Genetic microevolution of clinical Candida auris with reduced Amphotericin B sensitivity in China

ABSTRACT

The global rate of Amphotericin B (AmB) resistance in Candida auris has surpassed 12%. However, there is limited data on available clinical treatments and microevolutionary analyses concerning reduced AmB sensitivity. In this study, we collected 18 C. auris isolates from five patients between 2019 and 2022. We employed clinical data mining, genomic, and transcriptomic analyses to identify genetic evolutionary features linked to reduced AmB sensitivity in these isolates during clinical treatment. We identified six isolates with a minimum inhibitory concentration (MIC) of AmB below 0.5 µg/mL (AmB^0.5) and 12 isolates with an AmB-MIC of 1 µg/mL (AmB¹) or ≥ 2 µg/mL (AmB²). All five patients received 24-hour AmB (5 mg/L) bladder irrigation treatment. Evolutionary analyses revealed an ERG3 (c923t) mutation in AmB¹ C. auris. Additionally, AmB² C. auris was found to contain a t2831c mutation in the RAD2 gene. In the AmB¹ group, membrane lipid-related gene expression (ERG1, ERG2, ERG13, and ERG24) was upregulated, while in the AmB² group, expression of DNA-related genes (e.g. DNA2 and PRI1) was up-regulated. In a series of C.auris strains with reduced susceptibility to AmB, five key genes were identified: two upregulated (IFF9 and PGA6) and three downregulated (HGT7, HGT13,and PRI32). In this study, we demonstrate the microevolution of reduced AmB sensitivity in vivo and further elucidate the relationship between reduced AmB sensitivity and low-concentration AmB bladder irrigation. These findings offer new insights into potential antifungal drug targets and clinical markers for the “super fungus”, C. auris.

Introduction

Candida auris is rapidly emerging as a global pathogen due to its increasing incidence of drug resistance. Amphotericin B (AmB) is typically employed as the last line of defense in treating fungal infections, particularly in patients with C. auris infections that are resistant to echinocalcins or unresponsive to echinocalcins [1]. Unlike common Candida species or other yeasts, resistance to AmB, including in haploid organisms such as Candida glabrata remains extremely rare despite over 50 years of clinical use [2,3]. Recent meta-analyses have revealed an AmB resistance rate of 12% in C. auris isolates [4]. Currently, the molecular basis underlying reduced susceptibility to AmB or AmB resistance in any yeast species is poorly understood. Studies have indicated that the molecular mechanisms by which Candida spp. develops resistance to AmB are related to specific target genes, mainly ERG2, ERG3, ERG6, and ERG11 [5–11]. One study identified a mutation in ERG6 as a cause of AmB resistance in clinical strains of C. auris [12]. Additionally, transcriptional profiling of AmB-resistant isolates has shown a significant difference in the expression of ergosterol biosynthesis genes, suggesting that AmB resistance may be associated with changes in membrane lipid permeability and chromatin remodelling [13]. The aforementioned studies focused exclusively on AmB-resistant strains. Therefore, in this study we analyzed strains with elevated minimum inhibitory concentrations (MICs) for AmB as well as a series of strains at different evolutionary stages (i.e. AmB MIC ≤ 0.5μg/mL, AmB MIC = 1μg/mL, and AmB ≥ 2 μg/mL) to comprehensively investigate the microevolutionary mechanisms behind reduced AmB sensitivity.

We observed that these strains appear to be associated with the clinical application of AmB therapy (i.e. intravenous administration, atomized inhalation, and bladder irrigation), warranting further investigation. However, in China, there have been limited studies on AmB resistance [14,15]. Continuous surveillance in recent years has revealed the emergence of several strains of AmB-resistant C. auris in Shenyang, China. Given that studies have shown high levels of AmB resistance in C. auris following AmB intravenous administration [12], we initially focused on investigating cases of C. auris strains with reduced sensitivity as indicated by AmB MICs.

In this study, we systematically investigated the clinical characteristics of a series of AmB-resistant strains, as well as the evolutionary features of these strains, using genomic and transcriptomic analyses to uncover the evolutionary basis of AmB resistance in clinical C. auris.

Methods

Isolates

Clinical isolates of C. auris were regularly isolated from Sabouraud dextrose agar or CHROMagar Candida medium and incubated at 37 °C under atmospheric conditions from January 2019 to September 2022. Antifungal susceptibility testing of all clinical C. auris isolates was performed by broth microdilution (ATB Fungus 3, BioMerieux France). The MIC endpoints for AmB were defined as the lowest drug concentration that caused 100% growth inhibition after 24 h of incubation at 35 °C. Clinical C. auris isolates were screened for drug sensitivity, and those with MIC values for AmB greater than 0.5μg/mL were included in the analysis. Further, we also included AmB-sensitive strains originating from the same patient for comparative analysis. We performed the skin screening (axilla and groin) on September 15th, 2022 in the two cases relevant to this study (Supplementary Table 1).

