Cryptococcosis is primarily caused by infection with the Cryptococcus neoformans species complex (CNSC) or Cryptococcus gattii species complex (CGSC). CNSC has been classified as C. neoformans (serotype A, formerly C. neoformans var. grubii) which includes four genotypes: VNI, VNII, VNBI, and VNBII, C. deneoformans (serotype D, genotype VNIV, formerly C. neoformans var. neoformans) and C. neoformans × C. deneoformans hybrid (serotype AD, genotype VNIII). C. gattii was classified into distinct species, including C. gattii (genotype VGI, formerly C. neoformans var. gattii), C. deuterogattii (VGII), C. bacillisporus (VGIII), C. decagattii (VGIIIc/VGIV hybrid), C. tetragattii (VGIV) and other unnamed species (VGV), which can be classified into serotypes B or C (Hagen et al., 2015; Kwon-Chung et al., 2017; Montoya et al., 2021). Cryptococcus spp. often causes clinical invasive fungal diseases (IFDs), such as cryptococcal meningitis/meningoencephalitis (CM), pulmonary cryptococcosis and cryptococcal sepsis, and among these, CM is the most common and serious type of cryptococcosis. It is estimated that 223,100 cases of cryptococcal meningitis are reported worldwide every year (Rajasingham et al., 2017). Cryptococcosis is known as one of the most common opportunistic infections in adults living with HIV (Rajasingham et al., 2017), but it also occurs in non-HIV populations with underlying diseases, such as diabetes, chronic liver disease, kidney disease, lung disease, malignancies, long-term steroid therapy or solid organ transplants, that may affect the immune status (Xiaobo Feng et al., 2008; Liaw et al., 2010; Zhu et al., 2010; Williamson et al., 2017).

For the clinical treatment of cryptococcosis, particularly CM, the combination of amphotericin B (AMB) and 5-flucytosine (5-FC) followed by fluconazole (FLC) maintenance therapy is recommended (Perfect et al., 2010; Liu et al., 2018). Although research on antifungal drug resistance is important, there is no defined clinical breakpoint (CBP) for antifungal agents except AMB (resistant if MIC >1 mg/L). The epidemiological cut-off values (ECOFFs) indicated in the European Committee on Antimicrobial Susceptibility Testing (2022) guidelines for wild-type (WT) C. neoformans are defined as MIC ≤1 mg/L for AMB and MIC ≤0.5 mg/L for posaconazole and voriconazole, and no ECOFFs for other antifungal agents have been established. For C. gattii, the ECOFFs are 0.5 mg/L and 1 mg/L for AMB and posaconazole, respectively, no ECOFFs have been indicated for other antifungal agents, and no CBPs have been reported for any antifungal agents (European Committee on Antimicrobial Susceptibility Testing, 2022). It is very important to further study the mechanisms of resistance to these drugs because this resistance increases the failure rate of induction treatment (Perfect et al., 2010; May et al., 2016). Further studies are needed to aid the establishment of cryptococcal CBPs in the future.

Several molecular typing systems have been applied for Cryptococcus spp. The commonly used methods include genotype (Meyer et al., 2009), serotype (T G Mitchell JRP., 1995), and mating type (Chaturvedi et al., 2000) analyses. The genotype can be determined with many methods. In this study, we used two band-based methods, M13-based PCR fingerprinting (FP) and URA5-RFLP analysis, and two sequence-based methods, MLST and whole-genome sequencing (WGS). The serotype and mating type can be identified by target sequencing using PCR. MLST (Meyer et al., 2009) provides high discriminatory power, has good repeatability across different laboratories, and is inexpensive. Thus, MLST is a major and important method for strain typing in epidemiological studies of Cryptococcus spp. (Meyer et al., 2009). However, the MLST system (Meyer et al., 2009) is not suitable for the molecular typing of hybrids because it fails to amplify certain alleles (Li et al., 2012; Samarasinghe and Xu, 2018; Cogliati et al., 2020). To overcome this problem, new primers were designed to enable the typing of AD hybrids in 2020 (Cogliati et al., 2020).

Multilocus sequence typing was used to determine the internal 469 ~ 723-bp nucleotide sequence of seven housekeeping genes for further typing (Meyer et al., 2009). ST5/VNI is the major molecular type of CNSC in China. Its appearance suggests low genetic diversity because the current genotyping method covers few genes, but the genomes of Cryptococcus neoformans var. neoformans JEC21 and Cryptococcus gattii WM276 are 18.57 Mb and 18.4 Mb, respectively, and include 14 chromosomes (Loftus et al., 2005; D’Souza et al., 2011). Previous analysis showed that ST5 has different subtypes, which was observed following the detection of single nucleotide polymorphism (SNPs) variants (Zhou et al., 2022), and WGS can further be used to analyse resistance mechanisms and simultaneously acquire other typing features (Zang et al., 2022; Zhou et al., 2022). Considering the complexity of Cryptococcus spp. genomes and the development of genome sequencing technology, typing based on WGS has achieved great success in the discrimination of species (Firacative et al., 2016; Wongsuk et al., 2020) and may further reveal the evolutionary relationship and population structure of Cryptococcus spp. isolates.

