Introduction

Talaromyces marneffei (formerly Penicillium marneffei) is a life-threatening thermal dimorphic fungus endemic to tropical Southeast Asian regions and is the etiological agent of talaromycosis [1]. While primarily affecting people living with HIV, T. marneffei case reports have recently been described in HIV-negative, immunocompromised individuals [2,3,4]. T. marneffei has become one of the three most common opportunistic pathogens in Southeast Asia [5] and is the most frequent blood-borne fungus in individuals diagnosed with AIDS living in Southern China [6]. Talaromycosis brings about a substantial public health threat in endemic areas and is now beginning to affect other regions [7].

Talaromycosis is recommended to be treated with amphotericin B for 2 weeks, followed by itraconazole for 10 weeks and then as continuous secondary prophylaxis until CD4+ counts remain above 100 cells/mm3 for 6 months [8]. However, therapy with amphotericin B and itraconazole can be limited due to drug toxicity and cost. In these cases voriconazole may be a suitable alternative [9]. In addition, fluconazole is an alternative treatment agent in many economically developing countries, including China [10]. Recently, echinocandins have been introduced as first-line drugs to treat invasive candidiasis [11], but whether this class of antifungal agents possesses activity against T. marneffei has not been fully investigated. Posaconazole, a newly approved triazole antifungal drug by FDA for treating invasive infections with Candida, Mucor, and Aspergillus in severely immunocompromised patients [12], may also be of significance for possessing activity against T. marneffei. Despite there are some other influence factors, early initiation of therapy and selection of empiric antifungal agents according to local susceptibility are greatly important for improving clinical outcomes in systemic fungal infections. In order to investigate the susceptibility profile of T. marneffei strains against eight agents, the Sensititre YeastOne™ method, a ready-to-use commercial kit, is adopted and 189 T. marneffei clinical strains isolated from HIV-infected patients from January 2013 to December 2016 in Guangdong is determined through their MICs.

Materials and methods

T. marneffei strains

Out of the 189 strains, 41, 39, 38, and 71 strains were collected during 2013, 2014, 2015, and 2016, respectively. The collected specimens included were the blood, bone marrow, and bronchoalveolar lavage fluid. All isolates were identified by conventional culture and microscopic characteristics before being stored at − 80 °C. Internal transcribed spacer (ITS) were amplified with PCR and detected with direct DNA sequencing to determine the species [13].

Antifungal susceptibility testing

Sensititre YeastOne ™ YO10 assay (Thermo Scientific, Cleveland, OH, USA) was conducted to test antifungal susceptibility according to manufacturer’s instructions. The disposable Sensititre panels, which are 96-well plates incorporated with alamarBlue for colorimetric determination, contain serial twofold dilutions of the following dried antifungal agents: 0.015 to 8 μg/ml of anidulafungin, 0.008 to 8 μg/ml of micafungin and caspofungin, 0.008 to 8 μg/ml of posaconazole and voriconazole, 0.015 to 16 μg/ml of itraconazole, 0.12 to 256 μg/ml of fluconazole, and 0.12 to 8 μg/ml of amphotericin B.

Prior to testing, strains were sub-cultured onto Sabouraud dextrose agar and incubated at 37 °C for 4 days. Then, a 0.5 McFarland yeast suspension of each strain was prepared using the spectrophotometric method, with final concentrations of 1.5–8 × 103 CFU/ml. One hundred microliter of the yeast suspension was dispensed into each well of the dried panels for rehydration. Finally, the panels were enclosed with adhesive seals and incubated at 35 °C in a non-CO2 atmosphere. Quality control strains of Candida parapsilosis ATCC® 22019 were included throughout the experiments. MIC endpoints were read visually when the color had obviously changed from blue (with growth) to red (without growth) after incubation.

