Molecular Characterization of Candida parapsilosis by Microsatellite Typing and Emergence of Clonal Antifungal Drug Resistant Strains in a Multicenter Surveillance in China

Candida parapsilosis is an important species causing invasive candidiasis (IC) in China. The present survey was a national multicenter study of the molecular epidemiology and antifungal susceptibility profiles of C. parapsilosis. Non-duplicate C. parapsilosis isolates were collected from 10 hospitals across China in the CHIF-NET program 2016–2017. Isolates were genotyped using four highly polymorphic microsatellite markers, and susceptibility profiles determined using Sensititre YeastOneTM YO10. A total of 319 C. parapsilosis from separate patients with IC were studied; 49.2, 17.9, and 10.3% isolates were from patients in surgical departments, general intensive care units (ICUs) and neonatal ICUs (NICU), respectively. C. parapsilosis showed good susceptibility to nine antifungal drugs. Microsatellite analysis identified 122 microsatellite (MT) types. Most MT types had sporadic distribution. However, we identified 32 clusters across 10 hospitals; seven clusters were caused by seven endemic genotypes involving five or more isolates in hospitals designated as H01, H02, H06, and H10. These clusters mainly affected surgical departments and ICUs, except for genotype MT42 which was seen in 22 patients from NICU (hospital H06). Of 16 fluconazole-resistant isolates, seven from hospital H02 shared the same genotype MT70, and three from hospital H04 were of genotype MT47. For 37 isolates with non-wild type MICs to 5-flucytosine, 29 were from hospital H01 (genotype MT48). Here we present the first nationwide molecular epidemiology study of C. parapsilosis in China, identified several previously unrecognized clusters, which included antifungal drug resistant isolates. These findings provide important data for control of IC in China.


INTRODUCTION
Invasive candidiasis (IC) is an important fungal disease among hospitalized patients, associated with significant mortality and excessive medical costs (Kullberg and Arendrup, 2015). Although Candida albicans still overall accounts for the majority of IC, many studies have reported increased prevalence of infections caused by non-C. albicans species (Chow et al., 2008;Silva et al., 2012;Pfaller et al., 2014). Particularly, the incidence of C. parapsilosis infection is increasing among hospitalized patients and has become one of the main pathogens causing candidemia in certain geographical areas including China (Guinea, 2014;Pfaller et al., 2014;Puig-Asensio et al., 2014;Strollo et al., 2016). The National China Hospital Invasive Fungal Surveillance Net (CHIF-NET) program is an ongoing nationwide, multicenter surveillance network established in July 2009 to provide data on the epidemiology of invasive fungal infections in China (Wang et al., 2012;Xiao et al., 2015Xiao et al., , 2018. In the period from 2010 to 2015, C. parapsilosis was the second most common species isolated from bloodstream after C. albicans (32.3%), accounting for 28.9% of cases (Xiao et al., 2018).
Based on the unexpected observation of potential case clusters in the CHIF-NET study above, the present study aimed to present the findings of the molecular epidemiology of C. parapsilosis sensu stricto isolates obtained from patients from 10 hospitals surveyed between August 2015 and July 2017. We chose to use a four-locus microsatellite-based typing method to genotype C. parapsilosis sensu stricto isolates described by Sabino et al. (2010) as this method has proven to be of good discriminatory (power of 99.9%), and is a common tool for genotyping C. parapsilosis sensu stricto (Vaz et al., 2011;Romeo et al., 2013;Delfino et al., 2014;Sabino et al., 2015). We used these microsatellite markers to distinguish isolates from all participant hospitals, aiming to uncover the endemic genotypes in different hospitals, and to call genetic association to support epidemiologically linked isolates in case clusters.
All strains were forwarded to a central laboratory (The Department of Clinical Laboratory, Peking Union Medical College Hospital) for confirmation of species identification by using the Vitek MS system (BioMerieux, France) selectively supplemented by ribosomal DNA sequencing as required (Wang et al., 2012;Zhang et al., 2014). The program was approved by the Human Research Ethics Committee of Peking Union Medical College Hospital (S-263).

