Identification and in vitro antifungal susceptibility of causative agents of onychomycosis due to Aspergillus species in Mashhad, Iran

Aspergillus species are emerging causative agents of non-dermatophyte mold onychomycosis. In this study, 48 Aspergillus isolates were obtained from patients with onychomycosis in Mashhad, Iran, during 2015–2018. The aim is to identify the Aspergillus isolates to the species level by using partial calmodulin and beta-tubulin gene sequencing and MALDI-TOF MS, and to evaluate their in vitro susceptibility to ten antifungal drugs: terbinafine, itraconazole, voriconazole, posaconazole, ravuconazole, isavuconazole, caspofungin, micafungin, anidulafungin and amphotericin B according to CLSI M38-A3. Our results indicate that A.flavus (n = 38, 79%) is the most common Aspergillus species causing onychomycosis in Mashhad, Iran. Other detected species were A. terreus (n = 3), A. tubingensis (n = 2), A. niger (n = 1), A. welwitschiae (n = 1), A. minisclerotigenes (n = 1), A. citrinoterreus (n = 1) and A. ochraceus (n = 1). Aspergillus flavus, A. terreus and A. niger isolates were correctly identified at the species level by MALDI-TOF MS, while all cryptic species were misidentified. In conclusion, A. flavus is the predominant Aspergillus species causing onychomycosis due to Aspergillus spp. in Mashhad, Iran. MALDI-TOF MS holds promise as a fast and accurate identification tool, particularly for common Aspergillus species. It is important that the current database of reference spectra, representing different Aspergillus species is expanded to increase the precision of the species-level identification. Terbinafine, posaconazole and echinocandins were in vitro most active against the studies Aspergillus isolates and terbinafine could be the first choice for treatment of onychomycosis due to Aspergillus.

Scientific Reports | (2021) 11:6808 | https://doi.org/10.1038/s41598-021-86038-z www.nature.com/scientificreports/ morphology is difficult or even impossible, whereas a correct identification to species level is critical for optimal antifungal treatment because Aspergillus species can have variable drug susceptibility 8 . Partial calmodulin (CaM) and beta-tubulin (BenA) gene sequences are currently used for species identification within a given Aspergillus species complex 9 . Recent studies have demonstrated that matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) could be an useful alternative for sequence-based identification of Aspergilli 10,11 , particularly for phenotypically similar or indistinguishable Aspergillus species 11,12 . In this study, partial CaM and BenA gene sequences in combination with MALDI-TOF MS analysis are used to identify 48 Aspergillus isolates obtained from onychomycosis patients in two hospitals of Mashhad, Iran. Furthermore, antifungal susceptibility testing on those isolates to 10 antifungal drugs including terbinafine is performed using CLSI M38-A3.

