Genetic Diversity and In Vitro Antifungal Susceptibility of 200 Clinical and Environmental Aspergillus flavus Isolates

ABSTRACT

Aspergillus flavus has been frequently reported as the leading cause of invasive aspergillosis in certain tropical and subtropical countries. Two hundred A. flavus strains originating from clinical and environmental sources and collected between 2008 and 2015 were phylogenetically identified at the species level by analyzing partial β-tubulin and calmodulin genes. In vitro antifungal susceptibility testing was performed against antifungals using the European Committee on Antimicrobial Susceptibility Testing (EUCAST) broth microdilution method. In addition, genotyping was performed using a short-tandem-repeat (STR) assay of a panel of six microsatellite markers (A. flavus 2A, 2B, 2C, 3A, 3B, and 3C), in order to determine the genetic variation and the potential relationship between clinical and environmental isolates. The geometric means of the minimum inhibitory concentrations/minimum effective concentrations (MICs/MECs) of the antifungals across all isolates were (in increasing order): posaconazole, 0.13 mg/liter; anidulafungin, 0.16 mg/liter; itraconazole, 0.29 mg/liter; caspofungin, 0.42 mg/liter; voriconazole, 0.64 mg/liter; isavuconazole, 1.10 mg/liter; amphotericin B, 3.35 mg/liter; and flucytosine, 62.97 mg/liter. All of the clinical isolates were genetically different. However, an identical microsatellite genotype was found between a clinical isolate and two environmental strains. In conclusion, posaconazole and anidulafungin showed the greatest in vitro activity among systemic azoles and echinocandins, respectively. However, the majority of the A. flavus isolates showed reduced susceptibility to amphotericin B. Antifungal susceptibility of A. flavus was not linked with the clinical or environmental source of isolation. Microsatellite genotyping may suggest an association between clinical and environmental strains, although this requires further investigation.

INTRODUCTION

Invasive aspergillosis is an important opportunistic fungal infection, particularly in immunocompromised patients (1–3). After Aspergillus fumigatus, Aspergillus flavus (section Flavi) is the second most common opportunistic pathogen causing invasive and superficial infections in humans (4). The fungus is predominantly associated with infections of the respiratory tract, brain, sinuses, eye, and skin, particularly in hot and arid climates, such as North Africa, the Middle East, and India (4, 5). A. flavus is also known to produce aflatoxin in crops, resulting in significant economic damage globally (6, 7). Furthermore, experimental data suggest that A. flavus is significantly more virulent than A. fumigatus (8).

Irrespective of the causative agent, without adequate therapy, invasive aspergillosis is associated with high mortality rates (9). Triazoles are the preferred agents for treatment and prevention of infections caused by Aspergillus species in most patients (10). However, with the widespread use of azoles in clinical practice and agriculture, the emergence of triazole resistance in A. fumigatus has become a global public health concern. Moreover, mutations in A. flavus cyp51A (T788G) and cyp51C (Y319H), both associated with increased MICs to voriconazole, have also been reported recently (11).

Given the prominent role of azoles in the management of aspergillosis (12), it is important to determine the molecular epidemiology and resistance profiles of A. flavus isolates associated with Aspergillus disease (13). In the current study, we evaluated the genetic diversity and in vitro antifungal susceptibilities of a large collection of A. flavus strains obtained from clinical and environmental sources between 2008 and 2015, using PCR sequencing, microsatellite genotyping, and antifungal drug susceptibility profiles using the European Committee on Antimicrobial Susceptibility Testing (EUCAST) broth microdilution method.

(Parts of these results were presented at the 26th European Congress of Clinical Microbiology and Infectious Diseases, 9 to 12 April 2016, Amsterdam, the Netherlands, poster no. 0378 [14].)

RESULTS

All isolates were identified as A. flavus by sequence analysis of β-tubulin and calmodulin genes.

