citrinoterreus being the most prevalent (8.4%), followed by A. hortai (2.6%), A. alabamensis (1.6%), A. neoafricanus (0.2%), and A. floccosus (0.2%), whereas the occurrence of cryptic species varies depending on geographical origin (2, 100). Other cryptic species belonging to section Terrei (A. allahabadii, A. ambiguus, A. aureoterreus, A. bicephalus, A. carneus, A. iranicus, A. microcysticus, A. neoindicus, A. niveus, and A. pseudoterreus) were not identified in this international survey (100). Similar data have been reported by other authors (12, 73, 80, 101). Although A. terreus sensu stricto is the most commonly found species in clinical samples, intrinsic amphotericin B resistance and virulence potential are similar to that of cryptic species, suggesting that—from a clinical point of view—discrimination of cryptic species is not necessary for a routine laboratory (102).

BIOLOGY AND PATHOGENESIS

Pathogenicity

There are several infection models for the study of virulence and disease progression of invasive aspergillosis caused by Aspergillus species, mainly A. fumigatus. An invertebrate infection model that uses larvae of the greater wax moth Galleria mellonella is easy to handle. By infecting the hemocoel of larvae with conidia and monitoring survival curves, the model reflects the virulence attenuation of A. fumigatus mutants as observed in mouse infection models (103). This infection model has been successfully used to study A. terreus infections (104). However, this model lacks pharmacokinetic modeling and does not allow the application of immunosuppressive regimens, and the strong and rapid melanization of larvae succumbing to infection hampers subsequent investigations of tissue penetration and immune interactions. It is assumed that some features are comparable to immune reactions observed in the human host (105). Another invertebrate infection model utilizes Toll-deficient Drosophila melanogaster flies. The handling is more complicated, and the generation of Tl^−/− flies requires crossing of flies carrying a thermosensitive allele of Toll (Tl^r632) with flies carrying a null allele of Toll (Tl^I-RXA) (106). Comparative analyses of the virulence of A. fumigatus and A. terreus lacked significant differences in the survival of infected flies (107).

A more complex but still simple to use infection model that has been applied to A. fumigatus and A. terreus uses embryonated hen’s eggs (108). In this system, a hole is drilled into the eggshell, giving access to the chorioallantoic membrane that gets infected with conidia, which is followed by sealing of the hole with wax and further incubation of the eggs in a breeder. Egg candling, which is visualization of the blood vessels and developing embryo in the egg by use of a light source, allows easy and rapid assessment of embryo survival. Comparative studies of A. fumigatus and A. terreus revealed that a dose of 100 to 1,000 conidia (A. fumigatus laboratory or wild-type strain) is sufficient to cause 100% mortality (108, 109). In contrast, A. terreus strains required an approximately 100- to 1,000-times-higher infectious dose to achieve a survival pattern resembling that observed for A. fumigatus (110). In comparison to A. fumigatus, these experiments indicate a reduced virulence due either to a lower growth rate of A. terreus at >37°C or to a different interaction of the fungus with the embryonic immune system. A similar observation for virulence is found in murine infection models.

Murine infection models are currently the gold standard in fungal virulence and antifungal drug efficacy studies. However, the most frequently used wild-type mice, such as inbred BALB/c and C57BL/6, are naturally resistant to fungal infections. Immunosuppression is required to make these mice susceptible to fungal infections (111). Leukopenic models, in which leukopenia is induced by repeated injection of cyclophosphamide, as well as corticosteroid-treatment models (112) are most frequently applied. The leukopenic model is characterized by a paucity of tissue inflammation, a rapid germination of conidia, and a boundless invasive fungal growth in lung tissues (111). In the corticosteroid model, neutrophils are recruited to the site of infection and incite a strong inflammatory response (111). Depending on the infectious inoculum of A. fumigatus conidia, both models eventually result in 100% mortality. To follow the course of A. terreus infection in a murine model of bronchopulmonary aspergillosis, a bioluminescent A. terreus strain was generated (110) allowing in vivo real-time monitoring of disease progression (113). In the leukopenic model, a dose of 5 × 10⁶ A. terreus conidia resulted in 100% mortality 2 weeks postinfection. In contrast, a 50-times-lower inoculum dose of A. fumigatus presented a more rapid disease progression, with all mice succumbing within 4 days (113). A. terreus bioluminescence imaging lacked an increase in the pulmonary bioluminescence signal during the first 5 days of infection; thereafter, the signal steadily increased until mice died due to A. terreus infection. These data indicate a transient restriction of conidial germination. In agreement, histopathologic analyses found resting and partially swollen conidia inside epithelial cells and resident alveolar macrophages with invasive mycelium developing during subsequent days (110). In contrast, such a long lag phase is not observed in infections with A. fumigatus, which usually displays germ tubes within 24 h and hyphal tissue invasion within 3 days of infection (111).

Virulence analysis of the corticosteroid model showed that 1 × 10⁷ A. terreus conidia resulted in 50% mortality within the first week after infection; mice surviving this acute phase completely recovered within 2 weeks. Consistently, imaging studies showed a transient bioluminescence increase which declined in survivors (110). Histopathology confirmed an initial conidial persistence in epithelial and immune cells (up to 48 h), with germination and hyphal tissue penetration resulting in the death of some animals. Survivors cleared invading hyphae within 2 weeks with only scant immune cell infiltrates being visible. At this time point, viable conidia residing within macrophages were still detected (110). These data point toward a failure of the immune system to eliminate A. terreus conidia. The administration of 1 × 10⁷ A. terreus conidia in immunocompetent mice confirmed this assumption, as up to 5 × 10⁵ viable conidia were reisolated from lung tissue 5 days postinfection (114). In contrast, immunocompetent mice infected with 2.5 × 10⁷ A. fumigatus conidia cleared the number of viable conidia to 500 CFU within 4 days (115); mice infected with 5 × 10⁷ A. fumigatus conidia were free of pathogens after 10 days (116). These results show that A. terreus causes invasive aspergillosis in infection models previously developed for A. fumigatus. However, in comparison to A. fumigatus, disease progression with A. terreus is slower, a higher infectious dose is required for 100% mortality, and a significant proportion of conidia shows long-term persistence in infected lungs. These observations underline striking differences in the immune interaction between A. terreus and A. fumigatus conidia.

