Surveillance for azoles resistance in Aspergillus spp. highlights a high number of amphotericin B-resistant isolates

Abstract

Aspergillus spp. are the most common invasive mould infection and are responsible for high mortality. Aspergillus fumigatus is currently of interest because resistance to azole antifungals has emerged. The Campinas University Hospital (HC-UNICAMP) receives high-risk patients susceptible to opportunistic infections but there have been no reports of resistant A. fumigatus. This study aimed to assess the susceptibility profile of Aspergillus isolates, specifically looking for azole resistance. ITS and β-tubulin DNA sequencing was performed on 228 sequential clinical isolates. Broth microdilution susceptibility testing was performed for all isolates. A. fumigatus represented 74% of the isolates followed by Aspergillus flavus (12%). Nine A. fumigatus isolates from 9 different patients showed high MIC values to at least 1 azole, but cyp51A polymorphisms were detected in only 6 isolates and none correlated with known resistance mutations. The most troubling observation was that the minimum inhibitory concentration for amphotericin B was elevated (≥2 mg L⁻¹) in 87% of patients with A. flavus isolates and 43% with Aspergillus fumigatus isolates. Given that amphotericin B is used to treat azole-resistant infections, these data highlight the need for continuous surveillance in Aspergillus for all antifungal resistance to implement correct treatment strategies for the management of these pathogens.

INTRODUCTION

Aspergillus species are among the most abundant fungi worldwide.¹ These ubiquitous opportunistic pathogens are responsible for one of the highest mortality rates among patients with fungal infections. Over 30 million people worldwide are estimated to be at risk of invasive aspergillosis (IA) each year,² especially individuals with impaired immune function, those with low numbers of neutrophils, solid organ transplant recipients and patients on immunosuppressive therapies.^2-4 Over 200 000 patients develop IA annually,² with A. fumigatus the most common species isolated in patients, followed by Aspergillus flavus.⁵

Classical morphological species identification within each of the Aspergillus species complexes is difficult due to a lack of morphologically distinguishing features of these microorganisms.^{6, 7} There are similar species with differing antifungal susceptibilities so correct species identification is crucial; reports of cryptic and rare species causing aspergillosis in humans are increasing.^{8, 9}

The triazole antifungal drugs, itraconazole (ITZ), voriconazole (VRZ) and posaconazole (PSZ), are recommended as first line drugs in the treatment and prophylaxis of aspergillosis.^{10, 11} Despite the high mortality of IA, patient survival rates have improved due to better diagnostic tools and treatment regimens. Nevertheless, azole-resistant isolates A. fumigatus have been increasingly identified in the last decade. Data on azole resistance patterns include patients with non-invasive aspergillosis who were treated with long-term azole therapy and developed acquired resistance after 1-30 months of treatment. These isolates may exhibit a multi-azole-resistant phenotype and the underlying resistance mechanisms commonly involve point mutations in cyp51A gene, indicating that the fungus can adapt to azole compounds in patients exposed to them.¹² Another pattern of resistance occurs in patients with acute aspergillosis with no previous exposure to azole drugs. These patients most likely acquired resistant isolates from environmental sources. It is believed that these resistant isolates emerged due to the widespread use of agricultural azole antifungals.^13-16

Azole resistance in A. fumigatus is usually due to mutations in hotspot regions of the 14α sterol demethylase gene (cyp51A) which may or may not include tandem repeats in the promoter region of the cyp51A gene.^{17, 18} To a much lesser extent, non-cyp51A-mediated mechanisms of resistance, such as increased efflux, have been recognised as a mechanism for azole resistance in A. fumigatus.^19-21