Microsatellite genotyping of six short-tandem-repeat (STR) loci showed that all of the clinical isolates were genetically different from each other. However, identical STR genotypes were found between a clinical isolate and two hospital environmental strains. Of note, the patient with the clinical isolate stayed in the same hospital where the identical environmental strains were retrieved, which indicates similar phylogenetic origin and the possibility of the same source. Comparing the genetic relatedness by generating dendrograms of the STR profiles showed that all 200 A. flavus isolates clustered apart and were distinct from 10 A. flavus control strains isolated from patients in the Netherlands and Turkey.

The geometric mean (GM) MICs/MECs, the MIC/MEC ranges, the MIC₅₀/MEC₅₀, and MIC₉₀/MEC₉₀ distributions of the eight antifungals against 200 A. flavus strains are listed in Table 1. The MIC/MEC distributions of all tested antifungals are presented in Fig. 1. The overall results obtained from visual and spectrophotometric readings were not different for the MIC endpoints, and similar results were obtained in replicate experiments.

TABLE 1.

Geometric mean MIC/MEC, MIC/MEC ranges, MIC₅₀/MEC₅₀, and MIC₉₀/MEC₉₀ values of 200 clinical and environmental Aspergillus flavus strains to eight antifungal agents

Origin of isolation (no. of strains)	Druga	MIC/MEC (mg/liter)b
		Range	MIC50/MEC50	MIC90/MEC90	Geometric mean
All strains (200)	AmB	1–16	4	4	3.35
	ITC	0.031–4	0.25	0.5	0.29
	VRC	0.063–2	0.5	1	0.64
	POS	0.031–1	0.125	0.25	0.13
	ISA	0.125–4	1	2	1.10
	CAS	0.125–2	0.25	1	0.42
	AFG	0.016–1	0.031	0.25	0.16
	5-FC	4–64	64	64	62.97
Clinical (121)	AmB	1–8	4	4	3.40
	ITC	0.031–4	0.25	0.5	0.31
	VRC	0.25–2	0.5	1	0.69
	POS	0.031–1	0.125	0.25	0.13
	ISA	0.5–4	1	2	1.16
	CAS	0.125–2	0.25	0.5	0.36
	AFG	0.016–1	0.031	0.25	0.16
	5-FC	8–64	64	64	63.54
Environmental (79)	AmB	1–16	4	4	3.26
	ITC	0.031–2	0.25	0.5	0.25
	VRC	0.063–2	0.5	1	0.55
	POS	0.031–0.5	0.125	0.25	0.13
	ISA	0.125–4	1	2	1
	CAS	0.125–1	0.5	1	0.52
	AFG	0.016–0.5	0.063	0.5	0.16
	5-FC	4–64	64	64	62.08

^aAmB, amphotericin B; ITC, itraconazole; VRC, voriconazole; POS, posaconazole; ISA, isavuconazole; CAS, caspofungin; AFG, anidulafungin; 5-FC, flucytosine.

^bMECs were used for caspofungin and anidulafungin. MICs were used for all other drugs.

FIG 1.

MIC/MEC distributions for 200 Aspergillus flavus isolates from clinical and environmental origins. The x axis shows the MICs/MECs (in milligrams per liter), and the y axis shows the number of strains in the set with the given MIC/MEC. Arrow bars indicate epidemiological cutoff values proposed by EUCAST.

The geometric means of the MICs/MECs of the antifungals across all isolates were the following (in increasing order): posaconazole, 0.13 mg/liter; anidulafungin, 0.16 mg/liter; itraconazole, 0.29 mg/liter; caspofungin, 0.42 mg/liter; voriconazole, 0.64 mg/liter; isavuconazole, 1.10 mg/liter; amphotericin B, 3.35 mg/liter; and flucytosine, 62.97 mg/liter.

