Detection of Candida auris Antifungal Drug Resistance Markers Directly from Clinical Skin Swabs

Accurate and rapid assessment of Candida auris antifungal drug resistance is crucial for effective infection prevention and control actions, as well as for patient management. Here, performance of a molecular diagnostic platform, enabling rapid identification of FKS1 and ERG11 mutations conferring echinocandin and azole resistance, respectively, was evaluated on a panel of clinical skin swabs. Gene sequencing and antifungal susceptibility testing were used as the gold standard.

ABSTRACT

Accurate and rapid assessment of Candida auris antifungal drug resistance is crucial for effective infection prevention and control actions, as well as for patient management. Here, performance of a molecular diagnostic platform, enabling rapid identification of FKS1 and ERG11 mutations conferring echinocandin and azole resistance, respectively, was evaluated on a panel of clinical skin swabs. Gene sequencing and antifungal susceptibility testing were used as the gold standard. All swabs were correctly categorized as harboring wild-type or mutant C. auris.

TEXT

The recent emergence of drug-resistant Candida auris has become a global health care issue (1). C. auris has a high propensity to colonize human skin and contaminate inanimate surfaces in health care facilities, which results in transmission of clonal C. auris isolates and nosocomial outbreaks (2, 3). Thus, screening for C. auris and prompt identification of colonized individuals and facilities are important components of surveillance for C. auris, since they allow timely implementation of infection prevention and control measures. Disturbingly, control of C. auris spread is constrained by the inadequate testing capacities of the diagnostic laboratories. The problems include misidentification by commonly available diagnostic platforms and delay in diagnosis due to the requirement of the testing of pure cultures (a colony needs to be obtained from the patient specimen, which can take days or even weeks). Antifungal susceptibility testing (AFST) by the broth microdilution method, even though standardized (4, 5), is not implemented everywhere. Moreover, since C. auris is a relatively new pathogen and our understanding of its resistance to antifungal drugs is limited, the interpretation of AFST results for this yeast lacks detailed guidelines. Even though tentative MIC breakpoints are provided by the CDC (https://www.cdc.gov/fungal/candida-auris/c-auris-antifungal.html), AFST still requires some level of expertise, since the echinocandin Eagle effect (paradoxical growth effect) and azole trailing growth effect complicate C. auris MIC reading and may lead to an erroneous susceptibility category assignment (6).

Molecular-based assays that profile drug resistance-associated target genes provide direct information about potential drug resistance relevant to therapy. They enhance epidemiological surveillance capabilities by offering shorter turnaround time, higher testing throughput, and an evaluation of the prevalence of resistance mechanisms. Given that azole and echinocandin resistance in C. auris are closely associated with specific ERG11 and FKS1 mutations (6–10), we developed a molecular diagnostic platform, employing asymmetric real-time PCR in conjunction with allele-specific molecular beacons (MB) (11), to quickly identify them. Simply, wild-type (WT) and mutant (non-WT) amplicons can be differentiated based on their melting temperature. Here, we validated a C. auris real-time PCR assay for the detection of antifungal drug resistance markers in clinical skin swabs by a direct comparison with gold standard methods involving targeted gene sequencing and antifungal susceptibility testing by CLSI broth microdilution.

We processed 112 clinical axilla/groin composite swabs (BD Eswab, liquid Amies medium; BD Diagnostics) in an enrichment protocol before plating onto CHROMagar Candida (12). As a result, 27 samples were culture positive, with beige-to-mauve colonies growing on CHROMagar plates. DNA was extracted from these colonies with a quick boiling method (13), and species identification was performed by sequencing of the ribosomal DNA (rDNA) region, amplified with Fun-rDNAF (5′-GGTCATTTAGAGGAAGTAAAAGTCG-3′) and Fun-rDNAR (5′-YGATATGCTTAAGTTCAGCGGGTA-3′) primers (S. Katiyar, personal communication), and further nucleotide BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) analysis. We identified 24 isolates (21%) as C. auris and 4 isolates as Candida parapsilosis (one swab/plate contained both C. auris and C. parapsilosis). These results confirmed that processing of skin swabs in an enrichment broth does not completely exclude growth of species other than C. auris. C. auris colonies and those of other species (e.g., C. parapsilosis) may look alike on CHROMagar Candida. Hence, recovery of beige-to-mauve colonies can never be considered definite identification of C. auris (14).

Antifungal susceptibility testing with echinocandins (anidulafungin and micafungin) and azoles (fluconazole and voriconazole) for the recovered C. auris isolates was performed according to CLSI guidelines (4). All isolates exhibited low MIC values (MIC ranges: anidulafungin, 0.06 to 0.125 mg/liter; micafungin, 0.06 to 0.125 mg/liter; fluconazole, 1 to 2 mg/liter; voriconazole, ≤0.03 mg/liter) and were categorized as echinocandin and azole susceptible. Moreover, sequencing of amplified FKS1 and ERG11 genes (6, 10) revealed that all recovered isolates presented WT FKS1 and ERG11 genotypes (Table 1).

TABLE 1.

Results of the molecular resistance assay matched with the results of FKS1 and ERG11 gene sequencing and antifungal susceptibility testing

Sample	Quantity	Real-time PCR assay result						Gene sequence		MIC range (mg/liter)b
		FKS1 HS1		ERG11 Y132		ERG11 K143
		Tma range (°C)	Genotype	Tm range (°C)	Genotype	Tm range (°C)	Genotype	FKS1	ERG11	ANF	MCF	FLC	VRC
Clinical swabs	24	64.53–65.34	WT	60.03–60.85	WT	51.90–52.70	WT	WT	WT	0.06–0.125	0.06–0.125	1–2	≤0.03
Spiked swabs	4	58.99–59.10	S639F					S639F		4 to >16	16
	6			60.36–60.60	WT	58.03–58.49	K143R		K143R			32 to >128	0.25–2
	4			55.58–55.80	Y132F	52.11–52.18	WT		Y132F			>128	4

^aT_m, melting temperature.

