Azole Resistance and cyp51A Mutation of Aspergillus fumigatus in a Tertiary Referral Hospital in Taiwan

Azole resistance in Aspergillus fumigatus has increasingly been reported worldwide. Its major mechanism of resistance is mediated by mutations in cyp51A. The objective of this study was to test the antifungal susceptibilities of A. fumigatus isolates from Chang Gung Memorial Hospital (CGMH), the largest tertiary referral hospital in Taiwan, and to investigate cyp51A mutations in azole-resistant strains. A. fumigatus isolates preserved in the Research Laboratory of Medical Mycology of CGMH from 2015 to 2021 were used. Antifungal susceptibility testing was performed using the YeastOneTM method. Isolates with high minimal inhibitory concentrations (MICs) against antifungals were further tested using the Clinical and Laboratory Standards Institute (CLSI) broth microdilution method. Mutations in the cyp51A in azole-resistant strains were detected by Sanger sequencing. The overall prevalence of azole-resistant isolates was 1.77% (two out of 113 isolates). The two azole-resistant strains had tandem repeats (TR) in the promoter region and mutations in the cyp51A gene (TR34/L98H and TR34/L98H/S297T/F495I). One strain showed intermediate susceptibility to voriconazole, and its Cyp51A protein had five amino acid substitutions (F46Y/M172V/N248T/D255E/E427K). TR34/L98H and TR34/L98H/S297T/F495I are the most prevalent cyp51A mutations in Taiwan, mediating azole resistance based on current publications and our results. YeastOneTM was validated as a rapid tool for the antifungal susceptibility test; however, further confirmation by CLSI should be considered when MIC values of voriconazole, posaconazole, and amphotericin B are close to the clinical breakpoints or ecological cutoff values.


Introduction
Aspergillus is a genus of ubiquitous saprophytic fungi and is an important opportunistic pathogen in humans. The main route of infection is the respiratory tract through inhalation of fungal conidia [1]. The disease spectrum varies depending on the underlying immune status of the host, and the coincidence of acute bronchopulmonary allergy, severe asthma with fungal sensitization, allergic fungal rhinosinusitis, Aspergillus bronchitis, chronic pulmonary aspergillosis, and invasive aspergillosis [1,2].
Aspergillus fumigatus, A. flavus, A. terreus, A. niger, A. nidulans, and A. versicolor are important pathogens of the genus Aspergillus [3]. A. fumigatus is the predominant species causing disease in humans [4,5]. Azole resistance in A. fumigatus isolates is increasingly reported worldwide [6]. Extensive use of azole fungicidal agents in agriculture has been proven to be associated with azole-resistant isolates of A. fumigatus from an environmental source [7,8]. The molecular mechanisms of azole resistance have been extensively investigated in the past decades. One of the major mechanisms of resistance is mediated by mutations in the cyp51A gene, which encodes a key enzyme in the biosynthesis of fungal ergosterol. Mutation in the coding region results in the production of Cyp51A with low affinity to azole antifungals [9]. This resistance is exaggerated when there are tandem repeats (TR) in the promoter region of the cyp51A gene, which enhances the expression of Cyp51A and the translation of the mutated protein [10]. The 34-bp TR and 46-bp TR within the promoter region with add-on amino acid substitutions were recognized as the predominant resistance mechanisms in environmental A. fumigatus strains [11,12]. The prevalent mechanisms were TR 34 /L98H and TR 46 /Y121F/T289A, which conferred resistance or decreased susceptibility to multiple-azoles [13][14][15].
The first azole-resistant A. fumigatus in Taiwan was reported by Wu et al. in 2015. In their report, three azole-resistant A. fumigatus were identified in 38 isolates from two university hospitals, all of which carried TR 34 /L98H mutations in cyp51A [16]. Later, the same group launched a multicenter study that tested 375 A. fumigatus isolates from 11 hospitals collected from 2011 to 2018. Azole resistance was detected in 19 isolates (5.1%), and cyp51A mutations included TR 34 /L98H, TR 34 /L98H/S297T/F495I, and TR 46 /Y121F/T289A [17]. Chang Gung Memorial Hospital (CGMH) is the largest tertiary referral hospital in Taiwan, with about 3700 beds and 300 ICU beds. The isolates in this study are representative of Northwestern Taiwan.
In this study, we assessed clinical A. fumigatus isolates from the CGMH Linkou Main Branch from 2015 to 2021 and aimed to evaluate their susceptibility to different antifungals and examine the cyp51A mutations in azole-resistant isolates.