Immune Cell Interaction and Contribution to Dissemination

Observations in murine infection models revealed long-term persistence of A. terreus conidia in epithelial and immune cells of the infected lung. This was observed not only in corticosteroid-treated and immunocompetent mice but also in mice treated with cyclophosphamide, which prevents the recruitment of immune effector cells to the site of infection but keeps resident alveolar macrophages mostly intact (110, 114). Usually, conidia of A. fumigatus are phagocytosed by immune effector cells, germinate, and escape by producing hyphae that pierce the macrophage membrane (114, 117, 118). Comparative in vitro studies of A. fumigatus and A. terreus with human monocyte-derived macrophages and murine macrophage cell lines observed that phagocytosis of A. terreus conidia proceeds faster than that with A. fumigatus (114). Recognition of β-glucans on the surface of A. fumigatus conidia is shielded by the conidial pigment dihydroxynaphthalene-melanin (DHN-melanin), which also inhibits the so-called LC3-associated phagocytosis (119). The DHN-melanin layer enables A. fumigatus conidia to evade recognition by immune cells. Studies showed DHN-melanin to be essential for the inhibition of acidification of phagolysosomes, whereby phagocytosis of melanin-deficient pksP (also known as alb1) mutants resulted in phagolysosome acidification (120, 121). Wild-type A. fumigatus conidia initiate the swelling process, which causes the fungus to be vulnerable to antifungal defense mechanisms. In agreement, in vitro experiments with primary macrophages and macrophage cell lines revealed an inactivation of conidia (40% to 70%) 10 h after phagocytosis (114). The surviving A. fumigatus conidia produced germ tubes and escaped by piercing the membrane of macrophages (114).

Rapid phagocytosis of A. terreus conidia by macrophages occurs via Dectin-1 and mannose receptors, as the combined blocking of these receptors reduced phagocytosis of A. terreus up to 90% (114). Studies showed A. terreus conidia to be rapidly phagocytosed, followed by phagolysosome formation and acidification. Surprisingly, conidia remain dormant in acidified phagolysosomes and hardly produce hyphae (lack of germination) that penetrate the macrophage membrane (114). This dormancy is associated with resistance of the A. terreus conidia to acidification within the phagolysosome environment; 12 h after incubation with macrophages, conidia were still fully viable (114). These data agree with murine infection models in which viable conidia were successfully reisolated from both immunocompetent and immunosuppressed mice (110).

In contrast to macrophages, the primary purpose of dendritic cells (DCs) is not pathogen inactivation but the presentation of fungal antigens for further recruitment of effector immune cells, thereby forming a bridge between innate and adaptive immune response (122). Interestingly, while macrophages actively acidify the phagolysosome to a pH of 5 (123), DCs show an incomplete assembly of the acidifying vacuolar ATPase (vATPase) and, in addition, phagolysosomes recruit the NAPH oxidase 2 (NOX2), which consumes protons and inhibits acidification (124). Therefore, the phagolysosome of DCs does not show the typical acidification observed with macrophages. When A. fumigatus and A. terreus conidia were challenged with either monocyte-derived human DCs or DCs from the MUTZ-3 cell line, A. fumigatus conidia successfully escaped and A. terreus conidia persisted (125). Despite the lack of phagolysosome acidification in DCs, A. fumigatus conidia, having failed to escape, showed loss of fitness. Furthermore, the viability of A. fumigatus conidia was reduced by 99% within 24 h when coincubated with amphotericin B or voriconazole. In contrast, in the presence of DCs and antifungals, A. terreus conidia remained viable (40% to 60%) after 24 h. Even after 72 h in the presence of DCs and coincubation with antifungals, approximately 30% of A. terreus conidia were still viable (125). These A. terreus conidia remained in a resting state after phagocytosis by DCs and did not induce the expression of DC activation markers as observed with A. fumigatus (125).

Since A. terreus conidia remain dormant in both macrophages and DCs, the phagolysosomal pH is unlikely to account for dormancy within immune cells. For germination of A. terreus, a slightly acidic pH is preferred over an alkaline pH. Increasing the pH of growth medium from pH 6 to pH 8 reduced the germination speed. However, while phagolysosomes contain a number of nutrients, nutrient supply is restricted to the fungus, which results in a carbon-limited environment (126, 127). Interestingly, A. fumigatus conidia require a much lower nutrient threshold than A. terreus to initiate rapid germination. When the standard RPMI glucose medium was diluted 64-fold, hardly any delay in germination was observed with A. fumigatus and an increase in optical density was already observed after a lag phase of about 5 h (125). In contrast, with A. terreus, the lag phase in undiluted RPMI medium was about 8 h, indicating a comparably slower initiation of germination, but, most strikingly, every single dilution step further delayed germination and a 64-fold dilution extended the germination time to 24 h (125). Whether this difference is due to a reduced permeability of A. terreus compared to that of A. fumigatus conidia or to a general difference in nutrient sensing has not yet been investigated. Additional studies showed that limitation in either glucose or nitrogen is sufficient to delay germination of A. terreus conidia to a much greater extent than observed for A. fumigatus (125). This indicates that nutrient restriction in phagolysosomes alone is likely to be sufficient to keep A. terreus conidia in a resting state.