Antifungal Susceptibility Testing
In vitro antifungal susceptibility to nine antifungal drugs was performed using the Sensititre YeastOne TM YO10 system (Thermo Scientific, Cleveland, OH, United States) following the manufacturer's instructions. The quality control strains included C. parapsilosis ATCC 22019 and Candida krusei ATCC 6258. MIC values of anidulafungin, micafungin, caspofungin, voriconazole, and fluconazole were interpreted according to current Clinical and Laboratory Standards Institute (CLSI) M60 clinical breakpoints (CLSI, 2017). Where there are no CLSI breakpoints, species-specific epidemiology cutoff values (ECVs) were used to define isolates as wild type (WT) or non-wild type (NWT) for amphotericin B, posaconazole, itraconazole, and flucytosine (Pfaller and Diekema, 2012;CLSI, 2018).

Microsatellite Typing
Genotyping of all C. parapsilosis isolates was performed using a panel of four highly polymorphic microsatellite markers as described by Sabino et al. (2010), namely B5, CP1, CP4, and CP6. Amplification reactions and allele sizes analysis were performed as previously reported (Wang et al., 2016). The genetic relationships between the genotypes were studied by constructing a minimum spanning tree using the BioNumerics software v7.6 (Applied Maths, Sint-Martens-Latem, Belgium), treating the data as categorical information. Genotypes showing the same alleles for all four markers were considered identical. Endemic genotypes were defined as genotypes infecting ≥2 different patients in one hospital. A cluster was defined as a group of ≥2 patients infected by an endemic genotype (Escribano et al., 2013).

Clinical Characteristics of Patients
The characteristics of patients and of Candida species at the different hospitals are shown in Table 1. For the 319 isolates, 72.7% were cultured from male patients, and 27.3% from female patients. Cases of IC included candidemia (n = 230), abdominal Candida infection (n = 61), central nervous system infection (n = 6), and other invasive infections (n = 22). Generally, the numbers of isolates increased with age, with 38.6% isolates from patients >60 years. In the present study, 49.2, 11.6, and 17.9% of isolates were collected from patients located in surgical departments, medicine departments and ICUs, respectively; 2.5 and 10.3% isolates were from patients in the pediatric ward and NICUs, respectively.

Antifungal Susceptibility of C. parapsilosis Isolates
The antifungal susceptibilities of the isolates are shown in Table 2.
Microsatellite Typing (MT) of C. parapsilosis A total of 122 MT types for 319 isolates were identified (Supplementary Table S1). There were differences in the discriminatory power of four alleles -allele B5 (n = 29, 23.8%), CP1 (23, 18.9%), CP4 (59, 48.4%), and CP6 (60, 49.2%). Among the 122 MT types, 83 genotypes were distributed sporadically. The remaining 29 genotypes were identified within 32 clusters -MT81 formed four clusters in four different hospitals, with the remaining 28 clusters encompassing different genotypes) (see Table 3). In hospital H01, there were seven clusters, involving a total of 91 isolates. In H06 and H10, there were also endemic genotypes involving large number of isolates, like MT29 (n = 38) and MT42 (n = 22) ( Table 3). The sampling interval within each endemic genotype in the same ward differed with a wide range (1-566 days). The detailed timeline and the ward information for seven endemic genotypes involving five or more isolates are shown in Figure 1.
The relationship between genotypes was shown in the minimum spanning tree in Figure 2. Figure 2 shows the genetic relationship according to different hospitals where most genotypes in each hospital are sporadically distributed. The endemic genotypes in hospital H01 were genetically closely related. The isolates from most cluster cases shared the same susceptibility pattern, except for endemic genotype MT52, one isolate from H01 showed NWT MICs to 5-flucytosine, while the remaining five isolates had WT MICs. In hospital H01, there were 29 isolates belonging to genotype MT 48, all which exhibited high MIC (>64 µg/mL) to 5-flucytosine. Of a total of 16 fluconazole-resistant isolates, seven were from hospital H02 (all genotype MT70), and three were from hospital H04 hospital (genotype MT47).

