Candida auris Isolates of the Southern Asian and South African Lineages Exhibit Different Phenotypic and Antifungal Susceptibility Profiles In Vitro

Candida auris is a serious nosocomial health risk, with widespread outbreaks occurring in hospitals worldwide. Sequence analyses of outbreak isolates revealed that C. auris has simultaneously emerged as four distinct continentally restricted clonal lineages.

ABSTRACT

Candida auris is a serious nosocomial health risk, with widespread outbreaks occurring in hospitals worldwide. Sequence analyses of outbreak isolates revealed that C. auris has simultaneously emerged as four distinct continentally restricted clonal lineages. We previously reported multiple independent introductions of C. auris isolates from at least three of these lineages (the Southern Asia, South African, and Japanese/Korean lineages) into hospitals across the United Kingdom and that isolates circulating in the United Kingdom displayed two different cell phenotypes which correlated with differences in virulence in Galleria mellonella wax moths. Here, we compared the phenotypic characteristics and antifungal susceptibilities of isolates representative of the three geographic clades circulating in the United Kingdom. Isolates of the South African and Japanese/Korean lineages, but not those of the Southern Asian lineage, grew well on media containing actidione. However, unlike Southern Asian lineage isolates, they were unable to produce even rudimentary pseudohyphae in culture. Importantly, although all isolates were fluconazole resistant in vitro, fluconazole and voriconazole exhibited significantly higher MICs against isolates of the South African lineage than against isolates of the Southern Asian lineage. A similar trend was seen with minimum fungicidal concentrations (MFCs), with higher MFCs of the triazole antifungal agents being seen for the South African lineage isolates. Finally, the formation of large cellular aggregates was seen only with isolates of the South African and Japanese/Korean lineages, which correlates with the reduced virulence observed previously in Galleria wax moths inoculated with such isolates. Intriguingly, aggregation could be reversibly induced in isolates of the Southern Asian lineage by exposure to triazole and echinocandin antifungals but not by exposure to amphotericin B or flucytosine.

INTRODUCTION

The incidence of systemic infections caused by non-Candida albicans Candida spp. continues to rise, in part due to the increased numbers of immunocompromised patients and those undergoing invasive procedures (1–8). In 2009, Candida auris, a novel member of the Candida haemulonii complex, was described from the discharge from a human external ear canal in Japan (9), an association with chronic otitis media that was confirmed from South Korean studies (10). C. auris has subsequently been reported from diverse clinical manifestations, ranging from colonization and mucosal infections to deep-seated infections and candidemia (11–15).

Today, C. auris is recognized as an emergent nosocomial pathogen with evidence of clonal inter- and intrahospital transmission and is widespread across several Asian countries, South America, and South Africa (11–19), with additional outbreaks being seen in several European countries and parts of the United States (reviewed in reference 20). Whole-genome sequence comparisons of C. auris isolates from different geographical regions demonstrated that closely related but distinct clonal lineages predominate on different continents (10–13, 16, 21), with widespread resistance to fluconazole and sporadic resistance to various other classes of antifungal agents being detected (11, 13–23). Recently, we demonstrated that the C. auris isolates circulating in the United Kingdom have been independently introduced from diverse geographic origins, with representatives of the Southern Asian (Indian), South African, and Japanese/Korean lineages all being present in UK health care facilities (24). Additionally, although C. albicans is accepted as being the most pathogenic Candida species (25, 26), previous studies have suggested that at least some isolates of C. auris exhibit pathogenicity comparable to that of C. albicans in in vivo models (27, 28).

Since 2013, the UK National Mycology Reference Laboratory (MRL) has received over 240 independent patient isolates of C. auris from over 20 different hospitals, including a number of isolates suspected of being parts of local outbreaks. In addition, a number of isolates were referred directly from countries where C. auris is endemic or were recovered from UK patients repatriated from affected hospitals in India or South Africa. Here, we have used representative examples of this extensive panel to examine the phenotypic behaviors of isolates from the Southern Asian (Indian), South African, and Japanese/Korean clades recovered from the United Kingdom (24), including their growth characteristics and antifungal susceptibility profiles. Our results reveal consistent differences in C. auris behavior which are strictly clade specific.