In previous studies of Cryptococcus spp. in Zhejiang, 98.1% (51/52) of cryptococcemia isolates were identified as C. neoformans var. grubii ST5-VNI-α. Most patients (64.80–71.4%) with cryptococcosis were not infected with HIV, according to data from a general hospital in Zhejiang Province, which is also a designated HIV hospital (Fang et al., 2020; Zhao et al., 2021). To collect more strains to further and more comprehensively explore the clinical characteristics, drug susceptibility and molecular characteristics of Cryptococcus spp. in Zhejiang Province, we collected all preserved strains of clinically isolated Cryptococcus spp. in 4 hospitals in Zhejiang Province from 2012 to 2018, and antimicrobial susceptibility testing (AST), molecular typing, and genomic analysis were performed.

As revealed by previous studies (Zhu et al., 2010; Fang et al., 2020; Fu et al., 2020; Zhou et al., 2020; Zhao et al., 2021), there are many more non-HIV patients in China than in other countries. In our study, we observed the same phenomenon: the proportion of patients with HIV was 12.78%. This finding may be due to the lower prevalence of HIV infection and the implementation of primary FLC prophylaxis in patients with CD4 counts <100 cells/μL in China (Organization WH, 2011; Parkes-Ratanshi et al., 2011; Oladele et al., 2017; Williamson et al., 2017). More importantly, this phenomenon indicates the importance of monitoring non-HIV-associated cryptococcosis even in apparently immunocompetent individuals (Williamson et al., 2017). In the United States, 20% of HIV-negative patients have no underlying conditions, such as a history of steroid therapy or malignant tumours (Williamson et al., 2017). Among non-HIV patients in this study, the presence of diseases affecting the immune system did not appear to increase the risk of Cryptococcus spp. infection since the proportions of patients with and without underlying diseases were nearly equal (45.55% vs. 41.67%). CNS-related infections were the main type of disease observed among HIV-negative patients (67.52%, 106/157). We reviewed the previous reports of Chinese researchers (Zhu et al., 2010) reported the clinical characteristics of 154 patients with cryptococcal meningitis over a 10-year period, the major of patients (66.9%) were otherwise apparently healthy and only 33.1% patients had predisposition factors (Zhu et al., 2010), there are also more reports from China that about 60% of cryptococcosis were from immunocompetent individuals, it is speculated that Chinese population might be more susceptible to cryptococcal infections than other ethnic groups (Fang et al., 2015).

In previous Cryptococcus spp. studies, many typing methods, such as RAPD, PCR fingerprinting, AFLP, MLMT, and MLST, have been employed and have mainly focused on genotyping (Hong et al., 2021). Due to their variable discrimination power, each of these methods has expanded knowledge regarding the genomic diversity of Cryptococcus spp. (Meyer et al., 2009; Hagen et al., 2015; Hong et al., 2021). We used band-based methods with the molecular types VNI–VNIV, VNB of CNSC and VGI-VGV of CGSC to assess this type of genomic diversity. With the development of sequencing technology, sequence base typing methods have provided more discriminatory power, repeatability and comparability. Thus, MLST and WGS are becoming increasingly popular. Using the MLST system, we identified lineages using sequence types. However, because MLST only includes the analysis of seven housekeeping genes, MLST may provide unilateral results, unlike WGS typing. Genome analysis in eukaryotes, including assembly, annotation, SNP calling, and phylogenetic tree reconstruction, is extremely difficult and requires expert bioinformatic knowledge and high-performance computers (Desjardins et al., 2017). In our study, we proved that even with roughly processed genomes, the results presented real genomic diversity and were clearly distinguished at the species level and lineage level. Using genome assemblies, Mashtree was used to reconstruct the phylogenetic tree efficiently. Not only can CNSCs and CGSCs be clearly distinguished, but subclusters in the major cluster can also be classified, as reported by Ziyi Zhou et al. in 2022 in a study that showed that ST5 has different subtypes based on the detection of single nucleotide polymorphism (SNPs) variants (Zhou et al., 2022), and the ST5 cluster and ST31 cluster can be roughly considered CC5 and CC31 (Fan et al., 2016; Cogliati et al., 2019; Thanh et al., 2019). Furthermore, the relationship between AD hybrids and other CCs could be clearly distinguished. These results improved the typing methods and revealed the biodiversity of Cryptococcus spp. in China (Zang et al., 2022; Zhou et al., 2022).