Not only is a prolonged resting state inside the phagolysosomes of immune cells a prerequisite for the persistence of A. terreus conidia, but this dormancy protects them from antifungal treatment (125) and may account for the difficulties in treatment of A. terreus infections (128). In addition, persistence within the immune effector cells may contribute to fungal dissemination. Activated DCs are known to migrate to draining lymph nodes, to present antigens, and to stimulate T lymphocytes (129). However, this migration is counteracted by the fact that A. terreus conidia may not trigger DC activation, which is unlike A. fumigatus (125). The lack of activation is not associated with surface markers on resting conidia of both species but is rather a result of missing additional surface markers during swelling and germination. This finding is consistent with the observation that UV-inactivated conidia from both species failed to activate DCs (125). Therefore, the question remains open whether DCs will start to migrate from the site of infection to draining lymph nodes in A. terreus-related infections. In this respect, it is well known that the proinflammatory cytokine tumor necrosis factor alpha (TNF-α) by itself is sufficient to activate the migration of DCs (130). Such a TNF-α-mediated activation is also observed with DCs carrying internalized A. terreus conidia, whereby TNF-α stimulation did not affect the viability of the intracellular conidia (125). An increase in circulating TNF-α was found in corticosteroid-treated mice showing an acute A. terreus lung infection (110). As only a small proportion of extracellular conidia caused invasive aspergillosis in this model and conidia residing in immune cells were still detected, the inflammatory response could potentially trigger A. terreus dissemination by immune cells acting as a vehicle, as in the “Trojan horse” phenomenon. The “sit and wait” strategy of A. terreus conidia in comparison to the “hide and escape” strategy of A. fumigatus conidia may have a significant impact on disease manifestations, propensity for dissemination, and effectiveness of antifungal treatment. It may also explain the observation that uptake of A. terreus conidia could potentially occur long before hospitalization. The persistence of conidia may support the onset of infection following immunosuppressive regimens.

Role of Accessory Conidia in Dissemination

The production of a second type of asexual conidia, termed accessory conidia, is another possible contributor to dissemination in invasive aspergillosis caused by A. terreus. These accessory conidia are directly produced from vegetative hyphae (24) and do not require the formation of conidiophores (131). Environmental and clinical A. terreus isolates are capable of producing accessory conidia in submerged cultures (24). The production of accessory conidia is not limited to A. terreus but seems to be a general feature of the A. terreus species complex, as their formation has been described for A. citrinoterreus, A. hortai, A. alabamensis, and A. neoafricanicus (102). Accessory conidia have also been described for all species of the Aspergillus section Flavipedes, which seem to form an evolutionary sister branch of the section Terrei (9, 132); in addition, there are other similarities, such as the production of pale yellow to brown conidia (132) and the biosynthesis of some special secondary metabolites such as lovastatin (28, 133) or (iso-)flavipucine (134, 135). However, it is unlikely that accessory conidia are the major infectious propagules for any of these species, as they seem less adapted for dispersion. While accessory conidia, in contrast to phialidic conidia, present a thickened cell wall, they lack a melanin layer, and the smooth surface appearance implies the lack of water-repellent surface hydrophobins that prevent them from becoming airborne (24). Furthermore, dormant accessory conidia are 4 to 7 μm in size, which is approximately twice the size of phialidic conidia (24), are multinucleated, and are able to germinate within 2 h, implying a high basal metabolic activity (26). Accessory conidia may be produced from invasively growing hyphae in infected tissues, as demonstrated by histopathologic analyses of murine lung tissues (110). By displaying much more extensive immunogenic β-1,3-glucan on their surface than phialidic conidia (24, 26), accessory conidia elicit a very strong inflammatory response in infected lung tissues and in the interaction with macrophages (26). This inflammatory response may activate dendritic cells with engulfed phialidic conidia to migrate to draining lymph nodes (125). Moreover, hyphae penetrating blood vessels can release attached accessory conidia into the bloodstream, thus potentially adding to the dissemination of fungal elements in acute bronchopulmonary aspergillosis to secondary sites of infection.

Studying the role of cryptic species of section Terrei (A. alabamensis, A. citrinoterreus, A. floccosus, A. hortai, and A. neoafricanus) in G. mellonella infection models showed that all species except A. floccosus formed accessory conidia, whereby germination kinetics were variable. Accessory conidia were detected in larval tissue 96 h postinfection. In comparative analyses, a significantly higher virulence potential of accessory conidia was found for one A. hortai isolate, leading to the conclusion that cryptic species have a similar potential to cause infections as observed for A. terreus (102).

Aspulvinone E-Melanin Versus Dihydroxynaphthalene-Melanin

1,8-Dihydroxynaphthalene (DHN)-melanin is described as a major virulence determinant of A. fumigatus by contributing to resistance against UV light and oxidative stress (136, 137), increasing the attachment of hydrophobins to the conidial surface (138), inhibiting the acidification of macrophage phagolysosomes (118), and blocking the NADP (NADPH) oxidase-dependent activation of LC3-associated phagocytosis (119). However, in addition to these protective effects of A. fumigatus DHN-melanin, it has been shown that MelLec, a melanin-sensing C-type lectin receptor, promotes host immunity by recognizing 1,8-dihydroxynaphthalene units of DHN-melanin and polyketide-derived DHN-melanin precursor molecules (139). Figure 3A shows a detailed summary of the current knowledge on DHN-melanin biosynthesis in A. fumigatus (140). In mice, the MelLec receptor is found on murine lung and liver epithelial cells, whereas in humans, it is presented on myeloid cells (139). A deletion of this receptor in C57BL/6 mice resulted in reduced neutrophil recruitment in immunocompetent mice when challenged intratracheally with A. fumigatus. In addition, an increased fungal burden was observed in several organs, including the brain, which implies that MelLec plays a role in the control of fungal growth (139). Indeed, human stem cell recipients receiving donor cells with a common missense single nucleotide polymorphism at the C terminus of MelLec appear to be at higher risk for development of severe, frequently disseminated invasive aspergillosis (139). These observations indicate a mutual interaction of DHN-melanin in supporting both the pathogen in successful host infection and the host in recognition of and defending against fungal infections. Most non-terreus Aspergillus species seem to produce a conidial pigment that is derived from the DHN-melanin precursor heptaketide YWA1 (Fig. 3A) and hence is recognized by the MelLec receptor (139).

FIG 3.

Schematic representation of the DHN-melanin and Asp-melanin biosynthesis pathways. (A) DHN-melanin biosynthesis pathway from Aspergillus fumigatus. The polyketide synthase Alb1 (also called PksP) uses acetyl-CoA and malonyl-CoA units for the synthesis of the heptaketide YWA1. The enzyme Ayg1 is an α/β hydrolase that truncates YWA1 to the pentaketide 1,3,6,8-tetrahydroxynaphthalene (1,3,6,8-THN). The reductase Arp2 produces scytalone, which undergoes dehydration by Arp1 to form 1,3,8-trihydroxynaphthalene (1,3,8-THN), which acts as a substrate again for Arp2, producing vermelone. The vermelone dehydratase Abr1 produces 1,8-dihydroxynaphthalene (1,8-DHN), which is activated by the laccase Abr2 for polymerization. (B) Asp-melanin biosynthesis pathway from A. terreus. The nonreducing nonribosomal peptide synthetase-like enzyme (NRPS-like enzyme) MelA uses two molecules of p-hydroxyphenylpyruvate for the production of aspulvinone E. A tyrosinase oxidizes either one or both hydroxyphenyl rings of aspulvinone E in the m position, which results in autopolymerization.