Despite its high toxicity, amphotericin B (AMB) is still widely used and is considered the gold standard treatment of some invasive fungal infections.²² Polyenes such as AMB interact with the cell membrane ergosterol, producing pores that increase the cell permeability causing loss of electrolytes, leading to lysis and cell death.²³ Resistance of Aspergillus to AMB is not well understood. However, despite the difficulty of associating in vitro results with clinical outcomes, high MIC values of AMB have been associated with poor outcomes.²⁴ Two mechanisms seem to be involved in AMB resistance, decreased ergosterol concentration in the cell membrane due to mutations in the synthesis pathway and increased production of catalases, which protect the cell from oxidative stress caused by the drug.²²

At Campinas University Clinical Hospital, Aspergillus susceptibility testing is not routinely performed for clinical practice and, until now, there have been no reports of azole-resistant A. fumigatus. However, monitoring for the possible emergence of resistant strains is necessary, especially with reports coming from a bordering country,²⁵ as well as the recent introduction of voriconazole to replace fluconazole, to which moulds are already resistant, for prophylaxis of fungal infections. The aim of this study was to perform molecular identification and assess the antifungal susceptibility profile of clinical Aspergillus spp. isolates recovered in routine tests during a period of 16 years and to look for cyp51A gene mutations in phenotypically azole-resistant A. fumigatus isolates.

MATERIALS AND METHODS

Fungal isolates

Two hundred and twenty-eight Aspergillus spp. clinical isolates were collected from 91 patients treated at the Campinas University Clinical Hospital, Campinas, Sao Paulo, Brazil during 1998-2014. The isolates were cultivated from sputum (n = 150); bronchial lavage (n = 32); ocular secretion (n = 4); blood (n = 2); cornea (n = 2); urine (n = 2); lung biopsy (n = 2); pleural liquid (n = 2); endotracheal secretion (n = 2) and others (n = 30). Identification had been previously performed by conventional morphological methods. After that, the isolates were stored in distilled water at room temperature.²⁶ For this study, all isolates were cultured on Sabouraud dextrose agar (SDA; Difco, Sparks, MD, USA).

Microbiological identification

Macromorphology and micromorphology of each species complex was observed after growth on SDA and potato dextrose agar (PDA; Difco).²⁷

DNA extraction

Genomic DNA was extracted from isolates grown 48-72 hours on SDA plates using a QIAmp^® DNA mini Kit (Qiagen Sciences, Germantown, MD, USA), according to the manufacturer's instructions.

Amplification of ITS and beta-tubulin

Primers ITS4/5 and β-tubulin-2A/2B were used to amplify DNA from all Aspergillus isolates. Comparative DNA analyses of the ITS (ITS4/5) sequence was performed to achieve Aspergillus species complex-level identification followed by beta-tubulin (β-tubulin 2A/B) for species confirmation. The resultant nucleotide sequences were edited on Geneious^® 8.1 (Biomatters Ltd 2015, Newark, NJ, USA) and directly compared with external databases.

Broth microdilution test

Minimum inhibitory concentration (MIC) and echinocandin minimum effective concentration (MEC) were determined following the Clinical and Laboratory Standards Institute M38-A2 guidelines.²⁸ In vitro antifungal susceptibility testing was performed using preprepared dry plates (Eiken Chemical Co., Tokyo, Japan). The antifungal agents analysed were AMB (range 0.03-16 mg L⁻¹), ITZ (range 0.015-8 mg L⁻¹), VRZ (range 0.015-8 mg L⁻¹), micafungin (MCF) (range 0.015-16 mg L⁻¹) and caspofungin (CPF) (range 0.015-16 mg L⁻¹). CPF was not available on the plate and was prepared separately. CPF (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in Water and then diluted in RPMI 1640 (Sigma-Aldrich).

The MIC scores were obtained visually with the inoculums of 1.0 to 5.0 × 10⁴ CFU/mL after incubation at 35°C for 48 hours. MICs for AMB, VRZ and ITZ were defined as the lowest concentration causing 100% inhibition of growth, compared to the drug-free growth control. The MECs for MCF and CPF were read visually with the same inoculum concentration after incubation at 35°C for 24 hours. Quality control was ensured by including Candida parapsilosis ATCC 22019, Candida krusei ATCC 6258, A. flavus ATCC 204304 strains each time that a set of isolates was tested.