Results

Triazoles were found to have low MICs against the yeast phase of most isolated T. marneffei in vitro, most strains (> 88%) for posaconazole, itraconazole, and voriconazole with MICs ≤ 0.015 μg/ml, and 80.4% strains were found to have MICs≤4 μg/ml for fluconazole. Posaconazole MIC values ranged from ≤ 0.008 to 0.06 μg/ml, the lowest MIC (geometric mean value, 0.013 μg/ml) of all tested azoles, followed by voriconazole and itraconazole. Fluconazole had the highest MIC (4.074 μg/ml) of all tested azoles, and most remarkable is that two (1%) strains were found to have MICs of 16 and 32 μg/ml, respectively. Moreover, the results revealed that the antifungal activity of triazoles was superior to echinocandins against the yeast form of T. marneffei. Echinocandins had comparatively lower or almost no activity against T. marneffei in vitro. The MICs against all strains were > 8 μg/ml for micafungin and ≥ 2 μg/ml for caspofungin and anidulafungin (not shown). Besides, all strains had MICs ≤ 1 μg/ml for amphotericin B. An overview is shown in Table 1.

Table 1 In vitro MICs of 189 T.marneffei strains to five antifungals as determined by YeastOne method
Full size table

Discussion

Talaromycosis, a disease with multiple clinical manifestations and high rates of morbidity and mortality, has increasing incidence among HIV-infected patients [14]. However, data describing antifungal susceptibility of T. marneffei is limited [15, 16], and no corresponding data has been previously reported from Guangdong Province. In this study, 189 clinical isolations from patients diagnosed with talaromycosis in the referral hospital of Guangdong Province were included.

Sensititre YeastOne TM YO10 assay has been evidenced to be with high agreement in testing Candida compared with broth dilution method and Etest [17]. Our protocols of Sensititre method were modified from those standardized by CLSI for the determination of antifungal susceptibilities in Candida species [18], which represent the first time this method has been used to evaluate antifungal susceptibilities of T. marneffei. Based on the results observed in our study, we can believe this method is suitable to detect the susceptibility of the yeast phase of T. marneffei. Past research had shown echinocandins activity against T. marneffei differs between the yeast and mycelial phases [16, 19]. Anidulafungin exhibits potential activity (MIC ≤ 2 μg/ml) against the mycelial phase of T. marneffei [20], while MICs against the yeast phase of 57 T. marneffei strains are over 2 μg/ml [16]. Our findings support previous research demonstrating the limited antifungal activity of anidulafungin against the yeast phase of T. marneffei, the parasitic form that can be detected under regular clinical laboratory conditions [21]. Our results also suggest caspofungin and micafungin have similar activity as anidulafungin. T. marneffei may well be resistant to the entire echinocandin class while in its yeast phase; however, the relationship between in vitro susceptibility of the yeast phase and clinical outcomes should be further investigated. In contrast to echinocandins, the yeast forms of all T. marneffei strains were inhibited by itraconazole and voriconazole. Posaconazole, as another triazole, had been previously used for clinical treatment of other dimorphic fungi infections, including H. capsulatum and Coccidioides [22, 23]. And it was also observed to have low MICs against the T. marneffei strains in our study, which was close to the results released by Lau [16]. Those may collectively indicate posaconazole is a promising agent for treatment of talaromycosis. Furthermore, amphotericin B also possessed good activity against included strains with MICs in the susceptible range. The vast majority of T. marneffei strains were sensitive to fluconazole with MICs ≤ 8 μg/ml, but it is noteworthy that four clinical isolates were found to not be within the susceptible MIC range. Additionally, the surprising is the MICs detected by different laboratory are various greatly, we postulate this discrepancy may be due to the absence of standardized protocols for testing.

In brief, our data support the use of amphotericin B, itraconazole, voriconazole, and posaconazole in clinical management of talaromycosis, meanwhile alarm for the potential resistance to fluconazole. However, since clinical breakpoints or epidemiological cutoff values of antifungal agents against T. marneffei have not been developed, we only discriminated them as suspected resistant strains according to the reference of Candida. Aiming to accurately distinguish resistant and wild-type T. marneffei strains, further studies in multiple settings will be required.