Several observations indicate that pigment biosynthesis in A. terreus conidia is different from the DHN-melanin biosynthesis pathways in A. fumigatus (Fig. 3A) and related species. In macrophages, the acidification of phagolysosomes is not inhibited by A. terreus conidia; however, inhibition of acidification is observed when heterologously expressing the YWA1 synthase wA from A. nidulans in A. terreus (114). A search for a polyketide synthase with homology to PksP (also called Alb1) from A. fumigatus (Fig. 3A) or wild type from A. nidulans revealed a distantly related candidate that was later identified as terrein synthase TerA (141). Therefore, all genes encoding key enzymes of secondary metabolite biosynthesis were screened for their expression under conidium-forming conditions, which led to the identification of a small gene cluster composed of the nonreducing nonribosomal peptide synthetase-like (NRPS-like) enzyme MelA and its adjacent tyrosinase TyrP (142) (Fig. 3B). Deletion of the melA gene resulted in unpigmented white conidia (142, 143), whereas deletion of tyrP generated yellow conidia showing a bright yellow-green fluorescence under UV light (142). A more detailed analysis of the metabolites produced by MelA revealed two molecules of the aromatic alpha-keto acid p-hydroxyphenylpyruvate to yield aspulvinone E (Asp) (142, 144). Further hydroxylation of aspulvinone E at one or both aromatic rings via the tyrosinase TyrP results in an activation of aspulvinone E for autopolymerization, leading to the final pigment Asp-melanin (142) (Fig. 3B). As Asp-melanin is unrelated to DHN-melanin in terms of synthesis and structure, as shown in Fig. 3, it is unlikely that the pigment has the same properties in immune interactions as described for A. fumigatus DHN-melanin. However, Asp-melanin also protects from UV light, and phagocytosis studies with soil amoeba revealed a reduced uptake of pigmented versus unpigmented A. terreus (142), which could have relevance in environmental competition and survival. Nevertheless, its direct contribution to virulence appears limited. Wild-type A. terreus conidia are rapidly phagocytosed by macrophages, which implies that the Asp-melanin pigment does not prevent the recognition of surface polysaccharides by macrophages (114), but the phagocytosis rate, oxidative stress, and cytokine response of macrophages toward A. terreus albino mutants has not yet been studied. In addition, and in contrast to A. fumigatus, A. terreus mutants with a defect in pigment formation show no increased sensitivity toward oxidative stress, which is one of the major mechanisms of phagocytes to inactivate pathogens (142). Lastly, a virulence screening in an embryonated chicken egg infection model revealed no virulence attenuation of A. terreus pigment mutants (142). However, whether this lack of virulence attenuation also transfers to a mammalian infection model needs further investigation. Also, the postulated lack of recognition of Asp-melanin by the MelLec receptor requires further studies.

Secondary Metabolites: Lovastatin and Terrein

The cholesterol-lowering drugs lovastatin (145) and terrein are probably the most prominent metabolites produced by this fungus (141). However, investigation of specific secondary metabolites contributing to the pathogenesis of A. terreus has not been explored in detail (146). Therefore, it is only possible to speculate on the contribution of secondary metabolites to pathogenesis. Histopathologic analyses observed in mice succumbing to A. terreus invasive pulmonary aspergillosis showed a fatty degeneration of liver hepatocytes (110). Fungal elements were not detected in the liver, and hence, it appears more likely that secondary metabolites were produced during invasive growth in the lungs, then circulated in blood, and eventually resulted in liver damage.

Lovastatin is a competitive inhibitor of the hydroxymethylglutaryl-coenzyme A (hydroxymethylglutaryl-CoA) reductase (HMGR) (147), which is essential for the production of cholesterol in humans and the cholesterol equivalent ergosterol in fungi. Lovastatin belongs to the type 1 statins, like simvastatin, mevastatin, and pravastatin, and was the first FDA-approved statin drug (148). While the lovastatin biosynthesis gene cluster is composed of approximately 18 genes, only 5 appear to be directly involved in lovastatin biosynthesis. These are the genes coding for two protein kinases (PKs), LovB and LovF, the enoyl reductase LovC, the transesterase LovD, and the cytochrome P450 oxygenase LovA (149, 150). Interestingly, as an HMGR inhibitor, lovastatin might also have a toxic effect on A. terreus by inhibiting ergosterol biosynthesis. To counteract this effect, gene clusters contain the lvrA gene. The encoded protein shows homology to HMGR and confers resistance to lovastatin (150). Furthermore, the lovastatin biosynthesis gene cluster contains the transcription factors LovE and LovH, but it has not yet been investigated whether lovastatin biosynthesis is active during invasive growth in lung tissues.

It has been shown that lovastatin administration to mice results in an expected decrease in plasma cholesterol levels, but a paradoxical increase in liver cholesterol production accompanied by increased liver weight and fecal output was observed (151). Thus, lovastatin production during lung infection may be a contributing factor to the fatty liver degeneration observed in mice (110).