In addition, all strains of this experiment were tested for sensitivity to drugs such as echinocandins by applying a commercial chromogenic susceptibility plate (Sensititre YeastOne, Thermo Fisher Scientific). Candida parapsilosis ATCC22019 and Candida krusei ATCC6258 were used for strict quality control during testing.

Case patient information

Patient information data, including sex, age, ward, underlying diseases, and antifungal drugs administration (including AmB intravenous, atomized inhalation, and bladder irrigation), and antifungal drug dosage, were collected via the Hospital Information System of the First Hospital of China Medical University. This study was approved by the first hospital of China Medical University (ERC No. 2024-41). Clinical samples were obtained after verbal consent only as part of routine patient care and diagnostic workup for pathogen isolation and susceptibility testing. The case diagram was generated on https://app.diagrams.net.

Genome sequencing

Genome sequencing was performed by Shanghai Personal Biotechnology Co., Ltd using the Illumina NovaSeq platform as described previously [14]. For variant calling analysis, the reference genome (C. auris Respiratory Intensive Care Unit (RICU)1_A1) was downloaded from the National Center for Biotechnology Information (NCBI) genome database (NCBI accession: ASM1421745v1). The variants that were predicted to alter the amino-acid sequence in any coding sequence (nonsynonymous single nucleotide variants, stop loss, or gain variants, as well as indels) were annotated using the ANNOVAR software [16] and the RefSeq C. auris B11221 coding sequence.

Phylogenetic analysis

The maximum likelihood method based on the Tamura-Nei model was used to predict evolutionary history [17]. The location of blank and missing data was cleared for analysis and MEGA7 [18] software was used for the evolutionary analyses. A phylogenetic tree was constructed using https://www.chiplot.online/normalTree.html. Protein–protein interaction (PPI) network analysis of non-mutated genes was conducted using the STRING database (https://cn.string-db.org). PPI network diagrams were constructed on https://www.chiplot.online/normalTree.html.

Transcriptome analyses

RNA sequencing analyses were performed using the Illumina NovaSeq platform (Shanghai Personal Biotechnology Co.,Shanghai, China). For RNA extraction, C. auris cells were grown in YPD broth medium at 30 °C in a shaking incubator at 220 rpm. After 18 h, cells at the stationary phase were diluted with an equal volume of fresh YPD broth and incubated for 2 h at 37 °C to induce growth. Cells were then centrifuged for 10 min at 3000 × g before pellets were flash frozen and stored at −80°C. Total RNA was purified using a GeneJET RNA purification kit (Thermo Scientific) before RNA quality was assessed on a Bioanalyzer using the RNA6000 Nanochip (Agilent). Next, mRNA was enriched using oligo (dT) beads (New England BioLabs (NEB). Subsequently, double-stranded cDNA libraries were generated using the NEBNext Ultra directional RNA library prep kit for Illumina (NEB) according to the manufacturer’s instructions. The qualified libraries were subjected to Illumina sequencing with 150 bp paired end reads at the Novogene sequencing facility. Three biological replicates were included for each condition.

For data mapping analysis, the reference genome (C. auris B11221) and gene annotation files were downloaded from genome website. Differentially expressed genes (DEGs) were screened for expression difference fold |log₂FoldChange| > 1, and a significant adjusted p-value < 0.05.

Several important genes were selected for qPCR validation and the primers are shown in Supplementary 3.

Results

Clinical isolates and patient information

A total of 18 C. auris isolates were obtained between January 2019 and September 2022, including 13 from urine specimens, one from cerebrospinal fluid, and four from the skin. All the isolated strains exhibited high levels of resistance to fluconazole and sensitivity to 5-fluorocytosine. The MIC ranges for voriconazole and itraconazole were 1–4 μg/mL and 0.125-0.5μg/mL, respectively (Supplementary Table 1). Given that the MIC values for the two QC strains (Candida parapsilosis ATCC22019 and Candida krusei ATCC6258) tested by Sensititre YeastOne (SYO) against AmB were 1–2 dilutions higher compared to those tested by Fungus3 (F3), and considering that other studies suggest SYO overestimates AmB resistance in C. auris [19], we used MIC values from the F3 test for subsequent experiments. C. auris strains were grouped according to their MIC values for AmB as follows: ≤ 0.5, 1 μg/mL, and ≥ 2 μg/mL, designated as AmB^0.5, AmB¹, and AmB², respectively. Strains in the AmB^0.5 group included RICU40_A425, RICU41_A431, RICU43_A478, RICU40_Y024, RICU37_S006, and RICU37_S010. The AmB¹ group included RCIU38_A382, RICU40_A441, RICU40_A447, RICU40_A451, RICU41_A454, and RICU41_A457. The AmB² group consisted of six strains: Neurosurgical Intensive Care Unit (NSICU) 2_A120, RICU38_A397, RICU40_Y021, RICU43_A485, and RICU43_A485, which had an MIC of 2 μg/mL for AmB in the F3 assay and 4 μg/mL in the SYO assay. The remaining strain, RICU38_A398, had MICs of 4 μg/mL in the F3 assay and 8 μg/mL in the SYO assay. According to the Centers for Disease Control's tentative breakpoint, C. auris is considered AmB-resistant if the MIC is higher than 2 μg/mL. Therefore, the six strains in the AmB² group should be regarded as AmB-resistant.