MATERIALS AND METHODS

Fungal strains, identification, and culture.

The MRL collection includes 175 isolates of the Southern Asian (Indian) lineage, 77 isolates of the South African lineage, and 2 isolates (including the type strain [9]) of the Japanese/Korean lineage. C. auris isolates from the Southern Asian clade had been referred from 20 different primary hospitals, 6 of which were known to have received patients from other UK centers affected by outbreaks. Two isolates were recovered from patients repatriated to UK centers from India, and six isolates were received directly after isolation from infected patients hospitalized in Oman. South African clade isolates were received from 14 independent UK health care centers, only 1 of which was known to have received patients from other UK centers affected by outbreaks. One isolate was recovered from a patient repatriated to the United Kingdom directly from Kenya. All C. auris isolates were identified by a combination of ribosomal DNA gene sequencing targeting the 28S rRNA and/or ITS1 regions and matrix-assisted laser desorption ionization–time of flight mass spectrometry exactly as described previously (29), and clade delineation was achieved phylogenetically using the combined 28S rRNA/ITS1 loci (24). The purity of each isolate was confirmed by subculture on Mast CHROMagar chromogenic medium (Mast Diagnostics).

The ability of individual isolates to produce pseudohyphae or other structures was evaluated by streaking individual isolates on cornmeal agar plates and incubation of the inoculated streak under a sterile coverslip at 30°C for 48 h (Dalmau cultures). Aggregate-forming capacity was evaluated microscopically, by attempting to emulsify a single C. auris colony in approximately 20 μl of sterile water on a microscope slide and microscopic examination of the resultant emulsion at a ×400 magnification. Aggregation was scored as “none” (only individual budding yeast cells were visible; Fig. 1A, panel 1), “small” (most cells in the population were present individually, with a minor proportion being present in small aggregates; Fig. 1A, panel 2), or “large” (few individual cells were visible, with most cells being present as parts of large aggregates; Fig. 1A, panel 3). To determine resistance to actidione, isolates were streaked in duplicate onto plates of Sabouraud agar and Sabouraud agar containing actidione (0.5 g/liter) and scored for growth after incubation at 37°C for 48 h. To assess thermotolerance, individual isolates were spotted onto replicate plates and growth was assessed after 72 h of incubation at 30°C to 45°C as described previously (28). Of the 127 isolates of the Southern Asian lineage that were examined for pseudohyphal formation, 90 were also examined for aggregate formation and 81 were examined for growth on actidione. Similarly, of the 65 South African clade isolates examined for pseudohyphal formation, 50 were examined for aggregate formation and 40 were examined for growth on actidione (Table 1).

FIG 1.

Phenotypic behavior of C. auris isolates. (A) Aggregation capacity assessed by aqueous suspension on sterile microscopy slides. (B) Morphological appearance on Dalmau culture. (C) Isolates grown for 48 h at 37°C on Sabouraud agar with actidione (+A) or without actidione (−A). Bars = 20 μm. The isolates tested were NCPF 8983 (Southern Asian clade) (panels 1), NCPF 13001 (Southern Asian clade) (panels 2), NCPF 8977 (South African clade) (panels 3), and NCPF 13029 (type strain, Japanese/Korean clade) (panels 4).

TABLE 1.

Candida auris clade-specific phenotypic characteristics

Characteristic	Value for Candida auris strains of the following cladef :
	Southern Asian	South African	Japanese/Korean
No. of isolates with growth on actidione/total no. of isolates tested (%)a	0/81 (0)*	40/40 (100)*	2/2 (100)**
No. (%) of isolates with pseudohyphae characterized as followsb :
None	21 (16)	65 (100)*	2 (100)***
Poor	25 (20)*	0	0
Good	81 (64)*	0	0
No. (%) of isolates with aggregate formation characterized as followsc :
Large	0	42 (84)*	1
Small	18 (20)	8 (16)	0
None	72 (80)*	0	1d
Time to 50% killing (h) in G. mellonella wax moths (no. of isolates)e	22 (8)	56 (4)	ND
Maximum growth temp (°C)	42	42	42

^aDetectable growth on Sabouraud agar containing actidione measured after 48 h of incubation at 37°C.