Overall, posaconazole and anidulafungin showed the greatest in vitro activity among antifungal azoles and echinocandins, respectively, independent of the source of isolation. The highest MIC₉₀ value was 64 mg/liter for flucytosine, which was significantly different from those of the other seven antifungal agents (P < 0.01). Among azoles, the activity of posaconazole (GM MIC, 0.13 mg/liter; MIC range, 0.031 to 1 mg/liter) was at a level similar to that of itraconazole (GM MIC, 0.29 mg/liter; MIC range, 0.031 to 4 mg/liter), and this was followed by the activity of voriconazole (GM MIC, 0.64 mg/liter; MIC range, 0.63 to 2 mg/liter) and isavuconazole (GM MIC, 1.10 mg/liter; MIC range, 0.125 to 4 mg/liter).

No statistically significant differences in the susceptibility profiles of A. flavus were detected between the clinical and environmental sources investigated (Table 1).

DISCUSSION

This study investigated the genetic diversity and antifungal susceptibility profiles of a large collection of clinical and environmental strains of A. flavus. Importantly, the antifungal susceptibility of A. flavus was not linked with the clinical or environmental source of isolation. For all tested strains, antifungal triazoles and echinocandins had low MICs values, whereas flucytosine did not show inhibitory effects (GM MIC, 62.97 mg/liter). The efficacy of flucytosine against Aspergillus infection has long been controversial. Flucytosine was found to be active in vitro against Aspergillus isolates when the MIC was determined at pH 5.0 instead of pH 7.0 (15). In addition, the in vitro MIC at pH 5.0 corresponded to the in vivo efficacy of flucytosine monotherapy in a murine model of invasive aspergillosis (16, 17).

In our study, consistent with several previous reports (18–20), all tested azoles showed potent in vitro activity against all A. flavus isolates. Similarly, Shivaparkash et al. studied the antifungal susceptibility profiles of triazoles against a large collection of clinical (n = 178) and environmental (n = 10) strains of A. flavus originating from India, using the Clinical and Laboratory Standards Institute (CLSI) method (21). Posaconazole had the highest activity (GM MIC, 0.123 mg/liter; MIC range, 0.062 to 0.25 mg/liter), followed by itraconazole (GM MIC, 0.177 mg/liter; MIC range, 0.062 to 0.5 mg/liter), isavuconazole (GM MIC, 0.697 mg/liter; MIC range, 0.125 to 2 mg/liter), and voriconazole (GM MIC, 1.167 mg/liter; MIC range, 0.05 to 4 mg/liter). It is worth mentioning that CLSI clinical breakpoints have not been established for Aspergillus species (22). However, epidemiological cutoff values are available for Aspergillus spp. versus the triazoles, caspofungin, and amphotericin B. The epidemiological cutoff values that captured 95% of the A. flavus wild-type population (organisms in a species-drug combination with no acquired resistance mechanisms) were as follows: posaconazole, 0.25 mg/liter; itraconazole, 1 mg/liter; voriconazole, 1 mg/liter (19, 23); isavuconazole, 1 mg/liter (24); caspofungin, 0.25 mg/liter (25); and amphotericin B, 2 mg/liter (26).

The clinical breakpoint for A. flavus has only been proposed for itraconazole, with ≤1 mg/liter as susceptible and >2 mg/liter as a resistant phenotype, using EUCAST guidelines (27). For other systemic antifungals, there is, however, insufficient evidence to recommend a clinical breakpoint. In the absence of clinical breakpoints, epidemiological cutoff values, although they do not predict therapy outcome as clinical breakpoints do, may aid to distinguish wild-type isolates of Aspergillus spp. with no acquired resistance mechanisms detectable from those that may harbor resistance mutations (19, 28).

The EUCAST epidemiologic cutoff values for triazoles against A. flavus have been proposed according to MIC distributions of wild-type strains of A. flavus, as follows: itraconazole, 1 mg/liter; posaconazole, 0.5 mg/liter (27); voriconazole (29) and isavuconazole (30), 2 mg/liter; and amphotericin B, 4 mg/liter (27). As a comparison, the proposed EUCAST epidemiologic cutoff values for A. flavus were one step higher than those from CLSI.