^bANF, anidulafungin; MCF, micafungin; FLC, fluconazole; VRC, voriconazole.

DNA was extracted from 100 μl of clinical swab medium by using the Quick-DNA Fungal/Bacterial Miniprep kit (Zymo Research), eluted in 40 μl, and stored at −20°C postextraction. A SYBR green C. auris-specific real-time PCR assay was used to determine the presence of C. auris DNA in the swabs (15, 16). In the assay, 55% (62/112) of swabs were C. auris positive and 45% (50/112) of swabs were C. auris negative. Comparison of the real-time PCR results with the results of the enrichment culture led to the realization that C. auris cells most likely lost their viability in 34% (38/112) of swabs, but the DNA was still present (giving a positive real-time PCR result), which was possibly due to the long swab storage (4 to 5 months) before processing. Detection of C. auris was not affected by the presence of other microorganisms (skin flora) present in the sample.

To better evaluate assay performance, since all of the recovered C. auris isolates were categorized as FKS1 and ERG11 WT, swab medium, representative of the skin flora background (created by pooling aliquots from 20 C. auris-negative clinical swabs), was spiked with 4 × 10⁵ CFU/ml of the following C. auris drug-resistant isolates (representing clades I and IV): (i) FKS1 S639F ERG11 K143R, DPL 1359, DPL 1371, DPL 1373, and DPL 1442; (ii) FKS1 WT ERG11 K143R, DPL 1343 and DPL 1422; (iii) FKS1 WT ERG11 Y132F, DPL 1349, DPL 1367, DPL 1410, and DPL 1419. DNA from 24 culture- and real-time PCR-positive swabs and 10 spiked swabs was used to evaluate performance of the C. auris resistance markers real-time PCR assay (11). DNA isolated from DPL 1349 (ERG11 Y132F), DPL 1350 (ERG11 K143R), DPL 1373 (FKS1 S639F), and DPL 1384 (FKS1 WT ERG11 WT) C. auris strains was used as an allele melting temperature (T_m) reference standard. The assay was performed in a blinded fashion, such that the researcher running the samples was different from the one selecting samples for screening and had no knowledge of the C. auris genotypes. The C. auris resistance markers real-time PCR assay identified correctly all 24 C. auris culture- and real-time PCR-positive swabs as harboring FKS1 HS1 wild-type and ERG11 wild-type C. auris and all spiked swabs as harboring non-WT alleles: FKS1 S639F (4 swabs) and ERG11 K143R (6 swabs) and Y132F (4 swabs) (Table 1). The molecular diagnostic results from the assays were 100% concordant with the results of FKS1 and ERG11 sequencing and AFST. The cross-reactivity screens were automatically included in the tests, since we used clinical swabs that contained all kinds of microorganisms found on patient skin. The same applies to the spiked swabs, since they were prepared in swab medium representative of skin flora background. No unexpected/unspecific melting curves were obtained throughout the tests.

For limit of detection (LoD) determinations, swab medium representative of skin flora background was spiked with 4 × 10² CFU/ml to 4 × 10⁵ CFU/ml of DPL 1343 (FKS1 WT), DPL 1359 (FKS1 S639F), DPL 1419 (ERG11 Y132F), and DPL 1422 (ERG11 K143R). Genotyping confidence for FKS1 and ERG11 was obtained at the level of 10⁵ CFU/ml and 10⁴ CFU/ml, respectively (Table 2), which ensures reliable detection, since C. auris burdens on patient skin swabs are often higher, with up to 10⁸ CFU on a single swab (17).

TABLE 2.

Results of limit of detection determination experiments with 4 × 10² CFU/ml to 4 × 10⁵ CFU/ml of DPL 1343 (FKS1 WT), DPL 1359 (FKS1 S639F), DPL 1419 (ERG11 Y132F), and DPL 1422 (ERG11 K143R) C. auris strains prepared in representative skin flora background

Swab burden (CFU/ml)	CFU/reaction	Real-time PCR assay results (Tm [°C])a
		FKS1		ERG11 Y132		ERG11 K143
		WT	S639F	WT	Y132F	WT	K143R
4 × 105	1,000	+ + (65.62, 65.39)	+ + (59.02, 59.24)	+ + (60.47, 60.54)	+ + (55.76, 55.63)	+ + (52.17, 52.17)	+ + (58.03, 58.28)
4 × 104	100	+ + (65.71, 65.46)	+ − (59.24)	+ + (60.74, 61.13)	+ + (55.38, 55.82)	+ + (52.40, 52.34)	+ + (58.59, 58.76)
4 × 103	10	+ + (66.04, 65.25)	− −	− −	+ − (55.84)	+ + (52.40, 52.42)	− −
4 × 102	1	+ + (65.84, 65.67)	− −	− −	− −	− −	− −

^aSamples run in duplicates: +, positive result; −, negative result.

In summary, the C. auris resistance markers real-time PCR assay performed well in categorizing clinical axilla/groin skin swabs (original and spiked) as harboring WT or mutant C. auris. This assay can serve as a high-throughput diagnostic alternative, dramatically reducing the time needed to obtain information on isolate resistance genotype. The assay setup and readout do not require special expertise, since the thermocycling conditions can be saved and readily available anytime the assay needs to be performed. Interpretation of the assay results is straightforward: an unknown sample can be easily categorized as WT or mutant by comparing its amplicon T_m value with the reference T_ms. The assay is robust and can be conveniently expanded to enable novel mutation detection.