Based on these strains, we retrospectively reviewed relevant cases (Figure 1). Only one strain was isolated from a NSICU patient, while the remaining isolates were all obtained from RICU patients. With the exception of Case 1, in which the patient was not treated with any antifungal drugs for cerebrospinal fluid C. auris infections, all four patients were empirically initiated on 5 mg/L/12 h AmB atomized inhalation for respiratory tract Candida infections, followed by 5 mg/L/24 h AmB continuous bladder irrigation. Reduced AmB sensitivity gradually emerged following AmB bladder irrigation, while all respiratory cultures remained negative for C. auris after a course of respiratory atomized inhalation. During intravenous treatment with AmB, five blood cultures were taken from RICU38 and sent for testing, all of which were negative.

Figure 1.

Treatment regimens and isolates collection for five cases: one without AmB treated and four AmB treated patients. AmB, Amphotericin B; MCF, Micafungin. The rectangles above the timeline represent isolated C. auris with different MICs. Green rectangle: AmB ^0.5 isolates, yellow rectangle: AmB ¹ isolates, pink and orange rectangle: AmB ² isolates. The lower part of the timeline represents the different treatments. Blue arrow: AmB bladder irrigation, brown arrow: AmB atomized inhalation, pink arrow: AmB intravenous drip.

Case 1

NSICU2, a 65-year-old female, was admitted to the neurosurgery department on December 9th, 2018, due to a pituitary tumour. C. auris was cultured from her cerebrospinal fluid for the first time on February 18th, 2019. The isolate, A120, had an MIC of 4 μg/mL for AmB, indicating drug resistance. No antifungal drugs were used and cerebrospinal fluid specimens were not examined again. The disease did not worsen 18 days after the isolation of C. auris (A120), which was considered colonization rather than infection. The patient was discharged from the hospital on March 8th, 2019.

Case 2

RICU38, a 56-year-old male, previously diagnosed with hypertension and diabetes, was admitted to the ICU of the Department of Respiratory and Critical Care Medicine on October 12th, 2021, due to a right thalamic hemorrhage, Guillen Barré syndrome, and repeated pulmonary infections. During hospitalization, the patient received ventilator-assisted ventilation and anti-infection treatment. From October to December 2021, C. albicans and C. glabrata were repeatedly cultured from sputum samples. Fluconazole (0.4 g/24 h) was administered intravenously, and AmB (10 mg/12 h) was given as atomized inhalation therapy. No Candida growth was observed in the sputum after treatment. On January 13th, 2022, C. auris was isolated and cultured from the patient’s urine samples for the first time (AmB MIC ≤ 0.5 μg/mL), suggesting an AmB-sensitive C. auris urinary infection. On January 15th, 2022, he was given Micafungin 150 mg for 17 days (from January 15th, 2022, to February 2nd, 2022) due to Candida lung infection. On the same day, the patient was then treated with continuous bladder irrigation with AmB (5 mg/L) for 24 hours. On January 23rd, 2022, C. auris A382 (AmB MIC = 1 μg/mL and resistant to echinocandins) was isolated from the patient’s urine culture. Continuous bladder irrigation with AmB was administered until March 13th, 2022, followed by 5 mg of intravenous AmB infusion. C. auris A397 (AmB MIC = 2 μg/mL) and A398 (AmB MIC = 4 μg/mL) were isolated from urine specimens on February 27th and March 10th, respectively. Reduced AmB sensitivity gradually increased during irrigation, with MIC values rising from 1 μg/mL to 4 μg/mL. Finally, due to septic shock and multiple organ dysfunction, the family stopped further treatment and the patient left the hospital on March 13th, 2022.