As shown in Fig. 1, our results demonstrated that wild-type MIC distributions based on the proposed epidemiologic cutoff values of A. flavus were as follows: itraconazole, 98.5%; posaconazole, 99.5%; voriconazole, 100%; isavuconazole, 98%; and amphotericin B, 96%. Of note, the vast majority of A. flavus isolates had MIC values at or below epidemiologic cutoff values for triazoles and amphotericin B, including those from patients who had received these agents.

According to the recent guidelines of Infectious Diseases Society of America, the triazole voriconazole is still the recommended primary therapy for disseminated infections caused by Aspergillus species (10). However, acquired resistance is an emerging concern in A. fumigatus, which is now reported globally (31). Environmental exposure to azole fungicides has been found to be an important route for resistance selection (32), as several azole fungicides have a molecule structure similar to that of medical triazoles (33). Although the exact process of environmental resistance selection in A. fumigatus in not fully understood, exposure of the fungus to azole residues in its natural habitat is thought to be a critical factor. Although one can postulate that similar conditions, i.e., exposure to azole residues, may be present for the habitat of A. flavus, widespread resistance selection in A. flavus has not found. Acquired azole resistance in A. flavus is very uncommon and to date has been reported in only two patients receiving long-term azole therapy. Our result indeed indicates that unlike A. fumigatus, an environmental route of resistance selection is not a concern in A. flavus. Furthermore, in a recent study, using an experimental murine model of disseminated A. flavus infection, voriconazole monotherapy was shown to be effective against both a wild-type isolate without a mutation in the cyp51C gene and a voriconazole-resistant isolate harboring the Y319H substitution in the cyp51C gene (MIC EUCAST, ≥2 mg/liter) (34). The efficacy of voriconazole was correlated with voriconazole exposure in a dose-dependent manner and with the voriconazole MIC of the isolates. Interestingly, even lower exposure was required for A. flavus isolates with higher MICs (34).

We also observed that the MEC values of anidulafungin and caspofungin were relatively low, with MEC ranges of 0.016 to 1 mg/liter and 0.125 to 2 mg/liter, respectively. In agreement with our finding, Pfaller et al. also reported that all three echinocandins (caspofungin, micafungin, and anidulafungin) inhibited a collection of 64 A. flavus isolates (n = 64) at a concentration of ≤0.06 mg/liter using the CLSI method (35). In another study, the mean MIC₉₀ values for anidulafungin, micafungin, and caspofungin were 0.002 mg/liter, 0.002 mg/liter, and 0.032 mg/liter, respectively (36). Lockhart et al. also determined the echinocandin MEC values against 28 A. flavus isolated from patients with transplant-associated infections (37). The MEC ranges for all isolates were as follows: caspofungin and micafungin, 0.008 to 0.03 mg/liter; and anidulafungin, 0.008 to 0.015 mg/liter. Similarly, in our study, anidulafungin (MIC range, 0.016 to 1 mg/liter) appeared to be more potent than caspofungin (MEC range, 0.125 to 2 mg/liter).

Of note, amphotericin B showed relatively high MICs (GM MIC, 3.3.5 mg/liter; MIC range, of 1 to 16 mg/liter) against all A. flavus strains tested, which is in agreement with previous reports from the United States, Europe (38, 39), and the Middle East (36, 40), which reported amphotericin B MICs of >2 mg/liter for clinical A. flavus strains. The in vitro data were also supported by experimental models of systemic aspergillosis against clinical isolates of A. flavus, with various susceptibilities to the polyene (41, 42). The in vivo efficacy studies of amphotericin B demonstrated that a clinical isolate showing an MIC of ≥2 mg/liter may be reasonably considered resistant in vivo to any dose/formulation of amphotericin B (42).