^bAbility of individual isolates to produce pseudohyphae on Dalmau culture after 48 h of incubation at 30°C.

^cAggregate formation measured by making aqueous suspensions of individual colonies in sterile water on microscope slides (27).

^dWhile the Japanese/Korean clade strain isolated in the United Kingdom forms large aggregates, the type strain described by Satoh et al. (9) does not.

^eData are taken from reference 27.

^fStatistical significance was determined using Fisher’s exact two-tailed test. *, P < 0.0001; **, P = 0.0003; ***, P = 0.036; ND, not determined.

Antifungal agents.

Antifungal drugs were obtained from their respective manufacturers as standard powders. Amphotericin B and nystatin (Sigma Chemical Co., St. Louis, MO), terbinafine (Novartis Pharmaceuticals, Camberley, UK), isavuconazole (Basilea Pharmaceutica Ltd., Basel, Switzerland), and anidulafungin and voriconazole (Pfizer Central Research, Sandwich, UK) were dissolved in dimethyl sulfoxide. Itraconazole (Janssen Research Foundation, Beerse, Belgium) and posaconazole (Merck, Sharp and Dohme, Hoddesdon, UK) were dissolved in polyethylene glycol 400 by heating to 70°C. Fluconazole (Pfizer Central Research, Sandwich, UK) and flucytosine (Sigma Chemical Co., St. Louis, MO) were suspended in sterile water. Serial 2-fold dilutions of the various drugs were prepared in RPMI 1640 medium (with l-glutamine, without bicarbonate; Sigma Chemical Co., St. Louis, MO) and buffered to pH 7.0 using a 0.165 M solution of MOPS (morpholinepropanesulfonic acid; Sigma Chemical Co., St. Louis, MO).

Antifungal susceptibility testing and determination of MICs and MFCs.

MICs were determined according to CLSI methodologies (CLSI standard M27-A4 [30]) in round-bottomed 96-well plates with yeast blastospore suspensions that had been prepared in saline and then diluted into RPMI 1640 medium and adjusted to final concentrations of 2.5 × 10³ CFU/ml. Inoculated plates were incubated for 24 to 48 h at 35°C. MICs were read at 24 and 48 h as the concentration of drug that elicited 100% inhibition of growth (amphotericin B, nystatin) or significant (approximately 50%) inhibition of growth compared with that on a drug-free control plate (anidulafungin, itraconazole, fluconazole, voriconazole, posaconazole, isavuconazole, terbinafine, and flucytosine). All assays included the control strains Candida parapsilosis NCPF 8334 (ATCC 22019) and C. krusei NCPF 3953 (ATCC 6258).

Minimum fungicidal concentrations (MFCs) were determined after 48 h of incubation exactly as described previously (31) by removing 10 μl of the contents from broth microdilution wells showing no visible growth and spreading them onto Sabouraud dextrose agar plates. The plates were then incubated for 48 h, and MFCs were determined as the lowest drug concentrations which killed 95% of the inoculum. Initial inoculum concentrations were verified in parallel by serial dilution and plating.

Induction of reversible cellular aggregation using antifungal Etest gradient strips.

Candida auris isolates were subjected to susceptibility testing using Etest gradient strips (bioMérieux UK Ltd., Basingstoke, UK) exactly as described previously (23, 32). Briefly, five colonies from a fresh culture were emulsified in saline to achieve a turbidity corresponding to a 0.5 McFarland standard. The resulting suspension was used to inoculate RPMI 1640 solid medium plates (bioMérieux UK Ltd., Basingstoke, UK), which were then allowed to dry for 15 min prior to application of individual Etest gradient strips and incubation at 35°C in a moist incubator for 24 or 48 h. Cellular morphologies were examined by sampling various portions of the inoculated plate after 48 h of incubation and preparing cell suspensions in sterile water directly on microscopy slides.

RESULTS

Clade-specific differences in capacity of C. auris isolates to produce aggregates and pseudohyphae and to grow on media containing actidione.

