Correlation of Trichosporon asahii Genotypes with Anatomical Sites and Antifungal Susceptibility Profiles: Data Analyses from 284 Isolates Collected in the Last 22 Years across 24 Medical Centers

Trichosporon asahii is an opportunistic fungal pathogen that can cause severe infections with high mortality rates. Azole derivatives are the best-targeted therapy for T. asahii invasive infections, but azole-resistant isolates have been reported.

ABSTRACT

Trichosporon asahii is an opportunistic fungal pathogen that can cause severe infections with high mortality rates. Azole derivatives are the best-targeted therapy for T. asahii invasive infections, but azole-resistant isolates have been reported. To investigate peculiarities in the antifungal susceptibility profile (ASP) of T. asahii clinical isolates, we analyzed the genotype distribution, isolation sources, and ASP of 284 strains collected from 1997 to 2019 in different Brazilian medical centers. Species identification and genotype characterization were performed by analysis of the intergenic spacer (IGS1) region of the ribosomal DNA (rDNA). Antifungal susceptibility testing (AST) for amphotericin B and azoles was with the CLSI M27, 4th edition, microdilution broth method. Trends in the ASP of Brazilian T. asahii isolates were investigated using epidemiological cutoff values. Five different genotypes were found among the 284 isolates tested (G1, 76%; G3, 10%; G4, 3%; G5, 7%; and G7, 4%). The isolates were collected mainly from urine (55%) and blood/catheter tip samples (25%) where G1 was the most frequent genotype found (P < 0.05). The G7 isolates exhibited the highest MIC₉₀ values for azoles compared to those for the other genotypes (P < 0.05). Genotype 7 isolates also contributed to the increasing rates of voriconazole non-wild-type isolates found in recent years (P = 0.02). No significant differences were found among the AST results generated by isolates cultured from different anatomical sites. Monitoring T. asahii genotype distributions and antifungal susceptibility profiles is warranted to prevent the spread of azole-resistant isolates.

INTRODUCTION

First described by Akagi in 1929 (1), Trichosporon asahii is a basidiomycete yeast-like fungi that is considered an emergent opportunistic fungal pathogen capable of causing life-threatening invasive infections, mostly in neutropenic cancer patients, patients submitted for organ transplantation, or critically ill patients exposed to antibiotics, corticosteroids, or invasive medical procedures (2). The prognosis of patients with invasive trichosporonosis is a concern given that crude mortality rates as high as 42% to 87% have been reported by different series (2, 3).

Due to the lack of clinical trials, recommendations for the clinical management of invasive trichosporonosis rely mostly on data generated from animal models, in vitro susceptibility tests, and reports of small clinical series (3–6). In this scenario, most authors suggest that triazoles are the best therapy for invasive infections caused by Trichosporon spp., despite some reports illustrating that some species and strain variability may occur in their susceptibilities to fluconazole and voriconazole (3, 7, 8).

Nowadays, a total of 15 T. asahii genotypes has been characterized by different research groups based on the genetic diversity described in their intergenic spacer 1 (IGS1) ribosomal DNA (rDNA) sequences (9). Notably, there is a lack of robust data illustrating their distribution in different anatomical sites of patients colonized or infected by T. asahii. In addition, it remains controversial whether the genetic diversity described for T. asahii isolates has any impact in terms of virulence or antifungal susceptibility to different antifungal drugs (5, 6, 9).

The present study describes the prevalence rates of T. asahii genotypes found among 284 clinical isolates cultured from different patients as well as their in vitro susceptibility to four antifungal drugs.

RESULTS

Prevalence rates of T. asahii genotypes cultured from different anatomical sites.

Only five clinical centers were able to provide data on the annual incidence of T. asahii-positive cultures collected between 2010 and 2019. The pooled data showed that the incidence of cultures positive for T. asahii remained stable over the years, ranging from 0.5 to 2 isolates per 10,000 admissions each year. Clinical isolates were cultured from urine (n = 157 [55%]), blood samples/central venous catheter tips (n = 71 [25%]), respiratory tract samples (n = 16 [6%]), integumentary samples, including hair, nail, and skin samples (n = 11 [4%]), other sterile anatomical sites (n = 11 [4%]), and other sources (n = 18 [6%]).

