Species Distribution and Antifungal Susceptibilities of Aspergillus Section Terrei Isolates in Clinical Samples from the United States and Description of Aspergillus pseudoalabamensis sp. nov.

Aspergillus section Terrei consists of numerous cryptic species in addition to A. terreus sensu stricto. The treatment of invasive infections caused by these fungi may pose a unique challenge prior to diagnosis and species identification, in that they are often clinically resistant to amphotericin B, with poor outcomes and low survival rates in patients treated with this polyene. Data on the species distributions and susceptibility profiles of isolates within section Terrei from the United States (U.S.) are limited. Here, we report the species distributions and susceptibility profiles for amphotericin B, isavuconazole, itraconazole, posaconazole, voriconazole, and micafungin against 278 clinical isolates of this section from institutions across the U.S. collected over a 52-month period. Species identification was performed by DNA sequence analysis and phenotypic characterization. Susceptibility testing was performed using the CLSI broth microdilution method. The majority of isolates were identified as Aspergillus terreus sensu stricto (69.8%), although several other cryptic species were also identified. Most were cultured from specimens collected from the respiratory tract. Posaconazole demonstrated the most potent activity of the azoles (MIC range ≤ 0.03–1 mg/L), followed by itraconazole (≤0.03–2 mg/L), voriconazole, and isavuconazole (0.125–8 mg/L for each). Amphotericin B demonstrated reduced in vitro susceptibility against this section (MIC range 0.25–8 mg/L), although this appeared to be species-dependent. A new species within this section, A. pseudoalabamensis, is also described. Our results, which are specific to the U.S., are similar to previous surveillance studies of the Aspergillus section Terrei.


Introduction
Aspergillus species are able to cause a wide range of infections, with most affecting the lungs, such as acute invasive pulmonary aspergillosis, chronic pulmonary aspergillosis, aspergillomas, and allergic bronchopulmonary aspergillosis. In addition, exacerbations of chronic obstructive pulmonary disease may also be caused by these filamentous fungi. Patients at risk include those with oncologic or hematologic malignancies; solid organ transplant recipients; those receiving immunosuppressive therapy, including prolonged courses of corticosteroids; individuals with chronic pulmonary conditions, such as those with structural changes within the lung due to tuberculosis or sarcoidosis; and those with viral pneumonia (e.g., influenza or COVID-19) who require mechanical ventilation [1][2][3][4]. In epidemiologic studies, the primary causes of invasive aspergillosis include A. fumigatus, A. flavus, A. niger, and A. terreus [5][6][7]. A. terreus is of unique clinical importance since it is considered to be clinically resistant to amphotericin B, with poor outcomes reported in patients who have received this polyene as antifungal therapy [8,9]. Thus, in patients who are not able to be administered azoles due to adverse effects/toxicities or due to drugdrug interactions, treatment options are limited when infections are caused by A. terreus. Although the incidence of invasive aspergillosis caused by this species is relatively small compared to disease caused by A. fumigatus, higher rates of infection have been reported at certain centers, including ones in Innsbruck, Austria, and in Houston, Texas [10][11][12][13].
Aspergillus section Terrei is composed of numerous distinct species, including A. terreus sensu stricto, A. citrinoterreus, A. hortae, A. alabamensis, and A. neoafricanus, among others [12][13][14]. Relatively little is known about the species distribution of this section or their antifungal susceptibility profiles in clinical isolates from United States (U.S.). Although a large, international surveillance study has been published for Aspergillus section Terrei, which included a total of 370 isolates, only 10 of these were from North America [13]. Our objective was to evaluate the species distribution and antifungal susceptibilities of a large collection of Aspergillus section Terrei clinical isolates from institutions across the U.S. over a multi-year period in order to improve our understanding of this section and determine whether there may be differences between the susceptibility profiles of different drugs. We also describe a new species within the Aspergillus section Terrei, A. pseudoalabamensis sp. nov., which was cultured from different patients from a specific state in the U.S.

Fungal Isolates
In this prospective study conducted between October 2015 and January 2020, the sequential isolates of the Aspergillus species received in our reference mycology laboratory (Fungus Testing Laboratory, UT Health San Antonio) for species identification or antifungal susceptibility testing were included. Serial isolates from the same patients, and those from animals, environmental sampling, and nails and ears were excluded, with the exception of those included in the description of the new species. Subcultures of all isolates were prepared onto potato flakes agar on the day of receipt prior to further testing [15].