Materials and methods
Isolation. A total 48 clinical Aspergillus isolates were collected from patients diagnosed with onychomycosis during the period from 2015 to 2018 at Emam Reza and Ghaem hospitals in Mashad, Iran. Diagnosis of Aspergillus onychomycosis was made based on dystrophic nail appearance and mycological criteria following literature 4 : (1) positive direct microscopy with hyphae presented in nail specimen; (2) positive culture of NDM; (3) repeated culturing of the NDM with absence of dermatophytes and yeasts at different time and samples; (4) molecular identification of the Aspergillus isolates.
All specimens were collected using a scalpel and direct microscopic examination was performed using a 20% KOH solution. Sabouraud Glucose Agar (SGA) (Difco, Detroit, MI, USA) containing chloramphenicol (50 mg/L) and cycloheximide (400 mg/L) was used for the isolation of dermatophytes and SGA with chloramphenicol (50 mg/L) for molds. All isolation plates were incubated at 27 °C for up to 4 weeks.
Sequenced-based and MALDI-TOF MS identification. The isolates were identified by CaM and BenA gene sequencing 9,13 . The obtained sequences were compared to the NCBI nucleotide database (BLAST; http:// blast. ncbi. nlm. nih. gov/ Blast. cgi) and the internal sequence database of the Westerdijk Fungal Biodiversity Institute, Utrecht, the Netherlands, containing verified CaM and BenA gene sequences of all accepted Aspergillus species 14 .
MALDI-TOF MS was performed by the formic acid extraction method according to the manufacturer's instruction (AUTOF MS1000, Autobio, China) and a previous publication with minor modification 15 . All used chemical reagents were of LC-MS grade. Briefly, each Aspergillus isolate was cultured on SGA for 3-5 days at 35 °C. After growth, an appropriate amount of sample was collected in a 1.5 mL centrifuge tube containing 1.0 mL 75% ethanol (Sigma-Aldrich, St.Louis, MO, USA). After mixing, the sample was centrifuged at 15,000 × g for 5 min and the supernatant discarded. After drying of the residue at 37 °C, 40 µL of lysis solution 1 (containing formic acid) was added and air-dried at room temperature. Subsequently, 1µL was transferred on a target plate, dried naturally in a bio-safety cabinet, and afterwards 1µL of matrix solution was added and dried again. Eventually, for each isolate a mass spectrum was generated and integrated to give a sum spectrum using AUTOF MS1000 (Autobio, China) with in house database.
Species identification by CaM and BenA gene sequencing and MALDI-TOF MS of 48 Aspergillus isolates are shown in Table 1. GenBank accession numbers for BenA and CaM gene sequences generated in this study are listed in Suppl. Table 1.
Antifungal susceptibility testing and statistical analysis. All isolates were tested according to Clinical and Laboratory Standards Institute (CLSI) M38-A3 document 16 . The ten antifungal agents included in this study are itraconazole, voriconazole, posaconazole, ravuconazole, isavuconazole, caspofungin, anidulafungin, amphotericin B (Sigma, Poole, United Kingdom), terbinafine (Aladdin, California, United States) and micafungin (Toronto-Research-Chemicals, Toronto, Ontario, Canada). The antifungal agents were tested at concentrations ranging from 0.008-4 mg/L for echinocandins (micafungin, anidulafungin and caspofungin) and All methods were performed in accordance with the relevant guidelines and regulations as references were given.
Ethics approval. Ethical approval to conduct the study was obtained from the Ethics Committee of Mashhad University of Medical Sciences (IR.MUMS.MEDICAL.REC.1397.660) and all patients involved understood and agreed to the use of these clinical specimens in the present study.
An 89.5% concordance between MALDI-TOF MS and molecular identification was found MALDI-TOF MS allowed the identification of the four common non-cryptic species, but failed to correctly identify the four cryptic species.
Antifungal susceptibility. The ranges MIC/MEC, Geometric mean MIC/MEC, Modal MIC/MEC, distribution of MICs/MECs of ten antifungal agents against 48 Aspergillus isolates are presented in Table 2.
In general, among 48 Aspergillus isolates tested in this study, the lowest Modal MIC/MEC (< 0.008 mg/L) were those of anidulafungin and micafungin, followed by terbinafine and posaconazole For section Terrei, three A. terreus isolates had a similar antifungal susceptibility profile as A. flavus, except terbinafine, which has higher MICs (Modal MIC 0.25 mg/L, n = 3). In contrast to A. terreus, the isolate of the cryptic species A. citrinoterreus had a MIC of 4 mg/L to amphotericin B.
All section Nigri isolates revealed lower MIC and Modal MIC ≤ 1 mg/L to the five azoles tested. The majority isolates of section Nigri have an MIC value of 0.031 mg/L to terbinafine except one A. tubingensis isolate with an MIC value of 0.5 mg/L.