The 284 clinical isolates were successfully identified as T. asahii by sequencing of the IGS1 rDNA region, with 100% of query coverage and percent identity and E value of 0.0. As illustrated in Fig. 1 and Fig. S1 in the supplemental material, the median-joining network analysis showed five different genotypes (G1, n = 216 [76%]; G3, n = 28 [10%]; G4, n = 8 [3%]; G5, n = 21 [7%]; G7, n = 11 [4%]), including representatives of G7 that were never previously reported among Brazilian isolates.

FIG 1.

Median-joining genotypes network of T. asahii based on IGS1 rDNA sequences (A) and their distribution in anatomic sites (B). Circumference sizes are proportional to the frequencies of genotypes detected. Blue circles represent genotypes found in our study (G1, G3, G4, G5, and G7). Black circles represent genotypes that could not be grouped with any isolate from our collection. Red dots (median vectors [mv]) are hypothetical missing intermediates. CVC, catheter tips; Integ. samples, integumentary samples; S.S.N.B., sterile sites, nonblood.

Three to five distinct genotypes of T. asahii were found in each anatomic site. Genotype 1 was present in all evaluated sites and represented 76% of all typed isolates. In urine cultures, T. asahii G1 was responsible for 138 of 157 (88%) isolates studied (P < 0.001). Remarkably, all five genotypes were detected among the 71 isolates cultured from blood samples, including 48 isolates of G1 (68%) and 11 isolates of G3 (16%). The median-joining genotype network of T. asahii clinical isolates and the prevalence in each anatomical site are shown in Fig. 1.

We found a higher percentage of T. asahii G7 in blood/catheter tip samples than other anatomical sites (9.8% versus 1.8%, P = 0.005). Notably, the 11 T. asahii G7 isolates were cultured from patients admitted in 3 medical centers from different regions, instead of representing a cluster of infections documented in a single hospital.

Antifungal susceptibility tests.

Table 1 summarizes the MIC ranges (μg/ml), MIC₅₀, MIC₉₀, and geometric mean (GM) values obtained with the 284 T. asahii clinical isolates tested. At the 48-h reading, the MIC range values were 0.25 to ≥64 μg/ml for fluconazole (FLC), 0.03 to 2 μg/ml for both voriconazole (VRC) and posaconazole (POS), and 0.025 to ≥16 μg/ml for amphotericin B (AMB). The highest MIC₉₀ values were found for FLC, followed by those for AMB, POS, and VRC.

TABLE 1.

In vitro antifungal susceptibility of 284 Trichosporon asahii clinical isolates tested against triazoles and amphotericin B

Anatomic site or genotypea	No. of isolates	MIC (μg/ml)
		Fluconazole				Voriconazole				Posaconazole				Amphotericin B
		Range	50%	90%	GM	Range	50%	90%	GM	Range	50%	90%	GM	Range	50%	90%	GM
Anatomic sites
Urine	157	0.25 to ≥64	2	4	2.47	0.03–2	0.06	0.125	0.07	0.03–2	0.25	0.5	0.21	0.25 to ≥16	2	4	1.65
Blood/CVC	71	0.25 to 32	2	8	2.08	0.03–1	0.06	0.125	0.07	0.03–1	0.25	0.5	0.18	0.25 to ≥16	1	4	1.42
Respiratory	16	0.5 to 8	1	4	1.54	0.03–0.125	0.06	0.06	0.06	0.06–0.25	0.125	0.25	0.16	1 to 8	1	2	1.35
Integumentary	11	1 to 6	2	4	2.13	0.03–0.25	0.06	0.06	0.06	0.03–1	0.125	0.25	0.17	1 to 4	1	2	1.29
S.S.N.B.	11	1 to 16	2	4	2.57	0.03–1	0.125	0.125	0.12	0.06–1	0.25	0.25	0.21	0.5 to 4	1	4	1.55
Others	18	1 to 8	2	4	2.16	0.03–0.5	0.06	0.125	0.06	0.03–0.5	0.125	0.25	0.14	0.25 to ≥16	1	4	1.71
Genotypes
G1	216	0.25 to ≥64	2	8	2.30	0.03–2	0.06	0.25	0.06	0.03–2	0.25	0.5	0.18	0.25 to ≥16	2	4	1.57
G3	28	0.5 to 16	2	2	1.95	0.03–1	0.06	0.125	0.08	0.06–1	0.25	0.25	0.22	0.5 to ≥16	1	1	1.19
G4	8	1 to 4	1	2	1.54	0.03–0.125	0.06	0.06	0.05	0.125–0.25	0.125	0.125	0.13	0.5 to 2	1	2	1.09
G5	21	0.5 to 4	2	2	1.61	0.03–0.125	0.06	0.125	0.04	0.03–0.5	0.25	0.5	0.16	0.25 to 4	2	4	1.14
G7	11	1 to 32	2	16	4.83	0.06–1	0.125	0.5	0.19	0.06–1	0.5	1	0.36	1 to 4	2	4	2.13