Case 3

RICU40, a 70-year-old male patient, underwent right fossa dural arteriovenous fistula amputation and right temporo-parietal hematoma removal on April 19th, 2022. This patient also suffered from repeated postoperative pulmonary infections. He was admitted to the RICU on May 7th, 2022, and was placed on tracheal intubation with ventilator-assisted ventilation during hospitalization. On May 12th, the patient developed a fever and elevated inflammatory markers such as CRP and PCT. Anti-inflammatory therapy was administered with cefoperazone-sulbactam (3 g/8 h) and meropenem (1 g/6 h), but symptoms did not improve. In May 2022, repeated sputum cultures indicated C. auris growth (AmB MIC ≤ 0.5 μg/mL), and AmB (5 mg/12 h) atomized inhalation therapy was initiated. Micafungin 150 mg (2022-5-23–2022-6-19) was given to treat the C. auris lung infection for 26 days. C. auris A425 (AmB MIC ≤ 0.5 μg/mL) was isolated and cultured from urine samples for the first time on May 25th, 2022, and continuously isolated in urine thereafter. On May 30th, 2022, 24-hour AmB (5 mg/L) continuous bladder irrigation was added to the patient’s treatment regimen. Subsequently, C. auris A441 (AmB MIC = 1 μg/mL), A447 (AmB MIC = 1 μg/mL), and A451 (AmB MIC = 1 μg/mL), all with decreased AmB sensitivity, were continuously isolated from urine samples. During the screening of the patient's skin on September 15th, 2022, C. auris Y021 (AmB MIC = 2 μg/mL) was isolated from the patient's groin, and axillary strain Y024 (AmB MIC ≤ 0.5 μg/mL) was isolated. At the same time, S010 and S006 were isolated from the groin and axilla of a neighbouring RICU5 patient. After a period of treatment, the patient was discharged from the hospital on November 7th, 2022.

Case 4

RICU41, a 49-year-old female patient, underwent right meningioma resection in the neurosurgery department in March 2022 and suffered from repeated postoperative pulmonary infections. She was admitted to the RICU on May 24th, 2022 and was placed on tracheal intubation with ventilator-assisted ventilation during hospitalization. In May 2022, C. glabrata was repeatedly isolated from sputum specimens, so the patient was given atomized inhalation therapy with AmB (5 mg/12 h) on May 27th. C. auris A431 (AmB MIC ≤ 0.5 μg/mL) was isolated from urine samples for the first time on May 31st. On June 3rd, the patient’s catheter was replaced, and 24-hour AmB (5 mg/L) continuous bladder irrigation was administered. Following anti-infection treatment, the patient’s infection index and fever symptoms were improved, but the urinary tract infection persisted, and the indwelling urinary tube was removed on June 20th, 2022. C. auris A454 (AmB MIC = 1 μg/mL) and A457 (AmB MIC = 1 μg/mL) were isolated from urine samples on June 22nd and June 25th, 2022. After a period of treatment, the patient's fever symptoms and liver function improved significantly, and she was discharged on June 29th, 2022.

Case 5

RICU43 was a 78-year-old female with diabetes mellitus and hypertension. Intermittent dyspnea had occurred for nine months before admission, and no complete improvement was observed after anti-infection and ventilator-assisted ventilation treatment. Intravenous therapy was administered with 50 mg tigecycline, 0.5 g/q.d. levofloxacin, and 1.0 g/q.8 h meropenem. The patient was transferred to the RICU on August 1st, 2022, due to respiratory failure and severe pneumonia. During hospitalization, she was placed on tracheal intubation with ventilator-assisted ventilation. C. albicans was repeatedly cultured from sputum samples in the early stage of admission, and fluconazole (0.4 g/24 h) antifungal therapy was administered. From August 15th to September 22nd, C. auris (AmB MIC ≤ 0.5 μg/mL) was continuously isolated from the patient's sputum, and because these isolates showed sensitivity to AmB, the patient was treated with 5 mg/12 h AmB as an atomized inhalation therapy. C. auris A478 (AmB MIC ≤ 0.5 μg/mL) was isolated and cultured from urine samples for the first time on August 17th, and the patient was treated with 24-hour AmB (5 mg/L) continuous bladder irrigation. On August 25th and 26th, C. auris A483 (AmB MIC = 2 μg/mL) and A485 (AmB MIC = 2 μg/mL) were isolated. After this, the patient’s condition became critical due to septic shock and multiple organ failure. Subsequently, the family declined further treatment, and the patient left the hospital on September 23rd, 2022.

We also retrospectively reviewed 15 clinical cases prior to 2019. The six cases from the Neurosciences Intensive Care Unit (NICU) and the seven cases from the RICU were treated without bladder irrigation. To clear the urinary tract of C. auris infection or colonization, saline bladder irrigation was used in two cases from the NICU in 2018. There were no increases in the MIC values of AmB in patients treated with saline bladder irrigation compared to AmB bladder irrigation (Figure 2A). These results suggest a potential relationship between AmB bladder irrigation therapy and increased AmB MICs.

Figure 2.

Genome sequencing of clinically isolated C. auris. (A) Molecular phylogenetic analysis of 45 isolates (including 16 serial isolates from all six patients in this study and 29 previously studied isolates). The phylogenetic tree constructed using 403 SNPs has two main clusters (Cluster A and Cluster B). (B) Heatmap of the 18 C. auris strains isolated in this study plotted against their 46 non-synonymously mutated genes using heatmap tools in the genescloud platform (https://www.genescloud.cn).