Species Identification
All isolates were identified at the species level by combined DNA sequence analysis and phenotypic characteristics (i.e., colony and microscopic morphology). In preparation for DNA sequence analysis, isolates were suspended in Buffer G2 (Qiagen, Valencia, CA, USA) followed by lysis using a bead beater instrument (Precellys Evolution, Bertin Instruments, Rockville, MD, USA) and the addition of proteinase K and incubation at 56 • C [16,17]. An EZ1 DNA tissue kit and a BioRobot EZ1 instrument (Qiagen) was then used to perform DNA extraction. The partial calmodulin (CaM) and β-tubulin (BenA) genes were then sequenced using the previously published primer pairs CF1L (5-GCCGACTCTTTGACYGARGAR-3) and CF4 (5-TTTYTGCATCATRAGYTGGAC-3) and Bt2a (5-TGACCCAGCAGATGTT-3) and Bt2b (5-GTTGTTGGGAATCCACTC-3), respectively [18]. BLASTn searches of the sequences were then performed in GenBank, and results with at least a 90% query coverage were considered significant with an E-value of 0.0 at 98-100% identity.

Description of New Species
Further phylogenetic analysis and phenotypic workup was performed on five additional isolates, two from the same patient, that were unable to be identified at the species level. For phenotypic and morphologic observations, malt extract agar (MEA, Oxoid Limited, Hampshire, UK), creatine agar (CREA), Czapek's yeast autolysate (CYA), and potato flake agar (PFA) were used. The isolates were inoculated at three points on each plate of each medium and incubated at 25 • C, 37 • C, and 40 • C in the dark for 7 days. Colony features were examined after 7 days of incubation and colony color were recorded following the color descriptions in the Methuen Handbook of Colour [19]. For micromorphological observations, microscopic mounts were made in lactophenol cotton blue from Pathogens 2023, 12, 579 3 of 15 MEA colonies. Slide cultures were also performed, and scanning electron microscopic (SEM) images were also obtained from mounts taken from MEA colonies.
For phylogenetic analysis, cultures were grown on PFA slants and incubated at 25 • C for 7 days. DNA was extracted from the harvested mycelia following previously described methods [20]. The ITS region and parts of three gene regions, BenA, CaM, and the RNA polymerase II second largest subunit (RPB2) genes were amplified and sequenced as previously described using the following primer pairs: BMBC-R and NL4R (ITS), Bt2a and Bt2b (BenA), CF1 and CF4 (CaM), and 5F and 7CR (RPB2) [20]. Bidirectional sequencing was performed using a capillary electrophoresis system (3730 × L Genetic analyzer or 3500 × L Genetic Analyzer, Life Technologies, Carlsbad, CA, USA). The obtained sequences were manually inspected, contiguous sequences were generated, and sequence alignments were produced using the Muscle program implemented in SEQUENCHER v. 5.4.6., build 46289 (Gene Codes Corp. Ann, Arbor MI, USA) [21]. The IQ-TREE software was used to construct phylogenetic trees using maximum likelihood, and the program ModelFinder was used to find the best-fitting substitution model for each gene according to the Akaike Information Criterion (AIC). Programs to measure branch support (i.e., the approximate Bayes test (aBayes) and ultrafast bootstraps programs (BS)) were implemented [22][23][24][25][26]. The strain type of A. fumigatus (NRRL 163) was used as an outgroup. Single-locus analyses for ITS, LSU, BenA, CaM, and RPB2 as well as multi-locus were performed. FigTree ver 1.4.4 (http://github.com/rambaut/figtree/, accessed on 11 November 2020) was used to visualize trees and edited using Canvas Draw 3 for MAC (v. 3.0.5, build 274). To increase the robustness of the ML topologies observed in the datasets, values of aBayes implemented in IQTREE software v2.2.0 were also shown together with the BS values [24,25].

