Whole-genome sequence typing: A New molecular tool for uncommon fungal outbreak investigations
Whole-genome sequence (WGS) typing is a new molecular approach that enables genotyping of any microorganism without genetic insights or prior knowledge of the natural population diversity in that species. Thus, it can be used when no other conventional method is available for molecular genotyping. Indeed, when an outbreak is caused by an uncommon fungal species there is usually little epidemiological information available regarding dissemination, genetic diversity, and/or population structure for the species; standard methods for molecular genotyping are often not suitable and lack comprehensive resolution. For these unusual fungal pathogens, WGS typing can be highly valuable for allowing first molecular comparisons of strains [
32]. WGS typing is based on the analysis of whole single-nucleotide polymorphisms (SNPs) within each genome, which allows investigating genome-wide variation between isolates. The relationships between the isolates can then be inferred from the number of SNPs differing between them. SNP distances determine genetic relatedness. Another advantage of WGS typing is that it permits the production of “de novo assembly” for these unusual pathogens lacking published reference genomes. Whole-genome sequencing is also useful for getting information characterized important traits, such as the level of ploidy, the size of assembled genome, etc.
WGS typing has a high discriminatory power to relate strains. However, its current limitation is that it must be performed in a laboratory in which the different steps of the process are well established, including sequence acquisition and bioinformatics for handling and analysing the data [
33]. Considerable challenges remain due to the lack of common standards, particularly for investigating fungal genomes because they are much larger than those of bacteria and viruses and have a variable level of ploidy.
Recent years have been rich in large fungal outbreaks caused by very unusual fungal species (Table
1). These outbreaks, due to either yeasts or mold, have signaled the beginning of a new era of fungal outbreak investigation based on the comparison of the whole genomes of outbreak strains by using WGS. WGS typing permits investigation into several fungal outbreaks due to uncommon fungal species allowing researchers to decipher epidemiological clues [
34‐
36].
The largest fungal outbreak of healthcare-associated infections investigated using WGS typing was caused by the black soil mold
E. rostratum, responsible for a number of cases of meningitis in the USA. Owing to epidemiological investigations, the use of MPA injections contaminated with
E. rostratum was quickly demonstrated to be responsible for the outbreak [
37,
38]. Different lots of MPA from a single compounding pharmacy were positive with
E. rostratum. However, it was not possible to demonstrate the genetic relationship between the fungal isolates due to the lack of relevant typing methods for that given species. WGS typing was eventually used to analyse and compare the genomes of both the clinical isolates and those from MPA vials originating from different lots [
39]. It initially demonstrated that the isolates were highly clonal, strongly suggesting a single common source. Secondly, it demonstrated that clinical isolates could not be distinguished from those from the MPA vials, confirming that MPA was the source of infection. Interestingly, WGS analysis also showed that these
E. rostratum isolates were haploid and nearly identical, only differing by less than 2 SNPs between any pair of isolates. In contrast, hundreds of thousands of SNPs differentiated the isolates responsible for the outbreak and unrelated control isolates. These huge differences between unrelated and related genome strains unearths interesting questions and may indeed suggest that the unrelated control strains belonged to different cryptic species. Nonetheless, this study highlighted the power of the WGS typing to provide accurate information about the relatedness of the isolates of rare species, here
E. rostratum, when no genetic information was available.
The same approach was then used to investigate the healthcare-associated outbreak of
S. kiliense bloodstream infections associated with the administration of contaminated antinausea medication among oncology patients in Colombia and Chile [
24]. Once again, WGS typing confirmed a common source of infections in the two countries related to the contaminated medication. Indeed, genomes of isolates from infected patients of the two countries and from the different lots of contaminated medication were nearly indistinguishable from one another. No more than 5 SNPs were detected between any pair of isolates; for reference the genome size of
S. kiliense is approximately 36 MB.
In France,
S. clavata, another previously unrecognized fungal pathogen, was implicated for different clusters of infections in leukemic patients [
18]. The use of WGS typing identified a single clone for most cases. However, it was not possible to connect it to a common source of infection.