Phylogenetic analyses of clinical C. auris isolates

Consistent with previous phylogenetic and population structure analyses, all isolates in this study belonged to the South African clade (clade III) [14] A total of 45 isolates (including 16 serial isolates from all six patients in this study and 29 previously studied isolates [14]) were phylogenetically analyzed to explore the origin of AmB¹/AmB² C. auris cases. The phylogenetic tree was constructed using 403 single nucleotide polymorphisms (SNPs) and revealed two main clusters (Cluster A and Cluster B). Cluster A consisted mainly of C. auris isolates involved in this study (AmB¹ and AmB² C. auris). Nine strains (9/15) were isolated from four RICU patients treated with AmB bladder irrigation. Isolates from each patient formed small independent branches rather than being grouped as resistant and sensitive strains. Surprisingly, we found that the most genetically similar isolates to the AmB¹ and AmB² C. auris strains were those isolated from NICU patients (e.g. NICU8_A45, NICU12_A79, among others) and those taken from the bedrail of RICU patients (C12_A109) before 2019. Cluster B was mainly composed of from previously isolated C. auris strains from RICU and NICU patients. All of the 29 C. auris clinical isolates in this branch showed sensitivity to AmB. The patients who were not treated with bladder irrigation therapy and those who used saline bladder irrigation were included in this branch (Figure 2A).

Genome-wide SNP locus analysis

A total of 77 SNP sites were found to have mutations, of which 46 were nonsynonymous and 31 were synonymous (Supplementary Table 2). Among them, one gene, CJI97-005651, possessed the most SNPs, with 7–11 nonsynonymous mutations (c.g3a.M1I, c.g4a.A2T, c.g47a.R16H, c.t50a.V17E, c.t53g.L18R, c.g56t.G19V, c.t61g.Y21D, c.t1247g.V416G, c.c1262g.P421R, c.c1265a.A422E, and c.g1276c.G426R) in all strains except strain A398, as well as five synonymous mutations. Four genes had 2–3 nonsynonymous mutations, IFF6 (c.a146t.Y49F, c.g525c.M175I), NMA111 (c.c940t.L314F, c.t2432a.L811H), AMN1 (c.g845a.C282Y, stopgain c.c368a.S123X), and FKS1 (c.t1915c.S639P, c.c1916t.S639F), while the other 26 genes had nonsynonymous mutations at a single SNP. Notably, certain nonsynonymous gene mutations may be associated with reduced susceptibility to AmB, such as those genes encoding delta-5-6 steroid desaturase ERG3 (c.c923t.T308M), glycosylphosphatidylinositol (GPI)-anchored cell wall protein-encoding genes IFF9 (c.t360a.F120L), IFF6, and RBR3 (c.g4154a.G1385D), sphingolipid delta-8 desaturase SLD (c.g1717c.V573L), DNA-damage repair-related genes RAD2(c.t2831c.L944P) and RAD9 (c.a1964g.G655R), sterol uptake, and ergosterol biosynthesis genes UPC2 (c.g95t.R32M). RAD2 is a nucleotide excision repair (NER) gene, and NER mutants are very sensitive to UV-induced DNA damage [20]. The RAD9 gene encodes a chromatin binding protein that acts as a signal transducer at DNA-damage checkpoints [21] and plays a role in DNA repair and metabolism. UPC2 is a sterol uptake control protein and a transcription factor involved in the regulation of ergosterol biosynthesis and sterol uptake at the plasma membrane [22]. The tightly linked proteins encoded by 11 genes involved in C. albicans homologs in the protein interaction network (medium confidence = 0.4) are shown in Supplementary Figure 1A.

Nonsynonymous mutations in RAD2, PRO41 (c.g121c.A41P), and UPC2 were only found in AMB² isolates (NSICU2_ A120). Strains with nonsynonymous mutations in ERG3 were found only in the AmB¹ group (RICU41_A457) and in RAD9 in RICU41_A454 (Supplementary Figure 1B). Isolates RICU38_A382, RICU38_A397, and RICU38_A398 had mutations in SLD1, UTP22 (c.c406t.L136F), and MRPL10 (c.t833c.V278A). These three strains also had a nonsynonymous mutation in FKS1 (c.t1915c.S639P) as well as an echinocandin-resistant phenotype. As a result, these three mutations (SLD1, UTP22, and MRPL10) cannot be excluded from a potential association with echinocandin resistance in C. auris (Figure 2B) [23].

Transcriptomic analyses

To further investigate the relationship between AmB bladder irrigation and reduced AmB sensitivity, 11 clinical isolates from four patients treated with AmB bladder irrigation were selected, and control isolates were selected from three previously studied C. auris isolates (NICU8_ A45, NICU12_ A79, and C12_ A109) for phylogenetic analyses (Figure 3A).

Figure 3.