Preparation and Molecular Identification of Fungal Isolates
Fungal isolates preserved and identified as A. fumigatus in the Research Laboratory of Medical Mycology of CGMH from 2015 to 2021 were used in this study. The fungal isolates were subcultured on potato dextrose agar at 25 • C for 7 days. Morphological identification was performed by examining the macroscopic and microscopic features, followed by molecular identification for the isolates that could not be identified solely by morphological characteristics. Fungal genomic DNA was extracted using a Smart LabAssist (TANBead TM , Taoyuan City, Taiwan) automatic DNA extraction system. Internal transcribed spacers of ribosomal DNA (ITS) and the β-tubulin gene (benA) were amplified and sequenced. The calmodulin gene (CAL) was also used for identification. The primer pairs used were ITS1 (5 -TCCGTAGGTGAACCTGCGG-3 ) and ITS4 (5 -TCCTCCGCTTATTGATATGC-3 ) for ITS; Bt2a (5 -GGTAACCAAATCGGTGCTGCTTTC-3 ) and Bt2b (5 -ACCCTCAGTGTAGTGACCCTTGGC-3 ) for benA; and CMD5 (5 -CCGAG TACAAGGAGGCCTTC-3 ) and CMD6 (5 -CCGATAGAGGTCATAACGTGG-3 ) for CAL. The PCR conditions for each DNA segment were as previously described [18,19]. The sequences of the isolates were used as queries in the BLAST search against the NCBI GenBank database (https://blast.ncbi.nlm.nih.gov/Blast.cgi) (accessed on 1 March 2022).

Detection of cyp51A Mutations in Azole-Resistant Isolates
The analysis of cyp51A mutations was performed for isolates with high MIC values for the azole antifungals, as confirmed by the aforementioned methods. The entire cyp51A coding region and its promoter region were amplified by PCR. Primers used for the promoter region were TR34-F [5 -TAATCGCAGCACCACTTCAG-3 ] and TR34-R

Ethics Statement
This study was approved by the IRB of Chang Gung Medical Foundation (approval number 202200679B0 obtained on 21 February 2022) Patient consent was waived by the IRB.

