First Detection of TR₄₆/Y121F/T289A and TR₃₄/L98H Alterations in Aspergillus fumigatus Isolates from Azole-Naive Patients in Denmark despite Negative Findings in the Environment

Abstract

Azole-resistant Aspergillus fumigatus harboring the TR₃₄/L98H or TR₄₆/Y121F/T289A alterations is increasingly found in Europe and Asia. Here, we present the first clinical cases of TR₄₆/Y121/T289A and three cases of TR₃₄/L98H outside the cystic fibrosis (CF) population in Denmark and the results of environmental surveys. Four patients (2012 to 2014) with 11 A. fumigatus and 4 Rhizomucor pusillus isolates and 239 soil samples (spring 2010 and autumn 2013, respectively) with a total of 113 A. fumigatus isolates were examined. Aspergillus isolates were screened for azole resistance using azole-containing agar. Confirmatory susceptibility testing was done using the EUCAST microbroth dilution EDEF 9.1 reference method. For relevant A. fumigatus isolates, CYP51A sequencing and microsatellite genotyping were performed. Three patients harbored TR₃₄/L98H isolates. Two were azole naive at the time of acquisition and two were coinfected with wild-type A. fumigatus or R. pusillus isolates, complicating and delaying diagnosis. The TR₄₆/Y121F/T289A strain was isolated in 2014 from a lung transplant patient. Genotyping indicated that susceptible and resistant Aspergillus isolates were unrelated and that no transmission between patients occurred. Azole resistance was not detected in any of the 113 soil isolates. TR₃₄/L98H and TR₄₆/Y121F/T289A alterations appear to be emerging in the clinical setting in Denmark and now involve azole-naive patients. Two recent soil-sampling surveys in Denmark were unable to indicate any increased prevalence of azole-resistant A. fumigatus in the environment. These findings further support the demand for real-time susceptibility testing of all clinically relevant isolates and for studies investigating the seasonal variation and ecological niches for azole-resistant environmental A. fumigatus.

INTRODUCTION

Azole resistance in Aspergillus fumigatus has been increasingly reported, particularly during the last decade. As opposed to some sibling species of A. fumigatus with naturally occurring intrinsic resistance (e.g., Aspergillus lentulus, Aspergillus udagawae, and Aspergillus fumigatiaffinis [1]), two distinct routes of acquisition of resistance have been described for A. fumigatus, selection in a clinical setting during long-term azole therapy (2, 3) and primary acquisition of a resistant isolate from the environment (4, 5), the advent of which has been coupled to a considerable use of azole fungicides in agriculture and material preservation (6, 7). The majority of azole resistance is caused by alterations in the target protein sterol 14α-demethylase, encoded by the CYP51A gene (8). Two resistance mechanisms have been found both in the environment and in isolates from azole-naive patients. They are characterized by a tandem repeat in the promoter region of the target gene coupled with point mutations in CYP51A (TR₃₄/L98H and TR₄₆/Y121F/T289A). In Denmark, the TR₃₄/L98H genotype was found in the environment (in 2009) and in clinical samples (from 2007 and 2009), but the only humans in whom it has been found so far are azole-exposed cystic fibrosis (CF) patients (9, 10). This contrasts with the scenario in the Netherlands, where both genotypes are common causes of invasive aspergillosis in the hematological/transplant setting, including azole-naive patients (11, 12). Similarly, the TR₃₄/L98H alteration has been found in isolates from azole-naive patients in India and the TR₄₆/Y121F/T289A alteration in an isolate from an azole-naive patient from Belgium (13, 14). Several reports have documented in numbers the continuous spread of these “environmental” azole-resistant A. fumigatus strains in the Netherlands and in Asia (Iran, China and India) (11, 14–19), and together they account for the vast majority of the 6.8% triazole resistance in patient samples in the Netherlands from 2009 to 2011 (20). During 2012 to 2013, the first reports documenting the presence of TR₃₄/L98H in isolates from patients in Germany (21–23) and environmentally in Italy (24) have emerged. Clinical isolates with the emerging TR₄₆/Y121F/T289A mutation have to our knowledge been detected only in the Netherlands and Belgium, while environmental studies have documented the prevalence of this genotype in both countries and in India (13, 25).

Furthermore, both types of mutations have been found in isolates from domestic homes throughout the Netherlands (20). In 2013, the European Center for Disease Prevention and Control released a report integrating all evidence of the fungicide-driven hypothesis and stressing the need to test isolates for susceptibility (26).