The latest research has shown that the nonsusceptibility rate (intermediate and resistant) of Cryptococcus spp. with FLC reached 25.9% (Simwami et al., 2011), which was significantly higher than the value reported 9 years ago (9.5%) (Wang et al., 2012; Xiao et al., 2018). This finding raises concerns about the increased drug resistance rate. Mutation of the ERG11 gene is the most common mechanism underlying azole resistance. It has been reported that the G484S, Y145F and G470R mutations of ERG11 are related to azole resistance (Rodero et al., 2003; Sionov et al., 2012; Gago et al., 2017; Zhang et al., 2019). However, no mutation was found in ERG11 in these high-MIC strains in our study.

In particular, two strains with high MICs of 5-FC (>32 mg/L) were found in this study: 1 strain of C. gattii and 1 strain of C. neoformans. To date, the drug resistance rate of 5-FC with Cryptococcus spp. is 1–25% (Cuenca-Estrella, 2001; Govender et al., 2011; Espinel-Ingroff et al., 2012; Chang et al., 2021). The metabolic pathway of FCY2-FCY1-FUR1 is related to 5-FC biochemical reactions and may cause 5-FC resistance due to its mutations. The heterozygous G/T mutation at position 145 of the FCY2 gene and the nonsense mutation C505T have been reported. In addition, the T26C point mutation of the FCY1 gene leads to a change in the protein (M9T), which can lead to cross-resistance to 5-FC and FLC (Papon et al., 2007; Florent et al., 2009). Considering the mutations identified in the C. gattii isolate RM009, we assumed that the two nonsynonymous mutations in the FCY1 gene may be responsible for 5-FC resistance, but this finding needs further research. There were substantial differences regarding the C. neoformans isolate (ZY011); no exon mutation was found in the FCY2-4 and FCY1 genes, but the sequence of the FUR1 gene was significantly different from the reference sequence. The identical sequence ended in an exon sequence, and because there were no other ways to identify the sequence of mRNA or protein, we could not predict the actual change in this gene. Considering that the changed region is near the 5-FU-binding site, mutations in the FUR1 gene may cause the binding of pyrimidines to be blocked or inefficient, which may be the main factor underlying resistance.

The MICs of the 5 anti-cryptococcal drugs in this study were generally not high, there was no amphotericin B-resistant strain, most of the Cryptococcus spp. strains were sensitive to azoles, the MICs of several NWTs were generally not high, only two 5-FC-resistant strains were identified, these provides a more detailed reference for the empirical therapy of cryptococcosis, and our study supplements the available information regarding cryptococcal infections in Zhejiang Province for the global fungal database. These findings prove that simply processed genome sequences also provide convincing evidence of lineages.

In mainland China, the non-HIV population is the most commonly susceptible to cryptococcosis, and Cryptococcus neoformans (ST5-VNI-A-α) is the predominant pathogen. The isolated hybrids and CGSCs were rare; of particular interest, three strains of C. gattii, which are rare in the clinic, all caused meningitis and occurred in younger patients without underlying diseases. Phylogenetic analysis based on WGS is not band-based, covers more genomic information and therefore shows higher discriminatory power than traditional typing methods. Anti-cryptococcal drug-resistant isolates were rare, but isolates with high MICs were identified. Mutations in FCY2-FCY1-FUR1 were detected in the 5-FC-resistant isolate, and this finding deserves further study.

However, there are also some limitations in this study. First, data may be outdated or incomplete because many data were collected retrospectively through electronic medical records. Second, the proportion of HIV in underlying diseases would be biased in that only one of the four hospitals in the study is a designated hospital for HIV treatment. Third, due to the limitation of diagnosis methods and the lack of understanding of cryptococcosis in early years, not all Cryptococcus spp. strains in four hospitals were collected at the same time, resulting in sample bias. A better epidemiological investigation design is needed in the future.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

YY, YJ, and ZZ: conceptualization. JZ, QY, YF, FZ, XL, and LF: data curation. JZ and ZZ: formal analysis. YY: funding acquisition. JZ, ZW, and LF: investigation. YC, YJ, and YY: methodology. JZ, ZW, and YY: project administration. ZZ, QY, YF, FZ, QC, and LF: resources. ZW, YC, XL, and YJ: software. YJ: supervision. JZ and ZW: validation and visualization. JZ: writing—original draft. ZW, YJ, and YY: writing—review and editing. All authors contributed to the article and approved the submitted version.

This research was funded by the clinical research project of Zhejiang Provincial Department of Health (grant number 2018ZD030) and General Research Plan for Medical and Health of Zhejiang Province (Class A) (grant number 2020KY604).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.