Transcriptomic data analysis. (A) Relationship between isolates and bladder irrigation. The 11 strains above were treated with AmB bladder irrigation and the three following strains were not treated with bladder irrigation. (B) Principal-component analysis (PCA) of normalized RNA-seq read counts from three biological replicates per isolate displays the level of correlation and the reproducibility among different biological replicates. R package DESeq2 software for PCA analysis, ellipse parameters set as follows: topN < - “500”; type < - “t”; c_level < -0.95; segments = 101; (C) Amino acid sequences of 152 identified DEGs were used in protein-protein interaction (PPI) network analysis. Candida albicans SC5314 homologous protein was used as a reference. PPI network was constructed using https://www.chiplot.online/normalTree.html. Grey circle: upregulated genes, grey forked cross: downregulated genes. Different colours represent different gene function classifications. (D) Heatmap of the 152 identified DEGs in AmB^0.5, AmB¹, and AmB² isolates.

Multivariate principal- component analyses revealed similarities between biological replicates as well as differences among the C. auris isolates (Figure 3B). The biological replicates from each isolate clustered together, indicating a high level of data correlation, but the three groups did not cluster separately. There were 152 DEGs in the AmB¹ and AmB² C. auris isolates compared to the AmB^0.5 isolates (Supplementary Table 3). From these 152 DEGs, the isolates were mapped according to the following groups: AmB^0.5, AmB¹, and AmB². We found that 91 genes showed upregulated expression in the AmB² group, 18 genes in the AmB¹ group, and 43 genes in the AmB^0.5 group (Figure 3D).

We also performed Gene Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of the DEGs to gain insights into the overall differential transcriptional landscapes between the different isolates by RNA-seq (Supplementary Figure 2). KEGG analysis revealed that the DEGs were clustered in DNA replication-related pathways. Furthermore, we found that 13 genes (CDC46, PRI2, DPB2, RAD51, MCM2, MCM3, MCM6, RFA1, DNA2, SLD2, POL1, CaO19.2796, and CaO19.4030) were enriched in DNA-related pathways associated with DNA replication, DNA metabolism processes, DNA geometric changes, and DNA double-strand unwinding. Based on these annotations, a total of 31 genes were directly annotated for DNA replication by computational analysis. Nearly 9% of the DNA replication-related genes (13 out of 152 genes) were differentially expressed in AmB¹ and AmB² isolates compared to AmB^0.5 isolates. Furthermore, nine genes related to integral membrane components, such as ERG1, STL1, LAC12, HGT1, HGT7, GAP1, GAP2, FTR1, and RTA1, were also differentially expressed. The cell adhesion-related genesALS4 and ALS1 were also differentially expressed.

DEG analysis of three groups (Total, RICU, RICU single case) between AmB^0.5 vs. AmB¹ and AmB^0.5 vs. AmB²isolates

To explore the key factors of microevolution leading to reduced susceptibility to AmB in C. auris, we analyzed DEGs between the two groups (AmB^0.5 vs. AmB¹ and AmB^0.5 vs. AmB²) at three levels: (1) comparison analysis of the total strains (RICU and NICU) (Figure 5B), where the Total AmB^0.5 vs. AmB¹ group had 17 upregulated genes and 28 downregulated genes, while the Total AmB^0.5 vs. AmB² group had 87 upregulated genes and 36 downregulated genes (Figure 4A, 4D); (2) comparison analysis of the RICU strains, and RICU AmB^0.5 vs. AmB¹ group had 14/81 upregulated/downregulated genes, while the RICU AmB^0.5 vs. AmB² group had 76/36 upregulated/downregulated genes (Figure 4B, 4E); and (3) Individual case from RICUs (RICU41_A431 vs. A457 and RICU43_A478 vs. A483) were analyzed separately, where 175/179 upregulated/downregulated genes; and 1312/1217 upregulated/ downregulated genes were found, respectively (Figure 4C, 4F). Finally, we performed comparative analyses between four groups (Figure 5A, 5C) and six groups (Figure 5F), identifying27 important genes (ERG1, etc.), which were related to cell membranes, cell adhesion, cell wall synthesis, DNA replication, recombination, and sugar transport (Figure 5D). Seven genes, including ERG1, were selected for qPCR validation, and their expression was consistent with transcriptome data (Figure 5E and Supplementary Table 4).

Figure 5.

Comprehensive analysis of transcriptomic data. (A) Interactive venn network of four groups (total AmB^0.5 vs. AmB¹ group, RICU AmB^0.5 vs. AmB¹ group, total AmB^0.5 vs. AmB² group, and RICU AmB^0.5 vs. AmB² group) using venn diagram tools in the Evenn platform(http://ehbio.com/test/venn/). (B) Strain information for comprehensive analysis of transcriptome results. (C) Flower plot diagram of four groups (total AmB^0.5 vs. AmB¹ group, RICU AmB^0.5 vs. AmB¹ group, total AmB^0.5 vs. AmB² group, and RICU AmB^0.5 vs. AmB² group). (D) Heatmap of the six groups plotted against their DEGs using heatmap tools in the hiplot platform (https://hiplot.com.cn/home/index.html). (E) Correlation analysis of qPCR results and transcriptome data for seven genes including ERGs. (F) Flower plot diagram of six groups using venn diagram tools in the Evenn platform (http://ehbio.com/test/venn/). Important genes are labelled and key genes are circled in red boxes.