Fungal Isolates
A total of 114 isolates identified as A. fumigatus were used in this study. One isolate was later identified as A. flavus based on its ITS and benA sequences and excluded. Other isolates showing slow growth rates, atypical colony colors, or slow sporulation were all confirmed to be A. fumigatus based on their sequence characteristics. The final number of isolates enrolled in this study was 113. All 113 isolates were isolated from clinical specimens including sputum (n = 58), bronchial lavage (n = 39), wound (n = 10), deep tissue (n = 4), and cornea (n = 2).
(CGMHD 0641, CGMHD 0744, CGMHD 2261) and one with 8 μg/mL (CGMHD Two isolates showed high MIC values of >16 μg/mL for ITC (CGMHD 1652 and 22 these seven isolates, which showed high MICs for AMB, VRC, or ITC, the CLSI M microdilution method was used to confirm their MIC values.    Table 2 shows the MICs determined by the CLSI method and the comparison values with those of YeastOne TM . Three isolates (CGMHD 1497, CGMHD 15 CGMHD 2417) showed high MIC values for AMB by YeastOne TM and had lowe of 0.5-1 μg/mL by CLSI. The difference was up to three 2-fold dilutions. The MIC of VRC determined by the CLSI methods were generally lower than those by Yea (0.25-4 μg/mL vs. 0.5-8 μg/mL). However, the difference was within one 2-fold d The MIC values of ITC by the CLSI methods were higher than those by YeastOne TM >16 μg/mL vs. 0.06->16 μg/mL), and the difference was as high as three 2-fold d The MIC values of POS determined by CLSI methods were also higher than tho YeastOne TM (0.125-0.5 μg/mL vs. 0.03-0.5 μg/mL) and the difference was two 2-fo tions. CGMHD 1652 and CGMMHD 2261 are voriconazole-resistant strains acco the clinical breakpoint (CBP) of 2 μg/mL defined by CLSI M61 2nd edition [23]. sistance rate in this study was 1.77%. Both strains also had high MIC values for I μg/mL) and ISA (≥4 μg/mL). The isolate CGMHD 0744 was intermediately susce VRC, with an MIC value of 1 μg/mL.   Table 2 shows the MICs determined by the CLSI method and the comparison of their values with those of YeastOne TM . Three isolates (CGMHD 1497, CGMHD 1524, and CGMHD 2417) showed high MIC values for AMB by YeastOne TM and had lower values of 0.5-1 µg/mL by CLSI. The difference was up to three 2-fold dilutions. The MIC values of VRC determined by the CLSI methods were generally lower than those by YeastOne TM (0.25-4 µg/mL vs. 0.5-8 µg/mL). However, the difference was within one 2-fold dilution.
The MIC values of ITC by the CLSI methods were higher than those by YeastOne TM (0.125->16 µg/mL vs. 0.06->16 µg/mL), and the difference was as high as three 2-fold dilutions. The MIC values of POS determined by CLSI methods were also higher than those from YeastOne TM (0.125-0.5 µg/mL vs. 0.03-0.5 µg/mL) and the difference was two 2-fold dilutions. CGMHD 1652 and CGMMHD 2261 are voriconazole-resistant strains according to the clinical breakpoint (CBP) of 2 µg/mL defined by CLSI M61 2nd edition [23]. The resistance rate in this study was 1.77%. Both strains also had high MIC values for ITC (>16 µg/mL) and ISA (≥4 µg/mL). The isolate CGMHD 0744 was intermediately susceptible to VRC, with an MIC value of 1 µg/mL.

Detection of cyp51A Mutations in Azole-Resistant Isolates
The cyp51A gene of two azole-resistant strains and one intermediately susceptible strain was sequenced to detect mutations. Three different mutation patterns were detected in these three strains. Two resistant strains showed TR 34 /L98H (CGMHD 1652) and TR 34 /L98H/S297T/F495I (CGMHD 2261) mutations in cyp51A. The strain CGMHD 0744, which had intermediate susceptibility to VRC, had no tandem repeat in its cyp51A promoter region but had five amino acid substitutions (F46Y/M172V/N248T/D255E/E427K) due to point mutations in its coding regions.

Clinical Profiles of Patients from Whom the Azole-Resistant A. fumigatus Were Isolated
The clinical profiles of three patients, from whom azole-resistant A. fumigatus were isolated, are summarized in Table 3. CGMHD 1652 was isolated from the bronchoalveolar lavage of a patient diagnosed with bronchiectasis and bronchiolitis with symptoms of hemoptysis. The patient did not receive antifungal agents and had a favorable outcome. It is likely that the isolate was colonized and did not result in infection clinically. CGMHD 2261 was obtained from the bronchoalveolar lavage of a patient diagnosed with invasive fungal tracheobronchitis with a serum galactomannan value of 0.82. The patient died of respiratory failure and multi-organ failure, despite systemic VRC therapy. CGMHD 0744 was isolated from the bronchoalveolar lavage of a patient with necrotizing pneumonia caused by Klebsiella pneumoniae and pulmonary aspergillosis. The patient received systemic treatment with CAS but died of profound shock and multi-organ failure. The last two cases were considered as azole-resistant and intermediately susceptible strains related mortalities.