In our study, all the clinical isolates were concluded to be unrelated, as they had unique genotypes. The isolates were collected from different patients, from separate regions of Iran (Table 1), and were collected over an 8-year period. Similarly, the high genetic diversity in A. flavus was observed in clinical strains originating from humans (21) and animals (43). Furthermore, in a previous study, we also showed that clinical and environmental A. fumigatus strains from Iran clustered apart from each other (44). In the present study, however, microsatellite genotypes were matched between a clinical isolate and two environmental strains obtained from the same hospital environments, which may suggest that both clinical and environmental isolates belonged to the same origin and had the possibility of hospital-acquired colonization or infection. However, this requires further investigation. Understanding the correlation between the genotype of the isolates and the clinical disease (45), which may vary from country to country, and therapeutic modalities, including the antifungal susceptibility patterns of causative agents against a panel of available systemic antifungal compounds (46), is a valuable asset for clinicians, clinical mycology/microbiology laboratories, and health care professionals and may guide personalized therapy.

In conclusion, our data showed that triazoles and echinocandins had potent in vitro activity against both clinical and environmental A. flavus isolates. In contrast, 96% of the strains showed reduced susceptibility to amphotericin B, with MICs of ≥2 mg/liter.

MATERIALS AND METHODS

Fungal isolates.

A collection of 200 A. flavus strains was investigated, including 121 clinical and 79 environmental samples, as shown in Table 2. The clinical isolates were collected from patients who resided in different parts of Iran and who were admitted to four university hospitals in Tehran, Sari, Hamedan, and Mashhad between 2008 and 2015. The 121 clinical isolates were obtained from patients with various underlying diseases, including hematological malignancies, lung cancer, breast cancer, organ transplantation, chronic obstructive pulmonary disease, sinusitis, pulmonary and respiratory disorders, tuberculosis with cavity, renal failure, chronic granulomatous disease, autoimmune diseases, diabetes mellitus, and dermatomycoses. All the patient-related data were processed anonymously, and the ethics committee waived informed consent.

TABLE 2.

Origin of 200 Aspergillus flavus isolates from clinical and environmental sources in Iran

Origin by isolate type	No. (%) of samples
Clinical (n = 121)
Bronchoalveolar lavage	69 (34.5)
Nail scraping	17 (8.5)
Sinus biopsy specimen	14 (7)
Skin lesions	10 (5)
Ear swabs	5 (2.5)
Nasal mucosa	4 (2)
Lung biopsy specimen	2 (1)
Environmental (n = 79)
Hospital environment	56 (28)
Hospital's outdoor soil	23 (11.5)
Total	200 (100)

The collection of 79 environmental isolates of A. flavus obtained from 23 soil samples of the hospitals (11.5%) and the environment of 56 hospitalization wards (28%) were analyzed. In addition, 10 clinical A. flavus strains originally isolated from patients in the Netherlands and Turkey were used as the control strains in all experiments. All isolates were stored in 10% glycerol broth at −80°C at the Invasive Fungi Research Center of Mazandaran University of Medical Sciences in Sari. The isolates were subcultured on Sabouraud dextrose agar (SDA) supplemented with 0.02% chloramphenicol for 5 days at 35 to 37°C.

Strain identification and PCR sequencing.

The isolates were originally identified by experienced technicians on the basis of macroscopic colony morphology, microscopic morphology of conidia, and conidium-forming structures, and further confirmed by sequence-based analysis of parts of the β-tubulin and calmodulin genes, as described previously (47). To identify isolates to the species level and further analyze their phylogenetic relationships, the partial DNA sequence data from both genes were used as the BLAST query against two Web-accessible databases, those of the Centraalbureau voor Schimmelcultures (CBS-KNAW) Fungal Biodiversity Center, Utrecht, the Netherlands (http://www.cbs.knaw.nl), and the National Center for Biotechnology Information, Bethesda, MD (http://www.ncbi.nlm.nih.gov).

DNA extraction.