We here describe the first isolation of a TR₄₆/Y121F/T289A strain from a Danish patient and the finding of the TR₃₄/L98H genotype in routine samples from three Danish non-cystic fibrosis patients during a 20-month period (May 2012 to January 2014). Two of the patients were azole naive at the time of acquisition. Furthermore, we report the findings of two environmental follow-up surveys of azole resistance in Danish soil samples (obtained in March to April 2010 and September to October 2013).

MATERIALS AND METHODS

Clinical isolates.

Primary culture of respiratory samples or referred mold isolates was performed at the reference laboratory using Sabouraud glucose (pH 4) agar (SSI Diagnostika, Hillerød, Denmark) and yeast extract glucose chloramphenicol (YGC) agar (Oxoid, Germany). Agar plates were incubated at 37°C and examined daily for 5 days. Species identifications of mold isolates were done according to morphological criteria (27). A. fumigatus complex isolates were further incubated at 48°C to separate A. fumigatus sensu stricto from cryptic A. fumigatus complex species. Mucorales isolates additionally underwent examination of carbon assimilation patterns obtained using the commercially available ATB ID32C (bioMérieux, Marcy l'Etoile, France), which were read on day 2 as previously described (28).

Soil sampling. (i) Period 1.

During March to April 2010, 69 soil samples were collected from indoor flowerpots in the tertiary care university hospitals in Copenhagen (Rigshospitalet [∼1,100 beds]) (25 samples), flowerbeds in an amusement park in the center of Copenhagen (Tivoli Gardens) (17 samples), and both conventionally grown fields (12 samples) and organic grown fields (15 samples).

(ii) Period 2.

In September to October 2013, another 170 samples (130 samples from conventionally grown fields and 40 samples from organic fields) were collected, giving a total of 239 samples.

Processing of soil samples was done as previously described (10). Briefly, 2 g of soil from each sample was suspended in 5 ml 0.2 M NaCl with 1% Tween 20 and vortexed. Subsequently, 100 μl of this suspension was plated on Sabouraud dextrose agar (pH 4) (SSI Diagnostika, Hillerød, Denmark). All A. fumigatus isolates were subcultured to obtain pure cultures and identified as described above.

Susceptibility testing.

A four-well multidish plate containing RPMI 1640–2% glucose agar supplemented with itraconazole (4 mg/liter), voriconazole (1 mg/liter), posaconazole (0.5 mg/liter), and no antifungal (positive-control well) (Balis Laboratorium V.O.F., Boven-Leeuwen, the Netherlands) was used to screen for azole resistance in A. fumigatus. All Aspergillus isolates that grew on any one of the azole-containing agars were subjected to in vitro susceptibility testing according to the EUCAST microbroth dilution EDEF 9.1 reference method (29). Stock solutions (5,000 mg/liter in dimethyl sulfoxide [Sigma-Aldrich, Brøndby, Denmark]) of itraconazole (Sigma-Aldrich), voriconazole (Pfizer, Ballerup, Denmark), and posaconazole (MSD, Ballerup, Denmark) were prepared. Final drug concentration ranges were 0.03 to 4 mg/liter for posaconazole and voriconazole and 0.06 to 8 mg/liter for itraconazole. MICs were determined visually as a no-growth endpoint at 48 h of incubation. In addition, susceptibilities to amphotericin B and caspofungin were determined using E-test strips (BioMérieux, Marcy l'Etoile, France) on RPMI 1640–2% glucose agar (SSI Diagnostika). For caspofungin, microcolonies in the inhibition zone were ignored, according to the manufacturer's instructions (30). For interpretation of susceptibility the EUCAST clinical breakpoints for A. fumigatus were used (itraconazole, voriconazole, and amphotericin B MICs of ≤1 mg/liter [susceptible] and >2 mg/liter [resistant] and posaconazole MICs of ≤0.125 mg/liter [susceptible] and >0.25 mg/liter [resistant]). No clinical breakpoints for the echinocandins have yet been established for Aspergillus. No clinical breakpoints have yet been proposed for other molds, but the Aspergillus breakpoints were used as guides for likely clinical susceptibility.

PCR amplification and sequence analysis of the CYP51A gene.

Conidia were inoculated in 3 ml GYEP broth (2% glucose, 0.3% yeast extract, 1% peptone) and grown overnight at 37°C. Mycelial mats were recovered and subjected to DNA extraction. The promoter and full coding sequence of CYP51A were amplified by PCR, and both strands were sequenced as described previously (9, 31). The sequences were assembled and compared to the reference sequence of azole-susceptible wild-type A. fumigatus (GenBank accession no. AF338659) using CLC Main Workbench (CLC bio, Qiagen).

STRAf genotyping of clinical A. fumigatus isolates.