Discussion
Invasive aspergillosis (IA) is the most common invasive fungal infection in hematopoietic stem cell transplant recipients [24]. Other patient groups at risk of IA include those with prolonged neutropenia, cellular immunity deficiency, and those receiving immunosuppressive therapy, especially patients with graft-versus-host disease. In addition, increasing the use of biologics with tumor necrosis factor blockers, B cell-targeting monoclonal antibody (anti-CD20), and anti-rejection monoclonal antibodies (anti-CD52, CD25) have been associated with cases of IA [25].
The emergence of drug resistance in Aspergillus is an alarming issue in clinical practice because it may result in treatment failure or breakthrough infection, leading to patient mortality. Fungi can evolve different strategies to develop drug resistance. cyp51A encodes a key enzyme in the biosynthesis of ergosterol, and the Cyp51A protein is known as the target of azoles. The mutation of cyp51A caused altered affinity for azoles and mainly contributed to azole resistance [26]. The overproduction of the Cyp51A protein also led to the need for elevated effective drug concentrations and consequently resulted in azole resistance [27]. Resistance to polyenes has been reported with geographic specificity. Prevalence rates of AMB-resistant isolates have been reported to be 27% in Brazil and up to 94% in Hamilton, Canada [28,29]. Although still largely unknown, Ashu et al. proposed that the mechanism may be similar to the intrinsic resistance to AMB of A. terreus, associated with the upregulation of ergosterol biosynthesis genes and increased the expression of superoxide dismutase and catalase-encoding genes [29,30]. Decreased echinocandin susceptibility of A. fumigatus related to anidulafungin exposure was reported recently [31,32]. Mutations on FKS1, which encode the drug target β-1, 3-D-glucan synthase, are reported as the main mechanism. This results in the decreased sensitivity of glucan synthase to echinocandins and decreased echinocandin susceptibility and was associated with treatment failure in the case of chronic pulmonary aspergillosis with an aspergilloma [31,33].
Although isolates resistant to multiple azoles have been increasingly reported, CLSI did not establish CBPs for antifungal activity against A. fumigatus until the latest version of the 2nd edition of the CLSI M61 Performance Standards for Antifungal Susceptibility Testing of Filamentous Fungi was published in 2020 [23]. In this edition, the CBP of VRC is 2 µg/mL. Isolates were classified as susceptible, intermediate, and resistant if they had a VRC MIC value of <1 µg/mL, 1 µg/mL, and ≥2 µg/mL, respectively. This is the only antifungal drug for which CBP is defined in the CLSI. For other antifungals, the epidemiological cutoff values (ECVs) for A. fumigatus are provided in the 3rd edition of the CLSI M59 document [34]. The European Committee on Antimicrobial Susceptibility Testing (EUCAST) v. 10.0, also updated CBPs for ISA, ITC, VRC, and POS against A. fumigatus in 2020 [35]. Due to the different methodologies employed by CLSI and EUCAST, the interpretation of MIC values as susceptible or resistant should adhere to the protocol used. In this study, isolates CGMHD 1652 and CGMHD 2261 showed pan-azole resistance to VRC, ITC, and ISA, while CGMHD 0744 only showed elevation of VRC MIC but did not reach the resistance criteria.
Several mechanisms for cyp51A-mediated resistance have been reported including single point mutations (SNPs), multiple point mutations, and tandem repeats with or without point mutations. Some hotspot single-point mutations contribute to different phenotypes of resistance. These hotspots included G54, M220, G138, and G448. Most of these strains were resistant to multiple azoles [6,13]. Previous reports have proposed that these mutations may be associated with blocking the entry or modifying the binding site of the azoles, which further cause drug affinity reduction [14]. The combination of multiple SNPs, which cause amino acid substitutions, is another resistance mechanism. The most frequently reported mutations were the F46Y/M172V/D255E and F46Y/M172V/N248T/D255E/E427K substitutions. Strains possessing these two mutations have been reported to have different degrees of azole susceptibility. Even when regarded as "susceptible", strains with these mutations generally showed higher MIC values for azoles compared to wild-type strains [13]. The susceptibility test of CGMHD 0744, which had five substitutions, only showed an elevation