In Vitro Susceptibility
Antifungal susceptibility was performed via the Clinical and Laboratory Standards Institution (CLSI) broth microdilution methods for amphotericin B, isavuconazole, itraconazole, posaconazole, voriconazole, and micafungin [27]. Stock solutions at a 100X concentration of each antifungal were prepared in DMSO, and further dilutions were made in RPMI buffered with 0.165 M MOPS (pH 7.0), 0.2% glucose, and phenol red and without bicarbonate. Final testing concentrations ranged from 0.03 to 16 mg/L for amphotericin B and the triazoles and from 0.015 to 8 mg/L for micafungin. For amphotericin B and the triazoles, minimum inhibitory concentrations (MIC) were read at 100% growth inhibition after 48 h of incubation at 35 • C, while the minimum effective concentration (MEC) for micafungin, defined as the lowest concentration that resulted in morphologic changes (e.g., short, stubby, abnormally branched hyphae), was read after 24 h of incubation at the same temperature. Hamigera insecticola ATCC MYA-3630 (previously identified as Paecilomyces variotii) was included as a quality control isolate on each day of testing.

Data Analysis
Species distributions and sites of culture were described using descriptive statistics. MIC/MEC ranges, MIC/MEC 50 and MIC/MEC 90 values (i.e., MIC/MEC values that inhibited 50% and 90% of isolates, respectively), modal MIC/MEC values, and geometric mean (GM) MIC/MEC values were determined for each antifungal. When the MIC/MEC values measured were greater than the highest concentration tested, the next highest 2X concentration was assigned for statistical comparisons. Differences in GM MIC values were assessed for significance by ANOVA with Tukey's post hoc test for multiple comparisons and correlations between isavuconazole, posaconazole, and voriconazole MICs were assessed by Pearson's correlation using log 2 -transformed MIC values.

Species Distribution and Sites of Culture
A total of 278 isolates within Aspergillus section Terrei from institutions across the U.S. were included in this analysis. Of these, A. terreus sensu stricto was the most prevalent Pathogens 2023, 12, 579 4 of 15 species, followed by A. hortae, A. alabamensis, and A. citrinoterreus ( Figure 1). The lower and upper respiratory tracts (46.8% and 22.8%, respectively) were the most prevalent sites from which the isolates were cultured (Supplemental Table S1). All other sites had less than 10% of each isolate cultured in them. Five isolates (1.8%; one A. alabamensis, three A. hortae, and one A. terreus) were cultured from the blood, and these may represent true infections, since members of this section are known to form aleurioconidia, or accessory conidia, that may form directly from vegetative hyphae and may be released into the bloodstream [12]. sessed by Pearson's correlation using log2-transformed MIC values.

Species Distribution and Sites of Culture
A total of 278 isolates within Aspergillus section Terrei from institutions across the U.S. were included in this analysis. Of these, A. terreus sensu stricto was the most prevalent species, followed by A. hortae, A. alabamensis, and A. citrinoterreus ( Figure 1). The lower and upper respiratory tracts (46.8% and 22.8%, respectively) were the most prevalent sites from which the isolates were cultured (Supplemental Table S1). All other sites had less than 10% of each isolate cultured in them. Five isolates (1.8%; one A. alabamensis, three A. hortae, and one A. terreus) were cultured from the blood, and these may represent true infections, since members of this section are known to form aleurioconidia, or accessory conidia, that may form directly from vegetative hyphae and may be released into the bloodstream [12].

