Genomic Epidemiology Identifies Azole Resistance Due to TR34/L98H in European Aspergillus fumigatus Causing COVID-19-Associated Pulmonary Aspergillosis

Abstract

Aspergillus fumigatus has been found to coinfect patients with severe SARS-CoV-2 virus infection, leading to COVID-19-associated pulmonary aspergillosis (CAPA). The CAPA all-cause mortality rate is approximately 50% and may be complicated by azole resistance. Genomic epidemiology can help shed light on the genetics of A. fumigatus causing CAPA, including the prevalence of resistance-associated alleles. We present a population genomic analysis of 21 CAPA isolates from four European countries with these isolates compared against 240 non-CAPA A. fumigatus isolates from a wider population. Bioinformatic analysis and antifungal susceptibility testing were performed to quantify resistance and identify possible genetically encoded azole-resistant mechanisms. The phylogenetic analysis of the 21 CAPA isolates showed that they were representative of the wider A. fumigatus population with no obvious clustering. The prevalence of phenotypic azole resistance in CAPA was 14.3% (n = 3/21); all three CAPA isolates contained a known resistance-associated cyp51A polymorphism. The relatively high prevalence of azole resistance alleles that we document poses a probable threat to treatment success rates, warranting the enhanced surveillance of A. fumigatus genotypes in these patients. Furthermore, potential changes to antifungal first-line treatment guidelines may be needed to improve patient outcomes when CAPA is suspected.

Keywords: Aspergillus fumigatus; CAPA; COVID-19-associated pulmonary aspergillosis; azole-resistant Aspergillus fumigatus; coinfection; genetic epidemiology; genomic analysis.

Figure A1

Normalised depth of coverage by locus indicative of duplications and deletion events. (a) C440 Chromosome 1. The deleted region is also found in C408 and C439–C441. (b) C440 chromosome 4. Two regions, a 500 kbp deletion identified in all isolates, and a 1300 kbp deletion around the centromere of Chromosome 4. (c) C440 Chromosome 7. Two deletions in Chromosome 7, relating to the centromere and a region of 1600 kbp found in CAPA-A, CAPA-B, C403, 408, C436 and C439-C441, 444. (d) C444 Chromosome 8. Small deletion of 1400 kbp found in Chromosome 8 in isolates CAPA-B, CAPA-C, C424, C436, C441, 444.

Figure A2

Phylogenetic tree of CAPA isolates, A. fumigatus IA control isolates from London, UK, and isolates from colonising COVID-19 patients (Netherlands and Ireland). Rooted ML phylogenetic tree with bootstraps support over 1000 replicates performed on WGS SNP data. Branch length represents the average number of SNPs. * 0.04 = 3242 SNPs.

Figure A3

Phylogenetic tree of CAPA and A. fumigatus clinical and environmental isolates from Ireland and the UK. Rooted ML phylogenetic tree, showing: right track 1, if the isolate contains cyp51A polymorphism; middle track 2, the clade the isolate is located; and left track 3, the source of the isolate (CAPA, clinical, or environmental). Branch length represents the average number of SNPs. * 0.1 = 33,839 SNPs.

Figure A4

Phylogenetic tree of CAPA and A. fumigatus clinical and environmental isolates from Ireland and the UK. Rooted ML phylogenetic tree with bootstraps support over 1000 replicates performed on WGS SNP data. Branch length represents the average number of SNPs. * 0.1 = 33,839 SNPs.

Figure 1

Phylogenetic tree of isolates from CAPA, IA, and colonising patients. Unrooted maximum likelihood (ML) phylogenetic tree over 1000 replicates performed on WGS SNP data, showing the following: inner track 1, the source of the isolate (CAPA, colonising or IA); middle track 2, if the isolate contains cyp51A polymorphism; and outer track 3, the clade in which the isolate is located.

Figure 2

Multivariate analysis of isolates from CAPA, IA, and colonising patients. (A) Scatterplot of the principal component analysis (PCA) A. fumigatus genotypes using the first four principal components (PCs) illustrating the genetic identity of CAPA and IA control isolates. (B) Density plot of the discriminant PCA (DAPC), broadly identifying two distinct clusters, CAPA and control. (C) Composition plot highlights that all the isolates’ genotype is composed of genetic material identified as CAPA, IA, or colonising. Four isolates contain 70% control compared to CAPA, whereas there are 21 isolates the genome membership of which is mostly CAPA (>80%).

Figure 2

Multivariate analysis of isolates from CAPA, IA, and colonising patients. (A) Scatterplot of the principal component analysis (PCA) A. fumigatus genotypes using the first four principal components (PCs) illustrating the genetic identity of CAPA and IA control isolates. (B) Density plot of the discriminant PCA (DAPC), broadly identifying two distinct clusters, CAPA and control. (C) Composition plot highlights that all the isolates’ genotype is composed of genetic material identified as CAPA, IA, or colonising. Four isolates contain 70% control compared to CAPA, whereas there are 21 isolates the genome membership of which is mostly CAPA (>80%).

Figure 3

Phylogenetic tree of CAPA and A. fumigatus clinical and environmental isolates from Ireland and the UK. Unrooted ML phylogenetic tree with over 1000 replicates performed on WGS SNP data, showing the following: outer track 1, if the isolate contains cyp51A polymorphism; middle track 2, the clade in which the isolate is located; and inner track 3, the source of the isolate (CAPA, clinical, or environmental).

Figure 4

Multivariate analysis of CAPA and A. fumigatus clinical and environmental isolates from Ireland and the UK. Multivariate analysis, (A) scatterplot of the PCA A. fumigatus genotypes using the first five PCs illustrating the genetic identity of CAPA isolates, and clinical and environmental isolates from Ireland and the UK. (B) Scatterplot of the DPCA identifying 3 types form separate clusters, with CAPA and clinical with some overlap. (C) A composition plot comparing the genetic composition of each isolate. The plot highlights that each isolate is a mixture of genotypes identified as CAPA, environment, and clinical non-CAPA.

Figure 4

Multivariate analysis of CAPA and A. fumigatus clinical and environmental isolates from Ireland and the UK. Multivariate analysis, (A) scatterplot of the PCA A. fumigatus genotypes using the first five PCs illustrating the genetic identity of CAPA isolates, and clinical and environmental isolates from Ireland and the UK. (B) Scatterplot of the DPCA identifying 3 types form separate clusters, with CAPA and clinical with some overlap. (C) A composition plot comparing the genetic composition of each isolate. The plot highlights that each isolate is a mixture of genotypes identified as CAPA, environment, and clinical non-CAPA.