An interesting secondary outcome of these three studies was that a large panel of assembled genomes from the three species was produced for the first time. Some of them were from related strains, including genomes from the outbreak strains, while others were from unrelated control strains, from the same or different countries. These genome sequences served as the basis for constructing databases to study genetic variations in these unusual species, which had never been done before. Thus, in addition to their power for genotyping, WGS data generated during outbreak investigations can also to be used to better understand the biology, the virulence, and even possible therapeutic targets in these species.
Recently, the nearly simultaneous global emergence of a novel species of
Candida,
C. auris, capable of causing hospital-acquired multidrug-resistant infections, has likely represented one of the more complex challenges in terms of epidemiological investigation. In 2009, when
C. auris was first isolated from the external ear canal in a Japanese hospital patient, no information about this species or its putative reservoir was available. Since the publication of the first
C. auris genome sequence in 2015 [
40], a large number of
C. auris strains have been sequenced for investigating different outbreaks. Their analysis suggests that the pandemic is due to recent and independent emergences of different clonal populations of
C. auris, rather than to the worldwide spread of a dominant clone. Indeed, phylogenetic analysis of the whole genome SNPs from strains isolated from three continents showed that there is a strong phylogeographic structure, with the strains grouping into unique clades by geographic regions [
14]. The strains from the same geographic region, such as South Asia, South America, or South Africa, were highly clonal and belonged to a specific clade, with very few SNPs differences (16-70 SNPs). In contrast, tens of thousands of SNPs differentiated strains from different continents.
In addition, recent investigations showed that in a given country, strains isolated from patients hospitalized in different hospitals or towns can be virtually identical (only a few SNP differences) allowing precise tracking of clonal spreading in the different clusters. Finally, a recent study showed almost identical (less than 5 SNP differences)
C. auris isolates from the surface environment of one hospitalized patient and his own isolates, which further suggested that the spread within health care settings was possible [
41]. Overall, WGS exclusively provided unambiguous data on the genome sequences and on the number of SNPs differentiating isolates. Those can easily be shared between laboratories but also centralized for a global investigation of the pandemic [
42]. As mentioned above for other uncommon species, these data have also been used to further study genetics of the biology and the virulence of
C. auris [
43].
From infection control point of view
The various fungal outbreaks described within the last 6 years have highlighted not only the importance of general infection prevention measures but also the importance of warning alert systems. Indeed, the specificity of these outbreaks emphasizes the value of monitoring systems. Concerning
C. auris, the risk of misidentification was high and it has been suggested that
C. auris should be suspected in several situations, such as: a) when identification cannot be obtained using traditional biochemical mycological methods, and b) when resistance to more than one antifungal drug is detected for an isolate with an ambiguous identification [
44]. Furthermore, concerning the other described species such as
S. clavata, E. rostratum, and
S. kiliense, the published studies highlighted several points as the common source of contamination (
E. rostratum and
S. kiliense) and the specific populations at risk. It seems fundamental to introduce warning alert systems that could help practitioners and health authorities to react quickly to prepare in the event of an outbreak. These warning alerts could be implemented in specific populations e.g., immunosuppressed patients.
Given the risk of nosocomial transmission of
C. auris, it is necessary to promptly implement infection control measures to limit its spread. Similar to other
Candida species, colonized and infected
C. auris patients share the same risk factors, including diabetes mellitus, abdominal surgery, broad-spectrum antibiotics, and presence of CVCs [
17]. Healthcare acquired infections typically occur several weeks (10–50 days) into a patient’s hospital stay [
6].