Antifungal Susceptibility
The MIC and MEC results for each antifungal against all isolates tested are shown in Table 1. Against section Terrei as a whole, the most potent in vitro activity, as measured by GM MICs, was observed with posaconazole followed by itraconazole. The activities of isavuconazole and voriconazole were similar against these isolates, with both GM MICs, modal MICs, and MIC 50 and MIC 90 values being the same or within one dilution of each other. Not surprisingly, amphotericin B demonstrated less in vitro activity, as evident by higher GM MIC, modal MIC, and MIC 50 and MIC 90 values than the triazoles. This pattern of in vitro activity also held true for each of the individual species within section Terrei. Interestingly, the amphotericin B GM MIC and MIC 90 values were higher against A. citrinoterreus isolates (2.97 mg/L and 8 mg/L, respectively) than those of other species and were lowest against A. hortae isolates (0.862 mg/L and 2 mg/L, respectively). Overall, the amphotericin B and triazole MICs against A. neoafricanus and members of the section that were not able to be identified at the species level were similar to those observed for the  CLSI has not yet established breakpoints for any antifungal against members of section Terrei. However, epidemiologic cut-off values (ECVs) have been established for amphotericin B and the triazoles against A. terreus sensu stricto [28]. Overall, most isolates were categorized as wild-type since the majority MIC values were at or below the ECVs for posaconazole (ECV 1 mg/L; 100% wild-type), voriconazole (ECV 2 mg/L; 98.86% wild-type), itraconazole (ECV 2 mg/L; 100% wild-type), and amphotericin B (ECV 4 mg/L; 99.25% wild-type). However, the number of wild-type isolates was lower for isavuconazole (ECV 1 mg/L; 93.23% wild-type), although it was still greater than 90%.
Direct comparisons were also made between the triazoles isavuconazole, posaconazole, and voriconazole against isolates for which all three drugs were tested against at least ten strains ( Table 2). Posaconazole demonstrated significantly more potent in vitro activity, as assessed by GM MIC values (p-value < 0.001 for all comparisons), than that of both isavuconazole and voriconazole against Aspergillus section Terrei isolates as a whole and also against individual species. Lower modal MICs and MIC 50 and MIC 90 values were also observed with posaconazole. The correlations between isavuconazole and voriconazole, isavuconazole and posaconazole, and voriconazole and posaconazole MICs were also evaluated. The best correlation was observed between isavuconazole and voriconazole (Pearson's r-value 0.742) (Figure 2). In contrast, the correlations between posaconazole and either isavuconazole or voriconazole MICs were low (Pearson's r-values of 0.234 and 0.296, respectively). Micafungin also demonstrated potent activity, with MEC values at the lowest concentration tested (0.015 mg/L) for the vast majority of isolates tested. This was consistent regardless of species. However, since the endpoint for the echinocandins is different than that of the triazoles and amphotericin B, comparisons between this antifungal and the other drug classes were not made. and also against individual species. Lower modal MICs and MIC50 and MIC90 values were also observed with posaconazole. The correlations between isavuconazole and voriconazole, isavuconazole and posaconazole, and voriconazole and posaconazole MICs were also evaluated. The best correlation was observed between isavuconazole and voriconazole (Pearson's r-value 0.742) (Figure 2). In contrast, the correlations between posaconazole and either isavuconazole or voriconazole MICs were low (Pearson's r-values of 0.234 and 0.296, respectively).

Description of New Species
As previously described, five isolates underwent further phylogenetic analysis and phenotypic workup for the description of a new species, and each these isolates were cultured from four patients from three different institutions in Colorado, two in 2015 and one each in 2016, 2019, and 2020, respectively. All were presumptively identified as Aspergillus species belonging to the Aspergillus section Terrei based on phenotypic characteristics typical of species within this section. By BLASTn analysis of the BenA and CaM sequences, each isolate was found to be close to A. alabamensis, but were distant from other species within this section based on percent identity values (Supplemental Table S2). Subsequent phylogenetic analyses using additional regions (i.e., ITS and RPB2) were performed to study their relationship to species in section Terrei. Reference sequences (specifically from ex-types) of representative strains were obtained from GenBank and were included in the phylogenetic analyses [29][30][31][32]. Initially, phylogenetic analyses were performed on the datasets of individual genes (ITS, BenA, CaM, RPB2) to apply the genealogical concordance phylogenetic species recognition (GCPSR) concept [33]. The single-gene phylogenies showed that although the isolates were most phylogenetically close to A. alabamensis, they form a distinct lineage, which was confirmed by the phylogeny of the combined genes (Supplemental Figure S1 and Figure 3, respectively). The existing strains of A. alabamensis are diverse phylogenetically [30]. To show that our isolates are distinct, a separate CaM phylogenetic analysis was carried out, which included 24 additional strains of A. alabamensis. The resulting CaM phylogeny showed that our strains still constitute a distinct lineage. Here, we introduce a new species, A. pseudoalabamensis. The morphology of the new species is compared to A. alabamensis, and distinguishing characteristics are provided below.
Notes: A. pseudoalabamensis is phylogenetically closely related to A. alabamensis; however, it can be differentiated from this species by BenA, CaM, and RPB2 sequence data. The conidia of A. pseudoalabamensis are pale yellow to greyish to brownish orange en masse, while those of A. alabamensis are bright yellowish brown to cinnamon-brown [29]. Furthermore, as demonstrated by light microscopy and SEM, the conidiophores of A. pseudoalabamensis are shorter (150-325 × 2.5 µm) and finely roughened than those of A. alabamensis, which are longer (150-500 µm × 4.5-6 µm) and smooth ( Figure 4J-P) [29].