Epidemiological cutoff values (ECVs)

The ECVs for each antifungal were evaluated as described in recent studies.^29-31

Detection of cyp51A mutations

Aspergillus fumigatus isolates with azole VRZ MIC ≥4 mg L⁻¹ and ITZ MIC ≥2 mg L⁻¹ values and all sequential isolates from the patients who present 1 or more isolates with these MICs (a total of 24 isolates) were analysed. DNA was extracted from 48 hours fungal cultures as described above. Oligonucleotides used for amplification and sequence analysis of cyp51A were AF1P1 (F/R), AF2P1 (F/R), AF3P1 (F/R) e AF4P1 (F/R).^{32, 33}

Negative controls with no DNA were included in each amplification reaction. Sequences were aligned with an azole-susceptible strain sequence (GenBank accession number AF338659) and azole-susceptible clinical isolates.

RESULTS

Isolate identification

Based on morphological identification and DNA sequencing of ITS 4/5 and beta-tubulin (β-tubulin 2A/B) regions it was possible to identify 11 different species: A. fumigatus (n = 169); A. flavus (n = 27); Aspergillus parasiticus (n = 3); Aspergillus terreus (n = 11); Aspergillus niger (n = 8); Aspergillus tubingensis (n = 2); Aspergillus clavatus (n = 2); Aspergillus nidulans (n = 2); Aspergillus sydowii (n = 2); Aspergillus ochraceus (n = 1); Neosartorya hiratsukae (n = 1).

Broth microdilution test

MIC/MEC ranges, MIC/MEC50 and MIC/MEC90 of the antifungal drugs AMB; ITZ, VRZ, MCF and CPF evaluated against clinical isolates of Aspergillus and reference strains are shown in Table 1. The ECVs for AMB, ITZ and VRZ are shown in Table 2. Among the 168 A. fumigatus isolates, 3 (1.7%) had MIC values for ITZ above the ECV and 40 (24%) had MIC values for VRZ above the ECV. For AMB, 46 isolates from 32 patients had MICs ≥2 mg L⁻¹. The ECV for AMB was 1 mg L⁻¹ to A. fumigatus species and the, MIC/MEC50 and MIC/MEC90 were 1.0 and 2.0 mg L^−1, respectively.

Table 1.

MICs and MECs (mg L⁻¹) of 5 antifungal agents tested against Aspergillus species and reference strains

Aspergillus species (n)	MIC/MEC ranges (mg L−1)
	Parameter	AMB	ITZ	VRZ	Parameter	MCF 24 h	CPF 24 h
A. fumigatus (168)	MIC range	0.25-8	0.25-4	0.25-8	MEC range	≤0.015-0.03	0.06-0.5
	MIC50	1	1	1	MEC50	≤0.015	0.25
	MIC90	2	1	2	MEC90	0.03	0.25
A. flavus (27)	MIC range	1-4	0.125-1	0.5-2	MIC range	≤0.015-0.06	0.125-0.5
	MIC50	2	0.5	1	MEC50	0.03	0.5
	MIC90	4	0.5	1	MEC90	0.06	0.5
A. parasiticus (3)	MIC range	2 to >16	0.25-0.5	1-2	MEC range	0.03	0.25-0.5
	MIC50	8	0.5	1	MEC50	0.03	0.25
	MIC90	>16	0.5	2	MEC90	0.03	0.5
A. terreus (11)	MIC range	2 to >16	0.25-2	0.5-2	MEC range	≤0.015	0.25
	MIC50	2	1	2	MEC50	≤0.015	0.25
	MIC90	>16	1	4	MEC90	≤0.015	0.25
A. niger (8)	MIC range	0.5-1	1-4	1-2	MEC range	≤0.015	0.25
	MIC50	0.5	1	2	MEC50	≤0.015	0.25
	MIC90	1	2	2	MEC90	≤0.015	0.25
A. tubingensis (2)	MIC range	0.5-1	0.5-2	1-4	MEC range	≤0.015	0.25
A. clavatus (n = 2)	MIC range	0.25-1.0	1.0-2.0	2.0	MEC range	≤0.015-0.03	0.5-1.0
A. nidulans (n = 2)	MIC range	2.0-4.0	0.5-1.0	0.25-1.0	MEC range	≤0.015-0.03	0.125-0.5
A. sydowii (n = 2)	MIC range	1.0-4.0	1.0-2.0	1.0-2.0	MEC range	≤0.015	0.06-0.25
A. ochraceus (n = 1)	MIC range	4.0	1.0	0.5	MEC range	≤0.015	0.5
Neosartorya hiratsukae (n = 1)	MIC range	1.0	1.0	1.0	MEC range	≤0.015	0.25
Reference strains
Candida parapsilosis ATCC 22019	MIC range	1.0	0.125	0.03	MIC range	0.5	1.0
Candida krusei ATCC 6258	MIC range	1.0	0.125	0.25	MIC range	0.06	1.0
Aspergillus flavus ATCC 204304	MIC range	2.0	0.5	0.5	MEC range	≤0.015	0.25