^aCVC, catheter tips; SSNB, sterile sites, nonblood.

We found substantial variation in MIC₉₀ values when comparing certain combinations of drugs and genotypes tested. In this scenario, the isolates representative of T. asahii G7, which emerged after 2013 in 3 medical centers only, exhibited higher MIC₉₀ values for FLC (MIC₉₀, 16 μg/ml; P = 0.03), VRC (MIC₉₀, 0.5 μg/ml; P < 0.001), and POS (MIC₉₀, 0.1 μg/ml; P < 0.02) than all other genotypes. No significant differences were found in AMB MIC values for the different T. asahii genotypes (P = 0.1). MIC values tended to be similar when checking isolates cultured from different anatomic sites (P > 0.05). Blood/catheter tips and sterile site isolates showed FLC, VRC, POS, and AMB MIC₉₀ values of 8 μg/ml, 0.125 μg/ml, 0.5 μg/ml, and 4 μg/ml, respectively. Otherwise, isolates from all other anatomic sites showed FLC, VRC, POS, and AMB MIC₉₀ values of 4 μg/ml (P = 0.7), 0.125 μg/ml (P = 0.2), 0.5 μg/ml (P = 0.7), and 4 μg/ml (P = 0.5), respectively.

As shown in Fig. 2, by using the ECV values for T. asahii recently described by Francisco et al. (4), we found a trend for increasing rates of non-wild-type isolates of VRC in the second period of study (P1, 5/72 [7%] versus P2, 21/212 [10%]; P = 0.02). When isolates belonging to the G7 genotype were excluded from the analyses, the rates of non-wild-type isolates of VRC did not differ between the two periods (P1, 0/72 [0%] versus P2, 11/212 [5.2%]).

FIG 2.

Historical trends in the antifungal susceptibility of Brazilian clinical isolates of T. asahii collected over the last 20 years. The isolates that exhibited MICs higher than the epidemiologic cutoff values for each antifungal tested (yellow lines) were identified by red circles. FLC, fluconazole; VRC, voriconazole; POS, posaconazole; AMB, amphotericin B; P1, 1997 to 2008; P2, 2009 to 2019; *, P < 0.05.

DISCUSSION

Trichosporon asahii has recently attracted worldwide medical attention due to its capability to cause a broad spectrum of infections in humans, including life-threatening invasive infections in immunocompromised hosts and critically ill patients. Additionally, in contrast with Candida spp., Trichosporon is a poor target for echinocandins and amphotericin B (2–7).

Trichosporon asahii usually comprises the majority of all Trichosporon spp. found in urine and blood cultures (3–6). Data on the incidence of this pathogen in clinical samples are scarce, and our results show an approximate incidence of 0.5 to 2 cultures per 10,000 hospital admissions. While a higher incidence of positive cultures was reported in Taiwan, >12 per 100,000 admissions, much lower rates are reported in US or European centers (10, 11).

Despite the large number of T. asahii genotypes described so far, there is a lack of robust data on the real prevalence in different anatomical sites and potential peculiarities in terms of antifungal susceptibility.