Discussion
Aspergillus terreus is one of the four main species causing infections by the genus Aspergillus, although the number of infections caused by this species is relatively small compared to A. fumigatus [5][6][7]. However, invasive and chronic infections secondary to A. terreus have fewer treatment options due to poor response rates observed in patients treated with amphotericin B formulations in various studies [8,9]. In a retrospective cohort study of 83 cases of proven or probable invasive aspergillosis due to A. terreus, Steinbach et al. reported that patients who received primary therapy with voriconazole had a higher overall survival 64.7%) versus those who received an amphotericin B formulation as primary therapy (26.2%; p = 0.01) [8]. This difference in survival rate was even greater for voriconazole compared to patients who only received amphotericin B for the duration of treatment (16.7%; p < 0.01). Similarly, Hachem et al. conducted a retrospective chart review of patients between 1995 and 2001 at MD Anderson Cancer Center in Houston, Texas, whose cultures were positive for Aspergillus species [9]. In all, 32 patients with invasive aspergillosis due to A. terreus and 33 patients in whom the disease was caused by A. fumigatus were evaluated. Interestingly, response rates to therapy were low in both groups, with only 39% responding in the A. fumigatus group and 28% in the A. terreus group. In those who received treatment with amphotericin B, response rates were 29% and 16%, respectively. Susceptibility data were available for 30 isolates (16 A. fumigatus and 14 A. terreus), with amphotericin B showing reduced in vitro activity against A. terreus (MIC 90 4 mg/L) compared to A. fumigatus (1 mg/L).
The poor clinical outcomes associated with the amphotericin B treatment of A. terreus infections are often attributed to microbiologic resistance. Indeed, some studies have reported elevated amphotericin B MICs against A. terreus isolates. In an international, prospective multicenter survey of the prevalence of A. terreus species complex infections and amphotericin B susceptibility, Risslegger et al. reported amphotericin B modal and MIC 90 values of 8 mg/L and 16 mg/L, respectively, against 370 isolates [13]. Based on their calculated amphotericin B ECV of 4 mg/L, just over half (52.4%) were classified as non-wild-types. Similar results were observed in a multicenter study conducted in France where the amphotericin B MIC 90 was 8 mg/L [34]. However, not all in vitro studies have reported amphotericin B MICs as high as these levels. Several studies have noted greater amphotericin B in vitro susceptibility against A. terreus isolates, as reflected by MIC 90 values of 2 mg/L [5,[35][36][37]. The results of these later studies are consistent with our findings, as the MIC 90 for amphotericin B against all members of section Terrei was 4 mg/L. However, we also saw variability with amphotericin B MIC results between the different species of this section, with the highest MIC 90 and GM MIC against A. citrinoterreus. This is consistent with the results of Imbert et al., who also reported higher amphotericin B MIC 90 and GM MIC values against this species [34]. Risslegger et al. reported higher amphotericin B MICs against all cryptic species compared to A. terreus sensu stricto, although this is not consistent with our results. Because of the variability in amphotericin B MICs and the lack of correlation between low values and positive clinical outcomes, CLSI now recommends against antifungal susceptibility testing of this polyene for clinical diagnostic purposes [38].
Our azole susceptibility results are in line with those reported by others, with posaconazole having the most potent in vitro activity and the MICs of voriconazole and isavuconazole closely mirroring each other, as reflected by Pearson's r-value of 0.742. This is consistent with what our group and others have reported for voriconazole and isavuconazole against A. fumigatus [39,40]. As previously noted, CLSI has not yet established breakpoints for any antifungal against A. terreus, although ECVs are available. These ECVs were established, in part, based on MIC data provided by multiple laboratories from several different countries [41][42][43]. The MIC ranges, modal MICs, and percent of isolates at or below the ECV from these studies (amphotericin B 0.12-32 mg/L, 2 mg/L, 94.7%; isavuconazole 0.06-2 mg/L, 0.25 mg/L, 99.7%; itraconazole 0.03-1 mg/L, 0.25 mg/L, 100%; posaconazole ≤ 0.03-2 mg/L, 0.25 mg/L, 99.7%; voriconazole 0.03-≥ 4 mg/L, 0.5 mg/L, 99.1%) are very similar to the results presented here, with the exception of the isavuconazole MIC range, which increased to 8 mg/L in the current study, and the percent of isolates at or below the isavuconazole ECV, which was 94.7% here. The reasons for these differences are unknown but may warrant the continued monitoring of isavuconazole MICs against A. terreus isolates over time.