• AMB, Amphotericin B; CPF, caspofungin; ITZ, itraconazole; MCF, micafungin; MEC, Minimal effective concentration; MEC50/MEC90, Minimal effective concentration capable of 50 and 90% of isolates inhibition; MIC, Minimum inhibitory concentration; MIC50/MIC90, Minimum inhibitory concentration capable of 50 and 90% of isolates inhibition; VRZ, voriconazole.

Table 2.

Distribution of MICs (mg L⁻¹) and ECVs values of 3 antifungal agents tested against main Aspergillus isolates

Aspergillus species	Antifungal agent	MICs distribution (mg L−1)
		0.125	0.25	0.5	1	2	4	8	16
A. fumigatus (168)	AMB		1	15	106a	42	3	1
	ITZ		1	79	86a	2	1
	VRZ		1	38	90a	31	7	2
A. flavus (27)	AMB				5	16a	6
	ITZ	2	6	17a	2
	VRZ			13a	13	1
A. terreus (11)	AMB					8a	1		2
	ITZ			2	9a
	VRZ			1	1	5a	4
A. niger (8)	AMB			6a	2
	ITZ				5a	2	1
	VRZ				3	5a

• AMB, amphotericin B; ITZ, itraconazole; MIC, Minimum inhibitory concentration; VRZ, voriconazole.

• ^a Epidemiological cutoff values (ECVs) for each antifungal in this sampling.

Excluding the A. terreus isolates, which are naturally resistant to AMB, 75 (33%) of 227 Aspergillus spp. isolates had MIC ≥2 mg L⁻¹ to this drug. For A. flavus isolates, 81% showed MIC ≥2 mg L⁻¹ and, for A. fumigatus isolates, 27% of 168 isolates had MIC ≥2 mg L⁻¹. This equates to 33 of the 76 patients (43%) with A. fumigatus had an isolate with a MIC ≥2 mg L⁻¹ for AMB.

cyp51A mutations

Nine isolates of A. fumigatus from 9 patients had high MICs for VRZ or ITZ (7 for VRZ and 2 for both ITZ and VRZ). One of the isolates had high MIC values for VRZ and AMB. A total of 24 A. fumigatus isolates, including all sequential isolates from patients who had azole-resistant isolates, were analysed for cyp51A mutations. The following amino acid changes were detected: F46Y, M172V, N248T, D255E and E427K. These alterations were detected in 6 isolates (2 isolates with high azole MICs and 4 susceptible isolates from the same patients that showed high azole MICs). Seven isolates with high MIC values (VRZ ≥4 mg L⁻¹ and ITZ ≥2 mg L⁻¹) did not carry a cyp51A mutation.