Terrein is a polyketide-derived metabolite with the underlying polyketide synthase identified during a search for genes related to pigment production in A. terreus conidia (141). A. terreus isolates produce several grams of terrein per liter of culture medium (141, 152, 153). The terrein biosynthesis gene cluster contains its own transcriptional regulator, TerR (154), which is activated independently by either methionine, nitrogen, or iron starvation via the global transcription factors AreA, AtfA, and HapX (155). Terrein displays some antifungal properties, reduces ferric to ferrous iron, and supports breakdown of fruit tissues (155). Hence, terrein production is assumed to provide a competitive advantage for both growth in the soil rhizosphere and fruit infection. However, as iron and nitrogen are also limited in infected tissues, terrein production during invasive growth cannot be excluded. Besides antiproliferative and apoptotic properties of terrein against several cancer cell lines (29, 156–159), terrein has anti-inflammatory properties by inhibiting interleukin 6 (IL-6) signaling (160, 161). Blocking IL-6 signaling cascades reduces angiogenesis, which seems important for effective drug targeting of fungal elements in necrotic tissues (162). Therefore, A. terreus terrein might take over the role of the metabolites fumagillin and gliotoxin produced by A. fumigatus and, although targeting different pathways, also block angiogenesis (163, 164).

In conclusion, many other secondary metabolites produced by A. terreus could add to persistence within immune cells, evade effective antifungal immune responses, and counteract antifungal therapy. However, without a detailed analysis of the gene clusters induced during pathogenesis and the metabolites released, the role of secondary metabolites in pathogenesis remains speculative.

HUMAN INFECTIONS AND BEYOND

The signs and symptoms of aspergillosis vary with the underlying type of illness (1), and the disease occurs mainly in humans whose immune systems are weakened (165). A. terreus-related invasive infections are also associated with tissue damage, surgery, or foreign body implants (27).

The at-risk population for invasive infections due to Aspergillus section Terrei does not differ much from patients suffering from diseases caused by A. fumigatus (166) (Table 2). Similar to invasive aspergillosis caused by A. fumigatus, severe and prolonged neutropenia is the dominant risk factor for invasive aspergillosis due to A. terreus (167); the A. terreus species complex also affects nonimmunocompromised individuals (168–170). A recent worldwide prospective survey showed A. terreus to be associated with chronic lung diseases (39%), followed by, among others, hematologic or oncologic malignancies, solid organ transplantation, and diabetes mellitus, each 7% (2). The spectrum of A. terreus-related diseases comprises invasive aspergillosis (25%), allergic bronchopulmonary aspergillosis (12%), chronic aspergillosis (11%), chronic obstructive pulmonary disease exacerbation (5%), aspergilloma (4%), otitis externa (3%), and others. The incidence of A. terreus-related diseases varied remarkably from center to center and within centers (2, 171, 172). However, controlled clinical studies are necessary to define the real threat of A. terreus in the onset of chronic pulmonary aspergillosis.

TABLE 2.

Head-to-head comparison of A. fumigatus species complex and A. terreus species complex^a

Variable	Characteristics of:
	A. fumigatus species complex	A. terreus species complex
Pathogen-specific feature	Phialidic conidia	Phialidic conidia and aleurioconidia (accessory conidia)—thought to be a contributor to dissemination.
Environmental source	Predominant pathogen, ubiquitous in the environment	Found in soil, compost, and dust to a lesser extent than A. fumigatus; common in the marine environment, in India, and other restricted areas, such as Innsbruck, Austria; the reason for the relative frequency is not yet clear.
		Potted plants were associated with an outbreak within the hospital setting in Austria.
		Other efforts (Houston, TX) to identify a common source failed to be successful.
Epidemiology of infections	Worldwide occurrence	Worldwide occurrence but less frequent than A. fumigatus; third most important species in some regions; an emerging pathogen?
		Common in certain geographic locations (e.g., Innsbruck, Austria, and Houston, TX).
		Reports from Spain are on the rise.
		Cryptic species accounted for 10% to 15%, with A. citrinoterreus being the most prevalent.
Age, gender	No differences	No differences
Invasive aspergillosis	Pulmonary diseases most important, most common in hematopoietic stem cell transplantation, leukemia, and neutropenia.	Pulmonary diseases most important; most common in leukemia and hematopoietic stem cell transplantation.
	Mortality, 30% to 40%.	High propensity for dissemination; mortality, 70% to 51%; dissemination to heart and brain is common.
	Overall rate of responses ranges from 39% to 53%.	Overall rate of responses ranges from 16% to 28%.
	Overall incidence of invasive aspergillosis has decreased since the implementation of azoles as antifungal prophylaxis and early empirical or preemptive therapy with voriconazole.	Multivariate analyses showed poor outcome in patients with hematological malignancy with amphotericin B as well as azole therapy; posaconazole salvage treatment showed encouraging data.
		Occurs more often than A. fumigatus as breakthrough infection on antifungal azole prophylaxis (21%).
		Overall incidence has decreased since the implementation of antifungal prophylaxis or early empirical or preemptive therapy with voriconazole; in Austria, A. terreus infections in hematological patients disappeared but A. terreus is prevalent in nonpulmonary infections.
Solid organ transplantation	Leading pathogen in lungs or other solid organ transplantations.	Rare but, if it occurs, is most common in lung and liver transplantations.
Aspergilloma	Leading pathogen	Rarely caused by A. flavus, A. oryzae, A. terreus, or A. nidulans.
Chronic pulmonary aspergillosis	Leading pathogen	Few individual cases but is increasingly described (third most important cause in some countries).
Cystic fibrosis and allergic bronchopulmonary aspergillosis (ABPA)	Leading pathogen. Invasive cases are uncommon, several serological tests are the cornerstone for ABPA, and cross-reactivity may exist between antibodies to different Aspergillus spp.; therefore, results should be carefully evaluated.	A. terreus may colonize the airways of patients with cystic fibrosis; patients may be sensitized by several fungi, including A. terreus, and allergic manifestations are documented; isolated clinical cases of invasive aspergillosis in cystic fibrosis exist.
Onychomycosis	Few individual cases	Frequently involved; A. terreus along with A. flavus and A. niger the most common; prevalence ranges from 7% to 100%; the clinical presentation is unspecific.
Otomycosis	A. niger and A. flavus are most common; geographical differences exist.	Several individual cases; most common in refractory otomycosis in a retrospective study.
Sinus	A. fumigatus is most common; geographical differences exist.	A. terreus is rarely involved.
Other forms (e.g., endocarditis, meningitis, bursitis)	Unusual forms exist, with some in the immunocompromised host but most are not.	Unusual forms exist, and a small number of isolated cases exist but mainly in immunocompromised hosts.