From all clinical A. fumigatus isolates, nine microsatellite (or short tandem repeat [STR]) loci (STRAf 2A, 2B, 2C, 3A, 3B, 3C, 4A, 4B, and 4C) were amplified, fragment sizes were determined, and repeat numbers were assigned as previously described (32). Isolates with identical STRAf genotypes were considered isogenic.

RESULTS

The clinical characteristics of the four patients harboring the resistant isolates are summarized in Table 1. Full case reports are available in File S1 in the supplemental material. All patients had significant underlying disease, three harbored several mold isolates (either wild-type or resistant A. fumigatus or A. fumigatus in combination with a Mucorales species), and two were azole naive at the time of acquisition of the azole-resistant A. fumigatus isolate. Three of the four patients died within 6 weeks (Table 1).

TABLE 1.

Clinical characteristics of the four patients at the time of isolation of the first resistant isolate

Case no.	Sex/age (yr)a	Mo and yr of isolation	Region	Underlying condition	Aspergillus diseaseb	Previous azole exposurec	Treatmentd	Survival after diagnosis (days)e
1	F/77	May 2012	Zealand	COPD,f Wegener's granulomatosis	Probable IA	None (2 days of VRC)	VRC, L-AMB	23
2	F/60	Feb 2013	Jutland	Grade 1–2 follicular lymphoma with paraneoplastic pemphigus, previous breast cancer	Probable IA	VRC	VRC, CAS, L-AMB, POS	1
3	F/73	June 2013	Zealand	Invasive breast cancer (no immunosuppression)	None	None	None	41
4	M/42	Jan 2014	Jutland	Bruton's agammaglobulinemia, double lung transplant	Probable IA	FLC, ITR (5 days), VRC (toxicity issues)	VRC, CAS, L-AMB	100

^aF, female; M, male.

^bAs defined in reference 45.

^cVRC, voriconazole; FLC, fluconazole; ITR, itraconazole.

^dL-AMB, liposomal amphotericin B; POS, posaconazole; CAS, caspofungin.

^eSurvival (in days) after the initial demonstration of an azole-resistant A. fumigatus isolate.

^fCOPD, chronic obstructive pulmonary disease.

The susceptibility patterns, STRAf typing, and CYP51A sequencing data are summarized in Table 2. Of note, the STRAf typing and CYP51A sequencing revealed that the first isolate from patient 1 was a mixed culture of a susceptible and a resistant isolate. This was confirmed through subculturing of dilutions of the initial samples and susceptibility testing on 10 individual colonies, followed by repeated DNA analysis. For patient 2, the initial and third isolates were azole susceptible, in contrast to the second A. fumigatus isolate from this patient. The rapid growth of Rhizomucor pusillus in the last two of these samples complicated the detection and isolating of the Aspergillus isolate. Reisolation and renewed susceptibility testing of the second isolate were conducted before we concluded that it was indeed azole resistant. This was confirmed as sequencing demonstrated the TR₃₄/L98H alteration.

TABLE 2.

Mold isolates, diagnostic delay, MICs, resistance genotypes, and STRAf typing data from the clinical isolates obtained from the four patients

Case no.	Day	Sitea	Diagnostic delay (days)b	Species	MIC (μg/ml) ofc:					CYP51A genotype	STRAf type (2A-2B-2C-3A-3B-3C-4A-4B-4C)d
					POS	VRC	ITR	AMB	CAS
1	7	BAL	12/11	A. fumigatus	0.06	1	0.25	0.5	0.064	WTe	18-19-8-26-10-21-9-9-5
	7	BAL	12/11	A. fumigatus	1	4	>8	0.5	0.064	TR34/L98H plus S297T plus F495I	14-10-9-30-9-6-8-10-20
	17	BAL	14/7	A. fumigatus	0.5	1	>8	0.5	0.064	TR34/L98H plus S297T plus F495I	14-10-9-30-9-6-8-10-20
2	44	BAL	18/8	A. fumigatus	0.03	0.25	0.125	0.25	0.064	WT	14-20-11-34-9-7-8-10-12
	90	TS	20/18	A. fumigatus	0.5	4	>8	0.5	0.032	TR34/L98H	25-10-12-79-9-9-8-10-11
	90	TS	20/18	R. pusillus	0.25	>4	0.5	0.5	>32	NA	NA
	106	TS	13/9	R. pusillus	0.125	>4	0.25	0.5	NA	NA	NA
	110	TS	9/8	R. pusillus	0.125	>4	0.25	0.5	NA	NA	NA
	117	TS	9/8	A. fumigatus	≤0.03	0.5	0.25	1	0.064	WT	25-16-19-48-17-23-8-9-5
	117	TS	9/8	R. pusillus	0.25	>4	0.25	0.5	>32	NA	NA
3	6	BAL	16/10	A. fumigatus	0.5	4	>8	0.25	0.064	TR34/L98H	20-20-28-32-9-6-8-10-20
4	−7	BAL	26/11	A. fumigatus	0.06	0.5	0.125	0.75	0.094	WT	18-25-15-26-11-7-26-30-8
	36	Sputum	7/6	A. fumigatus	0.125	>4	0.25	0.75	0.032	TR46/Y121F/T289A	26-21-16-32-9-10-8-14-10
	58	Sputum	ND/7	A. fumigatus	0.25	>4	0.5	0.5	0.094	TR46/Y121F/T289A	26-21-16-32-9-10-8-14-10
	62	Sputum	10/9	A. fumigatus	0.25	>4	0.5	0.5	0.064	TR46/Y121F/T289A	26-21-16-32-9-10-8-14-10