Hand transmission and persistent contamination of environmental surfaces within healthcare facilities have been associated with outbreaks; thus, CDC recommendations are based on standard and contact precautions [
45]. As for other
Candida species, transmission could occur through healthcare workers (HCW) hands; however unlike others HCW seem not to be colonized long term. Indeed, as suggested by two different studies, carriage seems to be rare. The most recent study [
46] highlighted that during an outbreak only four among 145 HCW were hand carriers. However systematic sampling of the hands, nose, axilla, groin, and throat of 258 HCW conducted as part of the UK investigation identified only a single HCW with a positive nares swab, who later tested negative from the same site, suggesting transient carriage [
15].
All hospitalized patients with
C. auris infection or colonization should be housed in single rooms with environmental precautions. During outbreak investigations,
C auris has been isolated from medical equipment likely originating from patient’s skin shedding [
47]. Explaining why considerable attention has focused on the efficacy of disinfectants used for skin decolonization and environmental decontamination.
Indeed, several authors suggested that contaminated surfaces in healthcare facilities may be an important source of acquisition [
48‐
50]. Experimentally,
C. auris and other
Candida species persisted for 7 days or up to 4 weeks on moist and dry surfaces, respectively [
51]. Moreover,
C. auris exhibited a greater propensity to survive on surfaces than
C. albicans, but not
C. parapsilosis or
C. glabrata. In comparison to common bacterial pathogens,
Candida species were recovered with similar frequencies from dry surfaces and were recovered significantly more often from moist areas such as sinks. Moreover,
C. auris has the capacity to form antifungal resistant biofilms sensitive to the disinfectant chlorhexidine in vitro [
52]. Correspondingly, during an outbreak in London, Schelenz et al. [
15] implemented extreme environmental decontamination measures for cleaning and disinfection of the patient rooms and equipment using 1000 ppm chlorine-based products three times a day. On discharge or transfer of a
C. auris positive patient, the room was subjected to terminal cleaning with 10,000 ppm chlorine-based detergent and all cleaned equipment was left in the room to be disinfected with hydrogen peroxide vapour.
The CDC recommends thorough daily and terminal disinfection of room surfaces and shared medical equipment in rooms of patients with
C. auris infection (
https://www.cdc.gov/fungal/diseases/candidiasis/c-auris-infection-control.html). Although many disinfectants are registered as disinfectants against
Candida species by the Environmental Protection Agency, it is recommended that a disinfectant effective against
Clostridium difficile spores should be used, such as chlorine-based products.
Several recent studies evaluated the efficacy in vitro of disinfectants used for skin decolonization and environmental decontamination during hospital outbreaks [
53‐
55]. All showed differences between products; activity also varied with different formulations. Indeed, chlorhexidine gluconate, iodinated povidone, chlorine and H
2O
2 vapour demonstrated effective killing activity against
C. auris when used in clinical practice. However, among disinfectants utilised for skin decolonization, chlorhexidine was much less inhibitory at 3 min contact time compared to iodinate povidone used at 10% [
54]. Similarly, the widely used quaternary ammonium disinfectants had a relatively poor activity against all
Candida species [
55].
Limiting the spread of
C. auris needs rapid multifaceted interventions (Table
2) including contact isolation, enhancing environmental disinfection, screening contact populations, and skin decolonization for colonized patients. However, there is uncertainty as to how best monitor prolonged colonization (
https://www.cdc.gov/fungal/diseases/candidiasis/c-auris-infection-control.html). There are no clear data on the efficacy of decolonization measures for patients colonized with
C. auris, although chlorhexidine has been tested for that purpose during outbreaks [
15,
49].
Table 2
Interventions needed in case of Candida auris outbreak
1-Notify public health agency | Undoubted |
2-Place patient (colonized or infected) in a single room | Undoubted |
3-Institute Contact Precautions for colonized or infected patients | Undoubted |
4-Screen all contact patients (defined as roommates) once a week and before leaving the medical ward | Uncertainty how best to monitor |
5-Reinforce environmental cleaning 3× day with 1000 ppm chlorine based, vaporized H2O2 | Undoubted |
6-Reduce duration of invasive procedures in colonized patients | Undoubted |
7-Skin decolonization (colonized patients) with 10% w/w iodinated povidone | No clear data |