^aData compiled from D. W. Denning (207). Data also taken from references 1, 2, 83, 165, 167, 168, 208 to 212.

Diagnosing aspergillosis is difficult, as symptoms and signs lack specificity and sensitivity (1, 173). Chest X-ray and, even more importantly, computerized tomography (CT) scans are standards of care, but imaging does not differentiate between A. fumigatus, A. terreus, or other lung pathogens (174). Culture and microscopy in sputum or bronchoalveolar lavage fluid in combination with antigen- and possibly molecular-based assays are the modern way to diagnose aspergillosis (175). Importantly, A. terreus is the only cause of true aspergillemia, possibly because of the presence of adventitious conidia (173). A few studies focused on the performance of fungal biomarkers in A. terreus infections and observed that baseline serum galactomannan (GM) levels did not differ among the various Aspergillus species (171). In contrast, another report showed GM testing to be more suitable for detecting aspergillosis due to A. fumigatus than due to non-fumigatus Aspergillus species (172).

Treatment varies depending on the type of underlying fungal disease and includes antifungal therapy, surgery, or observation. Standard treatment of invasive pulmonary aspergillosis consists of mold-active triazoles (voriconazole, isavuconazole, posaconazole) or lipid formulations of amphotericin B (liposomal amphotericin B) (174). On the other hand, although echinocandins have in vitro activity against A. terreus, these agents are considered inferior for the therapy of invasive aspergillosis (174). Under treatment with amphotericin B, species identification is necessary, as the A. terreus species complex is usually resistant in vitro and in vivo against polyenes (56, 175). Overall, survival rates were greater for patients treated with triazoles (voriconazole, itraconazole) than for those treated with amphotericin B-based regimens (176). Limited data are available for other azoles; four out of nine patients (44%) were treated with posaconazole (salvage therapy) and responded favorably, while the success rate of lipid amphotericin B formulations was only 19%. Hachem et al. analyzed 513 patients with hematological malignancies and evaluated baseline characteristics, antifungal therapy, and outcome of individuals infected with A. terreus and non-A. terreus species (176). Diseases due to A. terreus were associated with a lower response rate to treatment and a higher rate of mortality than those non-A. terreus infections. The fact that A. terreus infections in comparison to non-A. terreus infections occur more often as fungal breakthrough infections is of interest (167). Treatment strategies should be based on the institutional epidemiology of fungal infections and the assessment of individual risk factors (56, 175). Fundamental differences between A. terreus and other, non-A. terreus species may explain their differences in antifungal susceptibility, the outcome of disease, and epidemiology (Table 2).

Most animal cases are sporadic, with the environment being the primary source of infection (177). Several well-defined clinical entities have been described, covering local and systemic diseases. Disseminated canine aspergillosis affects primarily German Shepherd breed dogs, with females being overrepresented (178). The main pathogens are A. terreus and, less frequently, A. carneus and A. alabamensis. A. fumigatus is the most prevalent etiology, although other species such as A. terreus have been reported in sinonasal aspergillosis, in dolichocephalic and mesocephalic dogs (179, 180), in the guttural pouch mycosis (181), in mycotic abortion in horses (182) and pigs (183), in mycotic keratitis (184, 185), and in avian aspergillosis (186). Heavy fungal loads are hypothesized to be the primary sources of massive poultry infections, whereas disseminated or rhino-nasal/rhino-orbital aspergillosis is associated with breed predisposition and immune deficiencies (187).

A. terreus—also known as a plant pathogen—has been responsible for severe loss of crops worldwide, annually damaging over 125 million tons of food grains (188). The saprobic nature of A. terreus enables this fungus to colonize sugar beet (Beta vulgaris), Jew’s mallow (Corchorus olitorius), cucumber (Cucumis sativus), purslane (Portulaca oleracea), and eggplants (Solanum melongena) (189). In contrast, A. terreus has been used as a biocontrol agent in protection against the snail Biomphalaria alexandrina (24), operates as a growth promoter of plants, and finally attenuates stem rot diseases caused by Sclerotium rolfsii (190).

AMPHOTERICIN B RESISTANCE/TOLERANCE

Amphotericin B shows a powerful and broad fungicidal activity against a vast variety of fungi and has a remarkably low microbial resistance rate. A. terreus is the prototype of an innately polyene-resistant fungus (191, 192). Hence, the majority of A. terreus species complex strains are intrinsically resistant to amphotericin B (amphotericin B-resistant A. terreus [ATR]), and MICs are generally ≥2 mg/liter, independent of the antifungal susceptibility testing method used (Clinical and Laboratory Standards Institute, European Committee on Antimicrobial Susceptibility Testing, Epsilometer test [E test], or others) (193). As amphotericin B resistance in A. terreus is not genetically transferable, the appropriate term is also tolerance or tolerance/resistance. In contrast, a few A. terreus strains show low MICs (0.06 to 0.125 mg/liter) and hence are referred to as amphotericin B-susceptible A. terreus (ATS) (194). Specifically, 8% to 13% of A. terreus isolates collected worldwide exhibit low amphotericin B MICs, suggesting possible genetic differences within the section Terrei (56, 73, 79). The mode of action of amphotericin B is complex and multifaceted (195, 196); it is assumed that polyenes interact with fungal ergosterol and/or induce reactive oxygen species (ROS), hence leading to cellular damage. To date, no A. terreus-specific genomic features have provided an explanation for amphotericin B resistance/tolerance mechanisms. The interference with the mitochondria by ROS/anti-ROS seems to be the main driver of amphotericin B efficacy (Fig. 4). In general, amphotericin B resistance in section Terrei is associated with modulating molecular chaperones, targeting ROS via mitochondria, and shaping the cellular redox homeostasis (Table 3). The underlying mechanisms of ATR may be associated with the level of catalase production by this species in comparison to that of A. fumigatus (191, 195). Evaluating ATR and ATS showed that ATR possessed double basal superoxide dismutase (SOD) activity compared to that of ATS. In addition, ATR presented an enhanced oxidative stress response, with significantly higher sod2 mRNA levels and increased catalase (CAT) transcripts upon amphotericin B exposition. The inhibition of SOD and CAT proteins rendered ATR susceptible to the polyene in vitro (191, 195). The coapplication of anti- and pro-oxidants significantly affects amphotericin B efficacy in an antithetic manner (196). Antioxidants and ROS-scavenging agents increased amphotericin B tolerance in ATR, while pro-oxidants rescued polyene tolerance in ATS. These in vitro data were confirmed by G. mellonella infection models. Whether such management is of clinical interest for A. terreus diseases needs to be analyzed in detail.