DISCUSSION

In this study, 228 clinical isolates of Aspergillus from a hospital in Campinas, Brazil, were identified to species and tested for antifungal susceptibility. Aspergillus fumigatus and A. flavus were predominant, representing 74% and 12% of the clinical isolates respectively. The A. flavus isolates from this study showed high MICs to AMB, with 81% of the isolates having MIC values ≥2 mg L⁻¹. MICs 2 mg L⁻¹ have been associated with treatment failure among patients with invasive aspergillosis.²⁴ Two isolates of the closely related species A. parasiticus also had high MICs to AMB of 8 and 16 mg L⁻¹. The susceptibility profile of section Flavi showing high values of MIC to AMB have been described in literature.^{34, 35}

As expected because of its natural resistance to polyenes, isolates of A. terreus showed high MICs to AMB. In addition, reduced susceptibility to VRZ was observed in 4 isolates. This susceptibility profile of A. terreus has been described in vitro and in vivo and has been associated with poor outcomes.^{36, 37}

Only 1 of the 10 isolates from the A. niger species complex, and a A. tubingensis isolate, had a high MIC to ITZ (4 mg L⁻¹). No isolates within this complex showed resistance to AMB, what was also previously described in literature.³⁸

Sequencing of the cyp51A gene was performed on a limited number of isolates. Numerous amino acid substitutions were detected, including F46Y, M172V, N248T, D255E and E427K. While these point mutations are frequently reported in A. fumigatus isolates from patients undergoing long-term azole therapy,^39-41 they have been detected in susceptible isolates and most likely do not contribute to resistance. Seven isolates showing high MIC values did not carry a cyp51A mutation. These results together indicate the involvement of other mechanisms like overexpression of cyp51B,^{42, 43} efflux pumps, drug efflux transporters, drug degradation or overexpression of the ABC transporter.²¹

The unexpected result from this study was that 27% of A. fumigatus isolates had AMB MICs ≥2 mg L⁻¹. Study published by Guinea and colleagues showed 1.5% of A. fumigatus with MIC ≥4 mg L⁻¹ to this drug ⁴⁴ and the same was observed in 2.4% of the isolates of this study. As reports of A. fumigatus isolates resistant to azoles increase, antifungals such as AMB may be required for the treatment of invasive aspergillosis. Since this agent is also ineffective against some isolates, leaving few or no therapeutic options, new drug development and drug combination strategies should to be explored.⁴⁵

Previous studies investigating the in vitro-in vivo correlation of AMB resistance in A. fumigatus resulted in inconsistent data.^{46, 47} However, in vivo studies have shown that a bigger efficacy of AMB was obtained when the maximum concentration of the drug serum exceeded 2.4 times the values of in vitro MIC,⁴⁸ which becomes infeasible when the microorganisms are resistant.

Micafungin and caspofungin were very effective against most species in this study. Echinocandins are effective in salvage therapy (alone or in combination), but they are not recommended for routine use as monotherapy for the primary treatment of invasive aspergillosis.⁴⁹ So far, little is known about Aspergillus resistance to the echinocandins class. The use of these compounds is restricted to cases involving combinations of antifungal agents or empiric treatment.²²

One of the caveats of this study is that the number of isolates did not correspond to the number of patients; some patients had more than 1 isolate and this duplication may have artificially raised the number of resistant isolates. However, with the consideration that, (i) drug resistance can develop under antifungal pressure over time and, (ii) that patients may carry unrelated isolates at any given time, it was felt that it was important to capture these isolates.

Despite the original aim of the study to detect azole resistance, the most significant finding is the high number of A. flavus and A. fumigatus isolates showing high MICs to AMB. Coupled with the A. terreus isolates, which are intrinsically resistant, this compromises an important therapeutic option for IA. Currently, knowledge about resistance mechanisms against amphotericin B in Aspergillus species is limited and this study has highlighted an area for future study. These data also highlight the need for continuous resistance surveillance in Aspergillus for correct management of these pathogens.