of the VRC MIC value, suggesting a marginal effect of this mutation. In a study of Cyp51A protein homology models, M172V, N248T, D255E, and E427K were nonsynonymous mutations that were located in non-conserved areas on the surface of the protein. Therefore, they were predicted to not interact with azole compounds or affect their structural integrity. In contrast, the substitution F46Y affected the transmembrane domain and substrate access channel of Cyp51A. Thus, the F46Y mutation was believed to be partially or totally responsible for the slightly higher azole MICs that these strains showed, due to the potential block of the substrate entrance channel [9]. In addition to point mutations, a TR in the promoter region, resulting in the overexpression of cyp51A, was related to azole resistance. Strains with TR or point mutations (L98H or Y121F) alone were found to have only a moderate increase in the MIC of the azoles. Strains with a combination of TR and amino acid substitutions can achieve significant azole resistance [27].
According to Wu et al. in 2020, the prevalence of azole-resistant isolates from 11 hospitals in Taiwan from 2011 to 2018 was 5.1% (19 of 375 isolates) [17]. The prevalence of azole-resistant isolates in our study was 1.77% (two out of 113 isolates). These two reports included the largest secondary and tertiary referral hospitals in Taiwan. Combining these two reports, the overall prevalence rate of the azole-resistant isolates was 4.3%. Among these azole-resistant isolates, six isolates had the TR 34 /L98H mutation, eight isolates had the TR 34 /L98H/S297T/F495I mutation, one isolate had the TR 46 /Y121F/T289A mutation, five isolates had the F46Y/G89G/M172V/N248T/D255E/L358L/E427K/C454C polymorphism, and one isolate had the F46Y/M172V/N248T/D255E/E427K polymorphism in the cyp51A gene. Another study by Chen et al. in 2019 reported that seven of 22 isolates of A. fumigatus from the environment were azole-resistant. Another two isolates of A. fumigatus isolated from patients in the study were azole-susceptible. Among the azoleresistant environmental isolates, three isolates had a TR 34 /L98H/S297T/F495I mutation, two isolates had TR 34 /L98H mutation, and two isolates hadTR 34 /L98H/S297T/F495I in the cyp51A gene [36]. Since most isolates were from the environment, we did not include the data from Chen et al. for the calculation of the clinical prevalence rate. The most prevalent two mutations reported by publications from Taiwan, TR 34 /L98H and TR 34 /L98H/S297T/F495I, which were found to be prevalent in the environment, are re-lated to azole fungicide use in agriculture [37]. This was also proposed to be associated with the increasing use of azole fungicides (mainly difenoconazole, tebuconazole, and propiconazole) in Taiwan over the last two decades [8,17].
Azole-resistant A. fumigatus isolates have been reported globally, with prevalence varying from less than 1% to up to 28% based on different geographic regions [11,38]. The major prevalence of azole-resistant isolates was reported in European countries, with the highest prevalence in the United Kingdom [11,13]. In Asia, lower rates of azole resistance have been reported in Taiwan, Japan, China, and India [11,13]. Most of the reported resistance rates in these countries are lower than 10%. The overall resistance rate and related mutation mechanisms in Taiwan were similar to reports from other Asian countries [11,15,39,40]. However, the resistance rate varies between countries and also within a single country [13,41]. The variation in the prevalence of resistance may result from the geographic location, difference in laboratory practice, study design, and the diseases enrolled in each study [41]. A higher prevalence of azole-resistant isolates was found in some patients with specific risk factors such as previous prolonged azole exposure. Singh et al. reported that the resistance rate of isolates from patients with chronic pulmonary aspergillosis was up to 59% [42]. The high percentage of resistant isolates was probably linked to previous ITC exposure, which may be encountered in patients with aspergilloma, chronic aspergillosis, cystic fibrosis, and predisposing lung cavities [43]. Previous reports have also found that patients with hematologic or oncologic disease were more likely to harbor azole-resistant isolates, which cause invasive aspergillosis and were associated with high mortality [44,45].