Figure 4.

Transcriptomic profiling of three groups (total, RICU, RICU single case) between AmB^0.5 vs. AmB¹ and AmB^0.5 vs. AmB² isolates. Important genes are labelled and key genes are circled in red boxes. (A)Volcano map shows up- and downregulated genes in the total AmB^0.5 group vs. total AmB¹ group. Upregulated genes are indicated by red dots, downregulated genes are indicated by blue dots. (B) Volcano map shows up- and downregulated genes in the RICU AmB^0.5 group vs. RICU AmB¹ group. (C) Volcano map shows up- and downregulated genes in the RICU41_A431 AmB^0.5 group vs. RICU41_A457 AmB¹ group. (D) Volcano map shows up- and downregulated genes in total AmB^0.5 group vs. total AmB² group.(E) Volcano map shows up- and downregulated genes in the RICU AmB^0.5 group vs. RICU AmB² group.(F) Volcano map shows up- and downregulated genes in the RICU43_A478 AmB^0.5 group vs. RICU43_A483AmB² group.

After two-by-two and four-group comparisons in the AmB^0.5 vs. AmB¹ groups, we identified important upregulated genes, including ERG1, ERG2, ERG13, and ERG24, which are involved in the ergosterol biosynthesis pathway. These genes were also associated with cell membranes (Figure 3C and 5D). In comparison, in the AmB^0.5 vs. AmB² group, important upregulated genes (i.e. SLD2, PRI2, MCM2, MCM3, MCM6, and DNA2) were related to DNA replication and recombination. These DNA-related genes are tightly linked in the network protein map (Figure 3C and 5D). RAD51, which is related to DNA repair, was also identified in the AmB² group.

Most downregulated genes were implicated in glucose transporter proteins (HGT1, HGT7, and HGT13). HGT1, a multifunctional complement evasion molecule, also regulates hyphae formation, leading to the downregulation of complement activation. Genes such as WOR1 and HXK1 were differentially expressed, and although the protein interactions are poorly correlated, they are all associated with virulence. Enhanced resistance appears to be compensated for by the downregulated expression of virulence genes.

After the six-group overlap analysis, we identified five key genes: two upregulated genes (IFF9, and PGA6) and three downregulated genes (HGT7, HGT13, and PRI32) (Figure 5F).

Discussion

Since 2022, MIC values of clinical C. auris isolates against AmB have significantly increased. Among them, six AmB¹ C. auris isolates (AmB MIC = 1 μg/mL) and six AmB² C. auris isolates (AmB MIC = 2 μg/mL or 4 μg/mL) were identified, indicating a low level of drug resistance. Importantly, we also identified 11 C. auris strains with elevated MIC values for AmB from urine samples of four patients, all of whom had been treated with AmB bladder irrigation. Most AmB-susceptible cases of C. auris prior to 2019 had not been treated with saline bladder irrigation, with the exception of two patients in the NICU. We hypothesize that AmB resistance in C. auris may be associated with bladder irrigation. In addition, there are usually two regimens regarding the dosage of AmB bladder irrigation: 5 mg/L (following standard instructions) and 25–50 mg/L (recommended by guidelines [1]). Based on safety and other considerations, all cases in this study were treated with 5 mg/L AmB bladder irrigation, and this regimen is largely effective in the treatment of most Candida-induced urinary tract infections in the clinic [24]. Unfortunately, the results of this study showed that continuous bladder irrigation with 5 mg/L AmB for 24-hour may induce elevated AmB MIC in C. auris, thereby suggesting that high-dose AmB bladder irrigation (25–50 mg/L) should be used to eliminate C. auris colonization or infection in the absence of clinical contraindications.