Concerning the intraspecific diversity of T. asahii, our results showed the presence of five circulating genotypes in Brazilian patients: G1, G3, G4, G5, and G7. Consistent with previous studies, including Brazilian series, T. asahii G1 (76%) was the most common genotype, followed by G3 (9.8%), G5 (7.5%), G7 (3.9%), and G4 (2.8%) (5, 6, 12–22).

Interestingly, a high proportion of G7 isolates were found in blood/catheter tip samples. The G7 genotype was described in 2004 from blood samples of Japanese patients (21) and later found in Thailand, Taiwan, Greece, Argentina, Mexico, India, Tunisia, and China (5, 15–18, 23–25). These results suggest a more virulent trait of G7, but in vitro and in vivo studies comparing the virulence of all different described genotypes are warranted to confirm this issue.

We were not able to find any representative of genotypes 2, 6, or 8 to 15 in our T. asahii collection. These genotypes have been reported in different geographic regions, with variable prevalences and distributions. The first seven genotypes (G1 to G7) of T. asahii have been extensively recovered from a large diversity of clinical samples and patients living in different countries (5, 6, 12–22).

Otherwise, G8 was only mentioned by a single report published in 2010, and G9 has only been reported from Turkish studies (6, 19). Two different research groups, one from China and another from Greece, reported the isolation of genotypes 10 and 11 in their collections (23, 26). There are additional reports of isolation of G12 in patients from China, and T. asahii isolates representative of G13, G14, and G15 were reported in a single publication from Brazil (9, 26).

Although 15 different genotypes for T. asahii have been described around the world, some of them have only been found in single reports (9, 19) and require confirmation to determine if they represent the genetic diversity of this pathogen or if their description was related to some artifacts faced during their sequence analysis.

Despite the lack of statistical differences in the AST results obtained with isolates cultured from several anatomical sites, attention is drawn to the high MIC₉₀ values for FLC (8 µg/ml) in T. asahii isolates obtained from blood samples/central venous catheters. These data suggest that MIC determination may have a role in defining the best strategy to treat patients with T. asahii fungemia.

Voriconazole exhibits the best in vitro activity against T. asahii clinical isolates and is considered first-line therapy, especially in hematological patients (3). High MIC values in T. asahii isolates are not commonly reported for this compound. In our study, the increase of rates of non-wild-type (NWT) T. asahii clinical isolates for VRC over the last 10 years was certainly influenced by the emergence of G7 isolates. Similar results were found in G7 clinical isolates tested against triazoles in a center from Greece (23). As reported in other yeast genera (Candida and Cryptococcus), we believe that the low susceptibility of G7 for azoles can be influenced by the selective pressure of their widespread use prophylactically or empirically (8). Otherwise, until now, these organisms have represented only a small proportion of the T. asahii clinical isolates found in Brazilian medical centers, and less than 15% in series published in other countries (5, 15–18, 23–25). A previous systematic review on Trichosporon fungemia shows a notable increase of azole-based treatments in the late 2000s, and the emergence of resistant isolates in recent years is a concern (27).

In conclusion, the present paper represents the largest series of T. asahii isolates characterized by sequencing IGS1 of rDNA region and antifungal susceptibility testing. We were able to demonstrate four main aspects: (i) despite the fact that 15 different T. asahii genotypes have been described by different authors, only 5 of them appeared in this large Brazilian isolate series; (ii) the five T. asahii genotypes found were cultivated from different clinical samples, illustrating their ability to colonize and infect several anatomic sites; (iii) isolates recovered from different anatomical sites exhibited similar MIC values to all tested drugs; (iv) isolates of G7 emerged throughout the last 10 years and exhibit limited susceptibility to all triazoles investigated as well as to amphotericin B.

Based on our findings, we may suggest that monitoring of T. asahii genotype distributions and antifungal susceptibility profiles is warranted to prevent the spread of azole-resistant isolates.

MATERIALS AND METHODS

Selection of T. asahii clinical isolates.

We selected a total of 284 T. asahii clinical isolates obtained from different patients that were admitted in 24 Brazilian medical centers at different times in the period between January 1997 and May 2019. The isolates were obtained from different anatomical sites and sent to Laboratório Especial de Micologia, Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, Brazil, for further identification and antifungal susceptibility testing.