Aspergillus section Terrei is now recognized as consisting of at least 14 cryptic species in addition to A. terreus sensu stricto [14,34,44], although less than half of these have been implicated in human disease [13,29,34,45,46]. In the current study, the majority of isolates were identified as A. terreus sensu stricto (nearly 70%), followed by the cryptic species A. hortae, A. alabamensis, and A. citrinoterreus. Similar species distributions have been reported in other large surveillance studies, where A. terreus sensu stricto was the predominant species cultured from humans [13,34]. As with other reports, the majority of isolates in our study were cultured from either the lower or upper respiratory tract.
The classification of Aspergillus species currently follows the polyphasic approach, wherein morphology, physiology, and molecular data are combined [18,47,48]. The isolates studied were obtained from the same locality (i.e., the state of Colorado, USA), from four different patients and at different times. These isolates were found to be conspecific and formed a distinct lineage within the Aspergillus section Terrei. Although there are various characteristics that can be used for the identification of section Terrei species, such as colony characteristics on various media and micro-morphologies, for example, the length of conidiophores, the color of conidial heads, and the production of accessory or secondary conidia, to name a few, making a distinction between species is still challenging as these features can change due to environmental conditions. Molecular data are advantageous in delimiting species in this section because these do not alter with changes in environmental conditions, unlike morphological characteristics. Based on morphological characteristics and the phylogenetic concordance of the various loci used, these isolates represent a new species named A. pseudoalabamensis, which is presented here. Utilizing the approach of congruent topologies derived from single genes has been found to be a good tool in species delimitation in the genus Aspergillus [29,30,[47][48][49]. ITS is an accepted barcode for fungi, but in this study, ITS was not reliable in distinguishing the species in the section, and BenA, CaM, and RPB2 are often used instead. Geographical distribution is not always known for many fungal species. This report, however, suggests that A. pseudoalabamensis is possibly endemic in Colorado. The knowledge of endemic mycosis will help in the research for disease prevention and management.
Although these results add to the literature on antifungal susceptibility and the species distribution of Aspergillus section Terrei throughout the U.S., not least due to the inclusion of the description of a new species, A. pseudoalabamensis, from a specific geographic area, this study is not without limitations. No clinical data were available to determine whether worse outcomes were associated with different species, sites of infection, or susceptibility profiles. In addition, we are unable to determine whether different cryptic species were associated with specific regions, as many of these isolates were received from other commercial or reference laboratories; hence, their geographic origin is unknown. Therefore, further studies are warranted.
In conclusion, our results demonstrate species diversity within the clinical isolates of Aspergillus section Terrei in the U.S., although the predominant species remains A. terreus sensu stricto. The azoles posaconazole, itraconazole, isavuconazole, and voriconazole have potent in vitro activity against members of this section as a whole and against the individual species. Amphotericin B demonstrated reduced susceptibility compared to the azoles, but this also varied between different species. Overall, these results from the U.S. are consistent with those of previous surveillance studies from other countries.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pathogens12040579/s1, Figure S1: maximum likelihood trees showing phylogenies of individual loci; Table S1: site of cultures of clinical isolates; Table S2: percent matches of A. pseudoalabamensis type strain to other types within section Terrei. Funding: Funding for this study and isavuconazole powder were obtained from Astellas Pharma, Inc. However, the research and publication were performed independent of this sponsor.
Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data will be made available to qualified researchers upon appropriate request.