In 2013, a 69-year-old man who was a former smoker with chronic obstructive pulmonary disease (COPD) and severe airflow obstruction sought care at the University Hospital in Århus, Denmark, because of gradually worsening dyspnea, cough, and expectoration. Previously, in 2011, imaging (Figure 1, panel A) and 2 thoracoscopies had been conducted because of suspicion of malignant mesothelioma. Further histopathologic examination and cultures revealed inflammation but no malignancy or mold infection. Subsequently, in 2012, a fistula between pleura and skin led to a persistent air-containing pleural cavity in the right side (Figure 1, panel B). In 2014, a fungus ball in the pleural cavity was found (Figure 1, panel C). Aspergillus IgG titer was 1:25,600 (reference range <1:200), and azole-susceptible A. fumigatus was cultured from sputum (P-1, May 2014). Voriconazole (200 mg 2×/d) was given, alternating with posaconazole (300 mg/d) for 2 years until clinical failure, and 2 azole-resistant A. fumigatus isolates were cultured from a new sputum sample (P-2 and P-3, June 2016). Despite amphotericin B inhalations followed by liposomal amphotericin B (3 mg/kg 1×/d), the patient died because of severe hemoptysis 1 year later in 2017.

Three A. fumigatus patient isolates (P-1, P-2, and P-3) were available for confirmatory species verification, reference susceptibility testing defined by the European Committee on Antimicrobial Susceptibility Testing using protocol for molds (E.Def 9.3), cyp51A Sanger sequencing (using wild-type reference sequence AF338659), and genotyping using the short tandem-repeat Aspergillus fumigatus (STRAf) assay (2,3) (Table). We included 4 A. fumigatus isolates representing relevant cyp51A profiles as control strains (SSI-3614 [wild-type], SSI-7828 [TR₃₄/L98H], SSI-5717 [TR₄₆/Y121F/T289A], and SSI-5197 [F46Y/M172V/E427K]). We detected 3 common Cyp51A variants (F46Y, M172V, and E427K) in the susceptible patient isolate P-1 (GenBank accession no. MG972984). Pan-azole resistance was observed for P-2 and P-3, and both shared cyp51A profiles with P-1 but also harbored a TR₁₂₀ mechanism (GenBank accession no. MG972983) in the promoter region (Table). All patient isolates had identical STRAf genotypes suggesting that they were isogenic (Table ) (4). Furthermore, the STRAf profile was unique among A. fumigatus isolates genotyped in Denmark (Appendix Figure).

We performed WGS for P-1, P-3, and all control strains to investigate relatedness and other potential mechanisms conferring azole resistance. We subjected total DNA (≈10 ng/µL) to WGS (NextSeq 550; Illumina, https://www.illumina.com) by using Nextera DNA library preparation kit (Illumina) and following the manufacturer’s instructions. We used NASP (5) to detect single-nucleotide polymorphisms (SNPs) after removal of duplicated regions in the A. fumigatus strain Af293 chromosomes (http://www.aspergillusgenome.org, genome version s03-m05-r09) using NUCmer (6). We inferred relatedness by using FastTree version 2.1.5 (7) and a 77.69% core genome (Table; Figure 2). To increase resolution, we conducted a subanalysis for P-1 and P-3 (core genome 79.71%), which identified 41 SNP differences; 6 of the SNPs were nonsynonymous in genes with no previous reported association to azole resistance (Appendix Table 1), and 35 were either synonymous or in noncoding regions (Appendix Table 2).

WGS revealed 41 SNP differences between the susceptible and the resistant patient A. fumigatus isolates that evolved during 2 years, similar to a previously described case of in-host microevolution of A. fumigatus (4). This finding substantiated an isogenic relationship between P-1 and P-3 and demonstrated that the TR₁₂₀ resistance mechanism emerged from P-1, probably during long-term azole therapy. Furthermore, WGS results supported the conclusion that the TR₁₂₀ was the sole mechanism of azole resistance in the azole-resistant patient isolates.

To our knowledge, the TR₁₂₀ is a novel azole-resistance mechanism in A. fumigatus, and the in vivo selection of a tandem repeat in the promoter of cyp51A is unique. The de novo acquisition of a TR has not previously been shown in vitro or in the environment (i.e., no isolates with L98H or Y121F+T289A combined with wild-type promoters have been reported). However, triplication of an existing TR₃₄ on tebuconazole exposure was selected in vitro, and a novel variant, TR₄₆³, found in clinical and environmental samples, has been derived from sexual mating between TR₄₆ parents (8,9).

Azole resistance involving TRs in the promoter region has been associated exclusively with environmental fungicide selection pressure in A. fumigatus and other plant pathogens. Furthermore, although asexual propagation of A. fumigatus with TR₃₄/L98H or TR₄₆/Y121F/T289A resistance mechanisms is widespread in the environment, the extent of de novo selection of TR₃₄/L98H and TR₄₆/Y121F/T289A is unclear (10). One hypothesis describes both environmental resistance mechanisms as being derived from single events of sexual reproduction (in environmental habitats) combining the TR with a cyp51A mutant. In addition, sexual reproduction might have led to a high genetic diversity among environmental azole-resistant A. fumigatus, which otherwise might have indicated multiple origins (10). Our finding might challenge the perception that TR azole-resistance mechanisms are exclusive to the environment and might warrant the question of whether TR₃₄/L98H and TR₄₆/Y121F/T289A derive from single events.

Hypothetically, the patient might initially have inhaled isogenic isolates with and without TR₁₂₀, the resistant one being undetected. However, a patient being co-infected de novo by a susceptible and an isogenic resistant strain has not been previously reported and is considered highly unlikely.

Long-term and subtherapeutic antifungal treatment might facilitate selection of resistance (11). Therapeutic drug monitoring was performed once in this patient but without information if the sample was taken according to guidelines as a trough level (lowest level after dosage). Thus, despite a concentration of 4.3 mg/L (within the recommended trough range), potential subtherapeutic levels during the 200 mg 2×/d dosing scheme cannot be ruled out. The F46Y/M172V/E427K substitutions in Cyp51A, found in both susceptible and resistant isolates, have been suggested to play no role or only a minor role in reduced azole susceptibilities (12,13). TRs in the promoter region of cyp51A have previously been linked to increased cyp51A gene expression and MICs because of duplicated srbA transcription factor binding motifs (SRE1 and SRE2) leading to increased expression of cyp51A (14,15). Taken together, our data suggest that TR₁₂₀ alone is an important driver of pan-azole resistance at a level comparable to that known to be mediated by the TR₃₄/L98H mechanism.

Our WGS results might obviate the desire for in vitro experiments testing the TR₁₂₀ mechanism in laboratory-engineered mutants. Further dissection of the WGS data can help elucidate potential genetic drivers of TR acquisition and add further knowledge as to whether the TR₃₄/L98H and TR₄₆/Y121F/T289A resistance genotypes derived from a single origin. This report adds another piece to the complex picture of emerging azole-resistant A. fumigatus and might serve to stimulate further research.

Additional information on in vivo selection of a unique tandem repeat mediated azole resistance mechanism (TR₁₂₀) in Aspergillus fumigatus cyp51A, Denmark.