^aOrigins of samples: BAL, bronchoalveolar lavage fluid; TS, tracheal sputum/aspiration.

^bDiagnostic delay (from initial sampling to microbiological diagnosis/from sample arrival at reference laboratory to final microbiological diagnosis). ND, not determined (sampling date not available).

^cPosaconazole (POS), voriconazole (VRC), and itraconazole (ITR) MIC determinations obtained using EUCAST microbroth dilution; amphotericin B (AMB) and caspofungin (CAS) MICs determined using E-tests.

^dNA, not appropriate (R. pusillus is intrinsically resistant, and thus caspofungin susceptibility testing was not deemed relevant). STRAf and CYP51A genotyping is applicable only to A. fumigatus.

^eWT, wild type.

Environmental studies.

In the first study, a total of 58 A. fumigatus isolates were recovered from 69 soil samples from Tivoli Gardens, potted plants at the Rigshospitalet, and from organic and inorganic fields. None were found to be azole resistant. In the second study, a total of 55 A. fumigatus isolates were recovered from 170 samples, of which 130 derived from fields where fungal pesticides were used and 40 from organically grown fields. But again none were found azole resistant. The detailed results are summarized and compared with previous data (2009) (10) in Table 3.

TABLE 3.

Soil samples and A. fumigatus findings from the two environmental surveys (2010 and 2013) compared with the data from a previously published study performed in 2009

Isolate type	No. (%) of isolates found at the indicated location in the indicated yr
	Tivoli		Hospital		Conventional farms		Organic farms
	2009a	2010	Outdoor 2009a	Indoor 2010	2010	2013	2010	2013
Soil samples	23 (100)	17 (100)	27 (100)	25 (100)	12 (100)	130 (100)	15 (100)	40 (100)
A. fumigatus	21 (91)	15 (88)	17 (63)	19 (76)	11 (92)	45 (35)	13 (87)	10 (25)
Azole resistant	3 (13)b	0	1 (4)c	0	0	0	0	0

^aData compiled from Mortensen et al. (10).

^bOr 14% of the A. fumigatus isolates.

^cOr 6% of the A. fumigatus isolates.

DISCUSSION

These clinical cases represent the first detection of A. fumigatus TR₄₆/Y121F/T289A and the first examples of A. fumigatus TR₃₄/L98H infection in azole-naive patients in Denmark. All four patients were functionally immunosuppressed either due to treatment (steroids, chemotherapy, and lung transplantation), congenital immunodeficiency, or invasive cancer and thus had recognized risk factors for pulmonary aspergillosis. Aspergillus is ubiquitous in the environment and exposure is unavoidable. Invasive aspergillosis (IA) with a resistant isolate harboring the TR₃₄/L98H mutation has been shown to be associated with a dismal outcome and a mortality of 88% (compared to ∼30 to 50% for aspergillosis with a wild-type susceptibility profile) (33). A similarly poor prognosis was reported in a small clinical series of probable or proven IA caused by isolates harboring the TR₄₆/Y121F/T289A alteration, as the five patients initially treated with voriconazole all failed and four died within 12 weeks (20). Among our patients, patient 3 had some habitual pulmonary symptoms related to her disseminated cancer; however, she had no obvious clinical signs of invasive aspergillosis and may thus have been only colonized. This contrasts with the situation in the three other patients, in whom several serial cultures gave rise to growth of molds, serum, or bronchoalveolar lavage (BAL) fluid that were found to be galactomannan (GM) positive. These patients displayed lung infection with no alternative agents but mold(s) found, and two of three died within 6 weeks of diagnosis. The last patient died 14 weeks after the initial diagnosis of the TR₄₆/Y121F/T289A isolate, with a persistent infection.