FIG 4.

Proposed model of amphotericin B mode of action and amphotericin B resistance in Aspergillus terreus. Amphotericin B affects the cell membrane and mitochondrial (mt) function, which promotes the generation of reactive oxygen species (ROS). Influencing the redox status with l-ascorbic acid (AA) and N-acetyl-l-cysteine (NAC) impacts ROS levels and subsequent amphotericin B-mediated fungal cell damage.

TABLE 3.

Potential molecular targets of amphotericin B resistance in A. terreus and non-A. terreus molds^a

Potential target(s)	Comments
Cell membrane ergosterol	The role of ergosterol as resistance mechanism is unclear, as the content of ergosterol was similar for two species, ATR and polyene-susceptible A. fumigatus.
	ATR and ATS showed the same level of ergosterol content.
Cell wall composition	A. flavus resistant to amphotericin B showed an altered cell wall composition, and an increased β-1-3-glucan content was detected in the mutant strain.
	In a study of protoplasts from A. fumigatus, ATR and ATS showed no significant differences.
Mitochondrial functions	ATS displayed a 3-fold increase in O2 consumption compared to ATR.
	ATS showed an elevated mitochondrial DNA content (23%) compared to ATR.
	Amphotericin B resulted in an upregulation of mitochondrial genome content in ATS, whereas it was decreased in ATR.
Oxidative stress pathways	Catalase production is higher in ATR than in A. fumigatus.
	Amphotericin B mediated elevated ROS levels in ATS in comparison to ATR.
	NAC rescued amphotericin B-induced ROS increase as well as amphotericin B tolerance in ATS.
	AA showed contrasting results.
Molecular chaperone machinery	ATS showed only slightly increased HSP90 and HSP70 basal levels; ATR showed a strong reaction upon amphotericin B treatment, while ATS did not.
	HSP70 member SSB was up-regulated in ATR by amphotericin B exposition in a 2D proteome analysis.
	Western blot experiments revealed that ATR exhibited high basal levels of SSA and SSB HSP70 proteins.
	Amphotericin B resulted in a robust induction of HSP70 in ATR in comparison to ATS.
	Trichostatin A was particularly active against ATR.
	HSP90 inhibitors resulted in significant improvement in amphotericin B activity against ATR; MICs decreased from 32 to 0.38 mg/liter.

^aData taken from references 191, 195, 196, 213 to 217. ATR, amphotericin B-resistant Aspergillus terreus; ATS, amphotericin B-susceptible Aspergillus terreus; ROS, reactive oxygen species; NAC, N-acetylcysteine; AA, l-ascorbic acid; HSP, heat shock proteins; 2D, two dimensional.

In general, the prevalence of ATS in clinical specimens seems to be low; we showed that ATS occurred spontaneously as a sector product of ATR on drug-free medium (194). Spontaneous culture degeneration (sectoring) of ATR leads to the emergence of multiple sectors (ATSec) (Fig. 2D), which showed significantly lower MICs against amphotericin B (194). Differences in in vitro antifungal susceptibility of voriconazole and posaconazole were not observed. Overall, ATSec features ATS (morphologically). Antifungal susceptibility and virulence studies of ATSec, ATR, and ATS strains showed ATSec harboring amphotericin B MICs ranging from 0.12 to 0.5 μg/ml, and G. mellonella survival studies revealed an enhanced virulence of ATSec (104). Whether ATS occurs as a lab by-product via spontaneous culture degeneration or may arise in vivo (e.g., during antifungal treatment) needs to be examined in more detail; several studies addressing this topic are ongoing. So far, the presence of sector formation in cultures needs to be carefully evaluated when preparing inocula for further studies; the accidental usage of “mixed inocula” (conidia from sector and nonsector parts) for MIC testing or virulence studies may have a crucial impact on study outcome.

Speth et al. studied the role of platelets on virulence and pathogenesis of ATR and ATS in an untreated mouse model (197); data obtained showed that ATS-infected mice died even faster than ATR-infected animals. Conidia, aleurioconidia, and hyphae of ATS were more potent than ATR to trigger thrombocyte stimulation, and thrombocytes adhered better to ATS than to ATR fungal structures. Hence, we hypothesize that the capacity of fungal isolates to modulate thrombocyte parameters contributes to its virulence in vivo. Perkhofer et al. (198) showed human platelets to potentiate the efficacy of amphotericin B against ATR in vitro.

Usually, A. terreus is fully susceptible against azole drugs (101, 199), displaying MICs lower than epidemiologic cutoff values (200). Azole resistance in Aspergillus section Terrei seems to be rare but was observed when a global set of 498 strains was screened (100). According to EUCAST epidemiological cutoffs, azole non-wild-type isolates were detected in 5.4% of all tested A. terreus section Terrei isolates, whereas 6.2% of A. terreus sensu stricto isolates showed in vitro resistance to posaconazole. Whether the latter finding correlates with an in vivo lack of efficacy of azoles is presently unknown. Austria, Germany, and the United Kingdom displayed a high frequency of posaconazole resistance (10.5%, 13.7%, and 12.5%, respectively) in all A. terreus sensu stricto isolates. Azole resistance among cryptic species as well as in vitro resistance against itraconazole and voriconazole was rare (100). Mutations of the lanosterol-14-α sterol-demethylase gene (Cyp51A), a key protein in the ergosterol biosynthesis pathway, were found (197, 198, 201).

CONCLUSION AND FUTURE PERSPECTIVES

Aspergillus is one of the most common causes of mold infections in humans and is able to cause a variety of diseases in both immunocompetent and immunocompromised patients. In general, A. fumigatus species complex is the most prevalent species, followed by A. flavus and A. terreus, but this depends on local epidemiology. This is especially true for the unique fungal epidemiology in certain areas such as Innsbruck, Austria, and Houston, Texas (USA). The exact reason for the high prevalence in these regions is not yet known but needs further analyses. Overall, there are multiple unanswered questions regarding the epidemiology, pathobiology, genomic structure, underlying resistance mechanisms, and impact of the environment on infections related to A. terreus species complex.