According to the updated practice guidelines of the Infectious Diseases Society of America (IDSA) in 2016, the primary treatment for most invasive aspergillosis remains to be VRC, while AMB and ISA are alternative choices. Combination antifungal therapy with VRC and echinocandin may be considered in select patients [46]. Case 1 with the CGMHD 1652 strain received no antifungal therapy and had a favorable outcome. Case 2 with the CGMHD 2261 strain received VRC treatment as per the guidelines. The MIC of VRC for this isolate was 2 µg/mL as determined using the CLSI method, and this isolate was resistant to VRC. No susceptibility data were available at that time, thus, no alternative or combined antifungal agents were used. The patient had poor treatment outcomes and died of respiratory failure. Case 3 with the CGMHD 0744 strain was infected with an intermediate azole-susceptible isolate and treated with CAS because of an underlying condition of liver failure. Although the MEC of CAS was low, the patient died of profound shock and multi-organ failure. None of the patients received a combination of antifungal agents. However, in a randomized, double-blind multicenter study, mortality rates were higher in patients diagnosed with invasive aspergillosis receiving monotherapy with VRC than in those receiving combined therapy with AMB [47]. Alternative therapeutic agents and/or combined therapy should be considered in countries with azole-resistant strains or in patients with poor treatment response. Antifungal susceptibility tests should also be considered and might be helpful for the choice of alternative agents.
The MIC values determined by YeastOne TM and CLSI slightly differed in this study. Generally, the MIC values of AMB and VRC obtained from YeastOne TM were higher than those from CLSI. The MIC values of ITC and POS obtained by YeastOne TM were lower than those obtained by CLSI. Similar observations were also noted in previous publications [48][49][50]. Likewise, higher MICs of AMB of some isolated by YeastOne TM in this study were in fact wild type strains by CLSI. Thus, although YeastOne TM had a good performance of an overall agreement of more than 90% with CLSI when testing the Aspergillus species [48], confirmation with CLSI or EUCAST to identify resistant isolates was necessary when the MIC levels of ITC, VRC, POS, and AMB were close to CBP or ECV.
There are several cryptic species in the A. fumigatus species complex such as A. lentulus, A. novofumigatus, A. fischeri, A. viridinutans, and A. udagawae, which may be intrinsically resistant to one or more antifungal agents [51,52]. According to a multicenter study from 19 countries, the overall prevalence of azole resistance was 3.2% among 2941 A. fumigatus species complex isolates. However, up to 21.7% of the azole-resistant isolates were not identified as A. fumigatus, but as cryptic species such as A. lentulus, A. thermomutatus, and A. udagawae [53]. This finding emphasizes the importance of the correct molecular identification of species when treating Aspergillus infections, especially invasive aspergillosis. Although some isolates demonstrated atypical morphology such as pale colony color or delayed sporulation, we did not find any antifungal-resistant cryptic species in this study using sequence-based identification.

Conclusions
Azole-resistant isolates of A. fumigatus are present in CGMH and all over Taiwan. TR 34 /L98H and TR 34 /L98H/S297T/F495I are the most prevalent cyp51A mutations based on current publications and this study. Accurate species identification with aggressive antifungal susceptibility plays an important role in the choice of antifungal agent. YeastOne TM is a rapid and useful clinical tool for the antifungal susceptibility test. Although high agreement between YeastOne TM and the CLSI method was reported, a discrepancy between the results is still possible. When utilizing YeastOne TM in a clinical setting, the results should be interpreted carefully, and further confirmation by CLSI should be always considered when the MIC levels of VRC, POS, and AMB were close to CBP or ECV. Modification of clinical treatment strategies based on susceptibility testing is necessary when encountering azole-resistant A. fumigatus, which may be helpful to prevent morbidity and mortality.  Informed Consent Statement: The patient consent was waived by the IRB because only medical charts were reviewed, and the patient's identity was kept anonymous. Data Availability Statement: Data presented in this study are available upon request from the corresponding author.

Conflicts of Interest:
All authors declare no conflict of interest.