DISCUSSION
Candida parapsilosis has been reported to be a significant clinical pathogen, being able to form biofilms on central venous catheters and other medically implanted devices, and which can persist in the hospital environment. This last characteristic gives propensity for this fungus to cause nosocomial case clusters via carriage on hands of staff (Trofa et al., 2008;van Asbeck et al., 2009;Toth et al., 2019). In the present study, we used microsatellite genotyping method to study the molecular epidemiology of C. parapsilosis, cultured from patients with invasive fungal infections, as part of the national surveillance program CHIF-NET 16-17. We chose to use microsatellite typing as this method has emerged as a powerful tool to study the genetic relationship between C. parapsilosis isolates in the outbreak setting (Vaz et al., 2011;Diab-Elschahawi et al., 2012;Pulcrano et al., 2012;Reiss et al., 2012;Romeo et al., 2013;Delfino et al., 2014;Sabino et al., 2015;Wang et al., 2016;Badali et al., 2017;Magobo et al., 2017;Desnos-Ollivier et al., 2018).
Data for 319 patients were analyzed; of these 230 patients suffered from candidemia. The patients were mainly adults from surgical departments and ICUs, except for the patients in hospital H06 hospital (mainly NICU). By microsatellite typing, we uncovered 32 cluster cases in 10 hospitals during a period of 2 years. The timeline of cluster cases indicates that certain endemic genotypes have persisted in the hospital setting, and presumably were difficult to eradicate, consistent with previous reports (Escribano et al., 2013(Escribano et al., , 2018. Wang et al. (2016) previously reported details of a C. parapsilosis outbreak in China, where 97 isolates comprising only two clones were mainly obtained from the ICU and surgical wards. Other C. parapsilosis outbreaks in ICUs have been reported from countries such as Brazil and Turkey (Dizbay et al., 2008;Pinhati et al., 2016). Notably, in the present study, strains of the MT42 genotype were responsible for a 22-case cluster in the NICU of hospital H06 (Figure 1); previously, case clusters in NICUs have been uncommonly reported in China. C. parapsilosis has been recognized as a significant pathogen in NICUs worldwide (Toth et al., 2019). A meta-analysis of neonatal candidiasis by Pammi et al. (2013) showed that C. parapsilosis comprised 33.5% of all neonatal Candida infections, and was associated with 10% mortality. Not unexpectedly, in the 24 patients from whom C. parapsilosis isolates were cultured simultaneously from blood and catheter tips, the isolates all exhibited the same genotype, consistent with the widely accepted notion that vascular catheters often represent the source of candidemia and the meticulous catheter care is important for source control (Escribano et al., 2018).
Notably, there were fluconazole resistant isolates responsible for the case clusters in the present study, particularly in two hospitals (three isolates from patients in hospital H04 and seven from patients in H02 hospital). Others have reported clonal emergence or outbreaks of fluconazole-resistant isolates of C. parapsilosis (Pinhati et al., 2016;Thomaz et al., 2018;Singh et al., 2019). Hence such clusters should signal an alert to clinicians and microbiologists alike. Raghuram et al. have reported that fluconazole-resistant strains of C. parapsilosis may be more pathogenic than fluconazole-susceptible strain in liver transplant patients (Raghuram et al., 2012). In another report, a patient infected with azole-resistant C. parapsilosis died despite appropriate antifungal treatment (Zhang et al., 2015). Overall, however, only 5% of C. parapsilosis isolates were resistant to fluconazole in the present study compared with 3.9% in the SENTRY program (Pfaller et al., 2019). Despite being overall uncommon, resistance of C. parapsilosis to fluconazole has been increasing reported, and the data from SENTRY Program suggest that fluconazole-resistant C. parapsilosis could emerge in the presence of drug pressure during fluconazole treatment and prophylaxis, with subsequent transmission between patients in the hospital environment (Pfaller et al., 2019).
There were some limitations to this study. First, information of clinical characteristics, risk factors, and outcomes of patients with C. parapsilosis infections and those associated with the cluster cases were not available. Second, the study was performed retrospectively, and we could not collect the isolates from the hospital environment or health care workers' hands, thus it was not possible to identify the mode of transmission. Third, there was no data about patients transferring between different wards within the hospitals, as well as the rotation of healthcare workers between different departments, and therefore this limited the explanation as to why there were certain endemic genotypes existing in different areas within any one hospital. Further work is needed to investigate the clinical and microbiological consequences of the current cluster cases, especially in hospitals H01, H06, and H10.
Nonetheless the present study was instructive in that it brought to light, several nosocomial cluster cases of C. parapsilosis infection including those involving azole-resistant strains. Lack of awareness of healthcare-associated infections and transmission, particularly those caused by fungi, is common in hospitals in China. Increasing candidemia morbidity and emergence of azole-resistant Candida species serve to emphasize the attention required for effective management of Candida infections including strategies for infection control.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.