DNA was isolated as described previously (48); in brief, the isolates were cultured on Sabouraud dextrose agar slants. Conidia were harvested and added to 200 μl of breaking buffer (100 mM NaCl, 10 mM Tris-HCl [pH 8], 2% Triton X-100, 1% sodium dodecyl sulfate, 1 mM EDTA [pH 8]) with ∼0.1-g glass beads (diameter, 0.4 to 0.6 mm). After shaking by vortexing, conidia were incubated at 70°C for 30 min while shaking. Then, 200 μl of phenol-chloroform-isoamyl alcohol (25:24:1) saturated with aqueous buffer (pH 8.0) was added, and samples were incubated for 5 min while they were shaken. After centrifugation for 5 min, the upper phase containing the DNA was transferred to a new tube. One microliter of DNA was used per PCR.

Microsatellite genotyping.

Genotyping was performed on all A. flavus isolates, using a short-tandem-repeat (STR) A. flavus assay, as described previously (13). Briefly, six loci consisting of three dinucleotide-repeat units and three trinucleotide-repeat units (A. flavus 2A, 2B, 2C, 3A, 3B, and 3C) were amplified by using fluorescently labeled primers (6-carboxyfluorescein [FAM], HEX, and 6-carboxytetramethylrhodamine [TAMRA]) (32, 49, 50). The sizes of the fragments were determined by addition of the GeneScan 500 LIZ marker and subsequent analysis of the fragments on the Applied Biosystems 3730 DNA analyzer. Assignment of repeat numbers in each marker was determined from the GeneScan data by using the Peak Scanner version 1.0 software (Applied Biosystems). The sizes of the fragments were determined based on the 500 LIZ marker, and the repeat numbers of these isolates were compared to a collection of 20 A. flavus control isolates. Allele-sharing distance matrices were generated from the tandem-repeat numbers and were used as input to the Neighbor program of the PHYLIP version 3.6 software package to produce dendrograms (51–53).

In vitro antifungal susceptibility testing.

In vitro antifungal susceptibility testing was performed using a broth microdilution method, according to the EUCAST guidelines (54). The final concentrations of the following antifungal agents ranged from 0.016 to 16 mg/liter: amphotericin B, itraconazole, voriconazole, posaconazole, isavuconazole, caspofungin, and anidulafungin. Flucytosine was assessed over a 2-fold concentration range, from 0.064 to 64 mg/liter. The MICs of amphotericin B, itraconazole, voriconazole, posaconazole, isavuconazole, and flucytosine were determined visually: an inverted mirror was used for comparing the growth in wells containing the drugs with that in the drug-free control well. The MIC results were also read using a microtitration plate spectrophotometric reader (Anthos HTIII; Anthos Labtec Instruments, Salzburg, Austria). The MIC was defined as the lowest antifungal concentration that inhibited growth by 100% after 48 h compared with that of the drug-free well. The minimum effective concentrations (MECs) of caspofungin and anidulafungin were read using a plate microscope (Olympus SZX9; Olympus Nederland, Zoeterwoude, the Netherlands), at ×25 to ×50 magnification. The MECs were defined as the lowest concentration at which atypical, short, and branched hyphal clusters were observed instead of the long unbranched hyphal elements observed in the control well. The ranges and geometric means (GM) of the MICs and MECs were determined for each species and drug after 24 and 48 h of incubation. The MIC/MEC at which 50% (MIC₅₀/MEC₅₀) and 90% (MIC₉₀/MEC₉₀) of the isolates were inhibited were calculated using the aforementioned MIC determination criteria. The MIC₅₀ and MIC₉₀ values were calculated for species present in 10 or more isolates from the same geographic regions. If the MIC values of the replicates were distinct, the GM values of the replicates were used for comparison with other isolates. Paecilomyces variotii (ATCC 22319), Candida parapsilosis (ATCC 22019), and Candida krusei (ATCC 6258) were used for quality controls in all experiments. All experiments on each strain were performed using three independent replicates on different days.

Statistical analysis.

Data were analyzed using GraphPad Prism, version 7.0, for Windows (GraphPad Software, San Diego, CA). Genotyping diversity and MIC/MEC distributions between the isolates originating from different locations were compared using Student's t test and the Mann-Whitney-Wilcoxon test; differences were considered statistically significant at a P value of ≤0.05 (two-tailed).