Furthermore, we characterized the genetic evolution of AmB¹ and AmB² C. auris by genomic and transcriptomic analyses. First, we found that AmB¹ and AmB² C. auris isolates in this study developed mutations in ERG3 (c.c923t.T308M) and RAD2 (c.t2831c.L944P), which are associated with cell membranes and DNA damage repair, respectively. Whether these two nonsynonymous mutations are simply polymorphisms or whether they play an important role in AmB resistance requires further experimental confirmation. Secondly, we analyzed differences in transcript levels between AmB¹ and AmB² isolates compared to AmB^0.5 isolates. We found that consistent with our genomic data, upregulated genes were related to two major categories: membrane lipids and DNA replication. Among them, the expression of membrane lipid-related genes (ERG1, ERG2, ERG13, and ERG24) was upregulated in the AmB¹ group. AmB has been shown to bind or trap membrane ergosterol to trigger changes in membrane permeability by chelating ergosterol or by regulating channel function [25–27]. In contrast, qualitative and quantitative changes in membrane lipids and a reduction in total ergosterol content are closely associated with AmB resistance [28,29]. Additionally, DNA-related genes such as those involved in DNA replication (e.g. DNA2) and DNA metabolic processes (e.g. PRI1) were upregulated in AmB² C. auris isolates. While the possibility that DNA replication may enhance AmB resistance in Candida albicans [30] has been suggested, the effect on reduced AmB sensitivity in C. auris has rarely been reported and needs further investigation. After the six-group overlap analysis, we identified five key genes: two upregulated genes (IFF9 and PGA6) and three downregulated genes (HGT7, HGT13, and PRI32). The correlation between increased or decreased expression of these genes related to cell wall or membrane metabolism and AmB resistance deserves in-depth experimental studies to identify potential therapeutic targets.

Our phylogenetic analyses suggest that the overall evolution of C. auris isolates in China (Shenyang) occurred between 2016 and 2022, and that the C. auris isolates obtained before and after 2019 formed two major clusters (A and B). Cluster A included AmB¹ and AmB² isolates that were associated with individual evolution and were not derived from the same clone. Moreover, isolates obtained from the RICU were genetically more related to isolates from pre-2019 NICU patients. Therefore, the origin of strains with reduced AmB susceptibility in this study is complex and requires further investigation. Furthermore, one C. auris strain from the RICU setting (bedrail) was genetically related to these AmB¹/ AmB² strains. Another AmB² strain was isolated from the groin of a patient (RICU40) in 2022, suggesting that AmB¹ and AmB² strains may be widely disseminated in healthcare settings, and the potential risk of an epidemic outbreak of AmB-resistant C. auris should not be underestimated.

We recognize several limitations to this study. First, the number of strains and cases involved was small. Therefore, further clarification is needed on whether low-dose AmB continuous bladder irrigation therapy is associated with increased resistance to AmB in C. auris by using a larger sample group and in vitro experimentation. Secondly, no ERG6 mutation sites were identified in AmB¹ and AmB² C. auris isolates compared to AmB^0.5. This may be related to the low level of resistance to AmB in the C. auris isolates in this study, pending further inclusion of high-level AmB-resistant isolates to explore the mechanisms of AmB resistance. Thirdly, this study used ATB Fungus 3 for testing drug sensitivity to AmB, which requires interpretation that may be subjective. Other widely used antifungal susceptibility tests, such as the SYO and automated Vitek 2 system (BioMerieux), greatly overestimate the resistance of C. auris to AmB [19,31]. Additionally, the true breakpoints or epidemiological cutoff values for AmB resistance in C. auris have not yet been established, which is an important clinical issue that needs urgent attention. Fourth, because the main target of AmB is ergosterol, Candida resistance to AmB should be accompanied by an absence of ergosterol in yeast cell membranes. In the present study, we were unable to detect sterol content due to experimental limitations. Therefore, gas chromatography-mass spectrometry (GC/MS) should be implemented in future studies to determine sterol content in isolate samples. Lastly, according to a recent study by Pezzotti G et al., different C. auris subclades have unique ergosterol/ergostane fractions analyzed by Raman spectroscopy, which significantly impact AmB resistance [32–34]. The C. auris clusters in this study all belonged to Clade III and had different MICs for AmB. Future efforts should focus on studying the Raman spectroscopy characteristics of C. auris with reduced AmB susceptibility, providing a powerful tool for the rapid and accurate detection of AmB-resistant strains of C. auris.

In conclusion, we have revealed important insights into the microevolution of reduced AmB sensitivity in C. auris cases treated with continuous AmB bladder irrigation in China (Shenyang). This suggests that clinicians should closely monitor C. auris isolates from urine specimens for changes in reduced AmB sensitivity or resistance, especially when using low-dose AmB bladder irrigation regimens. These findings will inform strategies for the elimination of C. auris and the prevention of further emergence of AmB resistance, which remains a challenging task.

Funding Statement

This work was supported by the National Key Research and Development Program of China (2021YFC2300400). This work was also supported by the National Natural Science Foundation of China 82202547.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

Sufei Tian, Chen Rong, Hailong Li, Yusheng Wu, Na Wu, and Yunzhuo Chu have made substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data; Ning Jiang and Jingping Zhang have been involved in drafting the manuscript; and Hong Shang have given final approval of the version to be published.

Data availability statement

The raw sequence data reported in this work have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2022), the China National Center for Bioinformation/Beijing Institute of Genomics, as well as the Chinese Academy of Sciences (GSA: CRA015722, CRA015704, CRA015934, and CRA016185). All data are publicly accessible at https://ngdc.cncb.ac.cn/gsa.