It appears most likely that the first and third patients had acquired their azole-resistant A. fumigatus isolates from the environment prior to admission. Thus, patient 3 had received no mold-active azole when the TR₃₄/L98H isolate was cultured, and patient 1 was mold culture positive on the day of her admission, with subsequent full identification of A. fumigatus TR₃₄/L98H from a sample obtained on day 7 (after only 2 days of voriconazole treatment). The pictures regarding the second and fourth patients were more complex, as they initially presented with wild-type azole-susceptible A. fumigatus and were treated with voriconazole (systemically or by inhalation) before presenting weeks later with a second, pan-azole-resistant strain of A. fumigatus (TR₃₄/L98H and TR₄₆/Y121F/T289A, respectively). Had susceptibility testing been performed only for the first isolate, the coinfection would have remained undetected.

STRAf typing failed to detect any relationship between the isolates either across the patients or when comparing susceptible and resistant strains from individual patients. Moreover, the patients were admitted at four different departments in three different hospitals in separate geographical locations in Denmark, at different time periods and without any known previous contact. The patient harboring the TR₄₆/Y121F/T289A strain had no foreign travel history for the past 4 years. This suggests that the emergence of resistance is geographically widespread in Denmark.

Following the initial detection of 8% TR₃₄/L98H isolates in soil samples during the study performed in June to August 2009 and due to a growing concern that the TR₃₄/L98H and TR₄₆/Y121F/T289A alterations might be emerging in Denmark, soil samples were collected in March to April 2010 and examined for the presence of azole-resistant A. fumigatus. As the overall sample size was low and the soil partly frozen at the time of sampling, we repeated the soil sampling in the autumn in 2013 after plowing. However, on both occasions we failed to detect any resistant A. fumigatus isolates. Several environmental studies have been performed over the past years in both Europe (10, 13, 20, 24, 34) and India (19, 25), all demonstrating the presence of one or both of the two resistance “environmental” genotypes (see Table S1 in the supplemental material). Whereas the failure to demonstrate resistant isolates in the spring of 2010 may have been due to that period being an unusually cold spring following a winter with several months with freezing temperatures, it is less clear why no resistance was found in the recent autumn 2013 soil samples. One hypothesis is that the environmental resistance in Denmark is actually declining. However, it appears less likely that the patients should have been colonized with environmental isolates years ago and only now developed clinical infection, and even more so as other European countries are reporting increasing rates of resistance. Another hypothesis is that the amount of resistant A. fumigatus in soil is subject to seasonal variation. In the vast majority of previous studies demonstrating azole-resistant strains directly in soil samples, sampling was conducted during the summer months (Europe) or in a hot climate (India). Moreover, in a recently published Italian survey (24), five out of six isolates harboring the TR₃₄/L98H alteration were detected in soil samples collected during the summer months and only a single isolate was detected in a soil sample from an apple orchard collected during October and from a field type unlikely to be plowed. Only a single European study performed in an area where environmental resistance mechanisms in Aspergillus fumigatus are highly endemic demonstrated several resistant isolates in samples collected during October 2010 and March 2011 (P. Verweij, personal communication). However, in this survey air sampling of great volumes (14,000 liters/site) was employed, giving rise to a greater number of Aspergillus isolates per sampled site. It thus appears likely that soil sampling in temperate climates should preferentially be conducted during the summer months rather than the colder months preceding crop maturation or following harvesting and plowing of the fields in order not to underestimate the actual incidence of environmental resistance.

The apparent emergence of azole resistance in the clinical setting calls into question our current strategies using azoles for prophylaxis and targeted treatment of invasive aspergillosis (35, 36). The outcome for patients with azole-resistant IPA is dismal following standard treatment (20, 33). On the other hand, multiple papers describing postmarketing experience have strongly suggested voriconazole is superior and therefore the preferred agent for infections involving susceptible isolates (37–42). This again highlights the need not only for rapid identification but also for real-time susceptibility testing of clinical isolates of A. fumigatus, regardless of prior azole treatment (12, 13, 19). Ideally, susceptibility testing should be performed not only for every species of mold but by sampling several colonies (when present) of A. fumigatus from each clinical sample, as patients may be coinfected with wild-type and resistant isolates (43). Although the optimal number of colonies that should be tested is not known, a pragmatic recommendation would be to sample 5 colonies (when present), in agreement with what is recommended for yeast testing (44). Also, direct detection of resistant genotypes by molecular methods would be a potentially valuable tool in the future, as culturing of patient specimens is well known to have low sensitivity.