Discussion
This study presented, for the first time, an overview of the occurrence of Aspergillus species causing onychomycosis in Mashhad, Iran, including a molecular characterization of the isolates and in vitro susceptibility to 10 antifungal agents. In total 48 Aspergillus isolates were obtained from nail in patients with onychomycosis in two medical centers of Mashhad, Iran.
Our  33 reported the first case of fungal keratitis due to this species in Iran in 2019, suggesting that the occurrence of A. minisclerotigenes might have been underreported due to identification based on morphology. The recently described A. citrinoterreus within section Terrei 34 has been reported as causative agent involved inhuman invasive aspergillosis 35 38,39 . So far, only two cases reported A. ochraceus as a pathogenone case was an invasive pulmonary aspergillosis from Poland 38 , another case was allergic bronchopulmonary aspergillosis 39 . In our study, 38 isolates were molecularly identified as A. flavus or A. oryzae. These two species are genetically very similar and indistinguishable by BenA and CaM gene sequencing 40 . Identification of these species is based on the origin and/or toxin production potential: A. oryzae is used in (food) fermentations and biotechnology and does not produce aflatoxins, while A. flavus is not domesticated and can produce aflatoxins, though nontoxigenic strains also occur 41 .Of note, all 38 A. flavus isolates were corrected identification by MALDI-TOF MS analysis based on the AUTOF MS 1000 (Autobio, China) in-house database. Similarly, De Carolis et al. 40 reported that MALDI-TOF MS easily differentiated A. flavus and A.oryzae on the species level. In contrast to some investigators 42,43 , who reported that only 18 of 200 isolates were confirmed as A. flavus using MALDI-TOF MS based on the Bruker score database among a set of 200 clinical and environmental A. flavus isolates identified by sequencing. Masih et al. 43 also showed that A. oryzae can easily be identified as A. flavus. Those discrepancy could be due to differences between MALDI-TOF MS machines, the species coverage in reference databases, interpretive cutoffs and methodology applied for sample preparation 11 .
In our study, A. minisclerotigenes was misidentified as A. flavus by MALDI-TOF MS. Similarly, A. tubingensis and A. welwitschiae were misidentified as A. niger and A. citrinoterreus was misidentified as A. terreus. These are due to the inadequate number of reference spectra in the used in-house database (Suppl. Table 2). Therefore, databases should be expanded with well-validated spectra of cryptic species in order to get an accurate identification of all (clinical relevant) Aspergillus species. Our study confirms the discriminatory power of MALDI-TOF MS for common clinical Aspergillus species 40 .
The antifungal susceptibility data generated in this study indicate that terbinafine has excellent in vitro activity against the eight species tested. Terbinafine appears more active against isolates of section Flavi (GM 0.032 mg/L   www.nature.com/scientificreports/ and section Nigri (GM 0.077 mg/L) than to section Terrei isolates (GM 0.25 mg/L), and has superior activity against A. flavus compared to the azoles and amphotericin B. Similar results are obtained in previous studies 44,45 . Thus, terbinafine could be placed as an alternative drug for treatment of onychomycosis caused by Aspergillus species with the confirmation of clinical efficacy of terbinafine in the management of onychomycosis due to Aspergillus spp. 4 . Based on the proposed epidemiologic cut-off values (ECV) of A. flavus (posaconazole 0.5 mg/L; itraconazole 1 mg/L; voriconazole 1 mg/L; isavuconazole1 mg/L and amphotericin B 4 mg/L) 46 , all azoles tested in our study exhibited good activity to 38 A. flavus isolates and this is in agreement with previous reports 18,19 . Posaconazole had the lowest GM value (0.097 mg/L), followed by voriconazole and isavuconazole (both 0.274 mg/L), itraconazole and ravuconazole (0.311 and 0.394 mg/L, respectively) in an increasing order. Although no antifungal susceptibility ECV are available for A. flavus to ravuconazole, in our study, all A. flavus isolates had MIC values less than 1 mg/L (MIC range 0.25-0.5 mg/L), indicating the potency of this antifungal against A. flavus, which is in agreement with those reported by Pfaller et al. 47  Among the section Nigri isolates, the azoles tested were active against all isolates, although an A. tubingensis isolate presented a slightly higher MIC value for azoles, which is similar to results reported previously 52 .
The three A. terreus and one A. citrinoterreus isolate displayed low MICs for the tested azoles. Posaconazole was the most effective azole against A. terreus (modal MIC 0.063 mg/L, n = 2) which is similar to previous studies 53,54 .
Although A terreus is intrinsically resistant to amphotericin B, 12-13% of the isolates have low amphotericin B MICs 55,56 . Our results also show that the three A. terreus isolates exhibited MICs < 2 mg/L, which is below the proposed ECV (MIC 4 mg/L). However, the A. citrinoterreus isolate had a MIC of 4 mg/L for amphotericin B, which is in agreement with those reported by Imbert et al. 35 .
Echinocandins showed good activity against the most Aspergillus isolates in this collection, while anidulafungin and micafungin showed a lowest Modal MEC of 0.008 mg/L, followed by caspofungin with Modal MEC of 0.125-0.5 mg/L. Anidulafungin and micafungin appeared more potent than caspofungin and this is consistent with several previous studies 44,57 . Of note, the four section Terrei isolate have higher MEC values (0.5 mg/L), above ECV (0.125 mg/L). The similar results were reported by Lass-Flörl et al. With a set of 48 clinical and 31 environmental A. terreus isolates, they showed that caspofungin has higher MECs (MEC 90 2 mg/L) than anidulafungin (MEC 90 0.03 mg/L) and micafungin (MEC 90 0.02 mg/L) 58 .

Conclusion
Based on molecular and MALDI-TOF MS identification, A. flavus (79%) is the most common Aspergillus species in NDM onychomycosis due to Aspergillus in Mashhad, Iran. The other isolates showed a wider species diversity. We emphasize the importance of using molecular methods to accurately identify Aspergillus at the species level because different species may vary in terms of susceptibility to antifungal agents. However, our results are limited by the relatively low number of clinical Aspergillus isolates obtained in Mashhad, Iran. Terbinafine, posaconazole, and echinocandins are shown in vitro to be the most potent antifungal agents against Aspergillus spp. Terbinafine could be first line drug for treatment of onychomycosis due to Aspergillus, the in vivo efficacy remains to be determined.

Data availability
The samples utilized in our study were obtained from preexisting samples isolated from patients who routinely referred to the hospital for direct examination and culture in Clinical Lab. And we used the culture for this research, and no additional samples were taken from the patient.