The more aggressive clinical behavior of this species than of A. fumigatus is associated with several unique features of A. terreus: (i) the production of aleurioconidia in infected tissues, which can be released into the bloodstream and foster dissemination; (ii) the persistence in immune cells, which allows acquisition of dormant conidia before hospitalization (sit-and-wait strategy) with possible breakthrough of infections under immunosuppressive regimen; and (iii) the production of unique secondary metabolites, such as the production of Asp-melanin rather than DHN-melanin, which leads to altered immune interactions and may contribute to increased dissemination rates.

It is well known that early diagnosis and Aspergillus-specific prophylaxis has decreased the incidence of invasive aspergillosis in the past. So far, the introduction of voriconazole has resulted in a reduction in invasive aspergillosis due to A. terreus (60) but seems to have caused a shift toward infections with Mucorales (202). In contrast, findings from our single-center study support that amphotericin B may select A. terreus infections. The true epidemiology of chronic forms of Aspergillus infections, including those with A. terreus, is unknown and potentially underestimated. The worldwide A. terreus survey showed chronic aspergillosis to be on the rise. The number of immunocompromised humans will increase in the near future as a result of improved medical care and aging, and hence, Aspergillus-related infections will continue to rise. A deregulation of the immune system builds the basis for aspergillosis, and understanding the mechanisms from the onset of colonization down to infection is greatly needed to prevent disease progression. Limited treatment options result in the growing awareness of drug-resistant fungi, including A. terreus. The interference with the fungal stress response pathways seems to be the main driver of regulating amphotericin B efficacy. As amphotericin B is often the first-line treatment for invasive fungal infections, a proper diagnosis of the underlying pathogen is of high importance. A. terreus is resistant in vivo and in vitro to amphotericin B. Whether the rare amphotericin B-susceptible representatives derive from culture degeneration (sectoring) and hence are simply a culture by-product or exist in vivo needs to be studied in more detail. So far, new drugs such as olorofim (F901318) are promising, as they show activity against a broad range of Aspergillus section Terrei (203) isolates in vitro and in vivo. The novel antifungal compounds T-2307, Fosmanogepix (E1210/APX001), ASP2397 (VL-2397), Rezafungin (CD101), and Ibrexafungerp (SCY078), which target A. terreus in vitro (204), need to be carefully evaluated for their clinical effectiveness.

Biographies

Cornelia Lass-Flörl, M.D., is the Director of the Institute of Hygiene and Medical Microbiology of the Medical University of Innsbruck. Her research interests focus on the epidemiology, diagnosis, prevention and therapy of fungal infections and antifungal susceptibility testing. Professor Dr. Lass-Flörl has authored several journal articles, abstracts, and book chapters and is a member of the Subcommittee on Antifungal Susceptibility Testing of the European Society of Clinical Microbiology and Infectious Diseases. She is a board member of the European Confederation of Medical Mycology (ECMM) Excellence Center Initiative. The Institute is designated Reference Centre for Aspergillus and Aspergillus Infections and ECMM Excellence Centre Diamond Status. Cornelia Lass-Flörl is the convenor of the Aspergillus terreus working group of the International Society for Human and Animal Mycology. She is a fellow of ESCMID, ECMM, and IDSA. In 2020, she was elected into the American Academy of Microbiology.

Anna-Maria Dietl, Ph.D., is a Postdoc at the Medical University of Innsbruck at the Institute of Hygiene and Medical Microbiology. She studied Molecular Cell and Developmental Biology and graduated from the Division of Molecular Biology at the Medical University of Innsbruck. Her research activities focus on the human-pathogenic molds Aspergillus fumigatus and Aspergillus terreus and on the identification and characterization of fungus-specific metabolic pathways, which could serve as potential targets in the development of new antifungal therapies. On these topics, she has authored several research articles in renowned international journals and was honored to receive some scholarships and awards. Her current research project focuses on the molecular epidemiology and antifungal susceptibility of A. terreus from environmental and clinical sources. Anna-Maria Dietl is a member of the ÖGMM and on the reviewer board of Fungi-Animal Interactions for Frontiers in Fungal Biology and Microorganisms.

Dimitrios P. Kontoyiannis, M.D., is the Robert C. Hickey Chair in Clinical Care, Deputy Head, Division of Internal Medicine and Deputy Head in the Division of Internal Medicine at MD Anderson Cancer Center in Houston, TX. Dr. Kontoyiannis has authored over 630 peer-reviewed manuscripts and has given over 350 lectures at national and international conferences and academic institutions in United States and abroad. He is considered the leading mycology expert worldwide (Expertscape), with an H index of 111 and over 50,000 citations. His research group is credited for many and sustained contributions to clinical, translational, and experimental mycology. He is the recipient of many national and international awards and is the past president elect of the Immunocompromised Host Society (2016–2018). He is the leader of the European Confederation of Medical Mycology (ECMM) Diamond Excellence in Mycology Center at MD Anderson Cancer Center, the only U.S. center to receive such a designation by the ECMM.

Matthias Brock, Ph.D., studied biology at the Johannes Gutenberg-University Mainz, Germany, and the Philipps University Marburg, Germany, where he graduated in 2001 in fungal metabolic physiology. He was then appointed lecturer at the University Hannover, Germany, where he investigated the impact of fungal nutrient acquisition on disease establishment. Matthias then worked at the Hans Knöll Institute in Jena, Germany, where he developed fungal bioluminescent reporter systems for in vivo real-time imaging of fungal infections and their application in in vivo monitoring of antifungal therapy. In Jena, he also started to investigate fungal natural product formation. He was then appointed lecturer at the University of Nottingham, United Kingdom, where he currently develops fungal expression platforms for the characterization of secondary metabolite biosynthesis gene clusters. Matthias is a member of the DMykG and the VAAM and is publications officer of the British Mycological Society. He also serves as editor of FEMS Microbiology Letters and the journal Pathogens.