Only one isolate per patient was selected for further molecular identification, genotyping investigation, and antifungal susceptibility testing. The medical centers were encouraged to provide the annual incidence of T. asahii cultures per 10,000 admissions/year. The study protocol was approved by UNIFESP’s ethics committee (CEP-UNIFESP number 401.893/2013).

Molecular species identification of T. asahii isolates.

Molecular identification was performed by sequencing the IGS1 region of rDNA using the 26SF and 5SR primer pair as previously described (12). Contigs were assembled based on two reads per isolates and edited using Phred-Phrap-Consed, targeting a Phred score of >30 (28, 29). The consensus was compared with the sequences deposited in the NCBI database (http://ncbi.nlm.nih.gov) using the BLASTn tool, looking for the percentage of identity and sequence coverage of ≥98% and an E value of <10⁻⁵ (4). Additionally, the nucleotide sequences were aligned with sequences of T. asahii type strains deposited in the GenBank (https://www.ncbi.nlm.nih.gov/genbank/). The phylogenetic tree was created by using the neighbor-joining method based on the Kimura two-parameter model with 1,000 bootstrap pseudoreplicates in MEGA7 (30–33).

Genotyping analysis of T. asahii.

Sequences of the IGS1 of rDNA region from 284 clinical isolates identified at the species level were aligned with sequences deposited in GenBank (http://www.ncbi.nlm.nih.gov/genbank/) representative of the 15 identified genotypes of T. asahii. Alignments were performed using the muscle algorithm implemented by SeaView 4.2.12 and adjusted by eye before performing haplotype analyses (33, 34). Genotype analysis was implemented in DNAsp 5.0, excluding gap positions, and the median-joining network was constructed and visualized using the software Network 5.0.1.1 (35, 36). To check the consistency of the network results, the T. asahii genotypes were also investigated by the neighbor-joining analysis performed by SeaView 4.2.12 (34), according to previous studies (5, 6, 13, 19, 21, 23–25).

Antifungal susceptibility testing of T. asahii.

Antifungal susceptibility testing (AST) for fluconazole (FLC), voriconazole (VRC), posaconazole (POS), and amphotericin B (AMB) (Sigma, St. Louis, MO, USA) was performed by using the CLSI broth microdilution method according to the document M27, 4th edition (37). The MIC values were determined after 48 h of incubation (4, 37). The following drug concentration ranges were tested: 0.125 to 64 μg/ml for FLC and 0.03 to 16 μg/ml for VRC, POS, and AMB. Quality control organisms were included in all experiments and were represented by the ATCC 6258 Candida krusei and ATCC 22019 Candida parapsilosis strains. Antifungal MIC distributions obtained with 284 T. asahii isolates cultured from different anatomical sites and representatives of different genotypes were compared using analysis of variance (ANOVA) performed with the aid of SPSS program, and P values of <0.05 were considered significant.

To check for historical trends in the prevalence rates of non-wild-type (NWT) clinical isolates regarding their susceptibility to antifungals, the total collection of isolates was divided into two different periods: (i) P1, including 72 clinical isolates collected between 1997 and 2008, and (ii) P2, including 212 clinical isolates collected between 2009 and 2019. The definition of wild-type and non-wild-type T. asahii isolates was based on the epidemiological cutoff values (ECVs) recently proposed by Francisco et al. (4): 8 μg/ml for FLC, 0.125 μg/ml for VRC, 0.5 μg/ml for POS, and 4 μg/m for AMB. The distributions of MICs obtained with all isolates collected during both periods (P1 versus P2) were compared by Student's t test (P < 0.05).

Data availability.

The rDNA IGS sequences of the 284 Trichosporon asahii clinical isolates were deposited in the GenBank database under accession numbers MT460677 to MT460907, FJ169359 to FJ169365, FJ172160 to FJ172173, FJ187679 to FJ187697, KM488282 to KM488284, EU934808 to EU938058, and EU980326. For complete information, see Table S1 in the supplemental material.