Previously, we showed that isolates of C. auris collected in the United Kingdom could be broadly divided into two groups, based on examination of their cellular morphology in culture (27). While the majority of isolates grew as individual, budding yeast cells, a subset of isolates produced large aggregates of cells that could not be physically disrupted, a property that was reproduced in the Galleria mellonella insect model of infection. In that model, single-cell inocula of the aggregate-forming isolates were substantially less pathogenic than their exclusively single-cell counterparts (27). Retrospective molecular analysis of the isolates employed in that study demonstrated that the aggregate-forming isolates were all members of the South African C. auris lineage (Table 1; Fig. 1A) (24). Examination of an additional 50 isolates belonging to that lineage, including an isolate recovered from a British patient repatriated directly from an African hospital, revealed that all members of this lineage tested reproducibly formed aggregates when grown in vitro on standard mycological media (Table 1), with the majority being capable of forming very large aggregates (42/50; 84%). Conversely, no isolates of C. auris from the Southern Asian (Indian) lineage demonstrated this capacity (Fig. 1A, panel 1), although 20% of isolates (18/90) did produce much smaller aggregations under certain growth conditions (Fig. 1A, panel 2; Table 1). Interestingly, aggregate formation was also a feature of the single example of the Japanese/Korean lineage that we recovered from the United Kingdom but was not seen with the type strain of the species, which was also of that lineage.

Examination of the capacity of isolates from different clades to produce pseudohyphae after culture on Dalmau plates (Fig. 1B) or to grow on media containing actidione (Fig. 1C) also revealed statistically significant clade-specific differences in isolate behavior. All isolates of the South African and Japanese/Korean clades that were tested grew well on Sabouraud agar containing actidione (40/40 South African isolates; 2/2 Japanese/Korean isolates; Table 1; Fig. 1C), whereas 0/81 isolates of the Southern Asian (Indian) clade that were tested demonstrated this capacity (Table 1; Fig. 1C). Conversely, none of 65 South African or 2 Japanese/Korean isolates examined appeared to be able to produce even rudimentary pseudohyphae when subjected to Dalmau culture (Fig. 1B, panel 3), whereas the majority of isolates of the Southern Asian (Indian) lineage produced such structures, with 25/127 isolates producing rudimentary pseudohyphae (Fig. 1B, panel 2) and 81/127 isolates producing well-developed pseudohyphae (Fig. 1B, panel 1) (total, 106/127 [83.5%]). Pseudomycelium formation in Southern Asian clade isolates did not appear to correlate with the ability to form aggregates. For the 90 isolates that were examined for the presence of aggregates and pseudohyphal formation, pseudomycelium was detected in 72.2% (13/18) and 87.5% (63/72) of aggregate-forming and single-celled isolates, respectively (Table 1 and data not shown).

Clade-specific differences in antifungal MICs and MFCs.

To investigate whether resistance to various antifungal agents might correlate with the clonal origin of C. auris isolates, we established the antifungal MICs of 10 antifungal agents against a panel of >120 C. auris isolates from the Southern Asian and South African lineages (Table 2). Isolates were selected to represent as many different referral centers as possible, to avoid possible bias caused by the inclusion of multiple isolates from single outbreaks. Selection of the panel of antifungal drugs that were tested was based on the site of isolation of the strain, in conjunction with specific requests from the referring clinicians. MICs were read as the lowest concentration of drug that exhibited significant (>50%) inhibition of growth. To date, we have not observed significant trailing of MIC endpoints with any isolates of Candida auris, irrespective of the clonal lineage. Thus, MICs determined as the concentration of drug that elicited 80% to 90% inhibition of growth were at most a single doubling dilution higher than those measured as the concentration that elicited >50% inhibition (data not shown). In agreement with previous reports (11, 13–23), all tested isolates exhibited in vitro resistance to fluconazole (MICs, 8 to >64 mg/liter), based on C. albicans interpretive breakpoints (≤2.0 mg/liter for susceptible, ≥8.0 mg/liter for resistant). However, the modal and geometric mean MICs with fluconazole were significantly higher against isolates of the South African lineage than against those of the Southern Asian lineage, with fluconazole MICs of >64 mg/liter being found for all except one of the South African lineage isolates (Table 2 and data not shown). Similar patterns were observed with voriconazole, with significantly higher MICs being found for South African lineage isolates than for Southern Asian lineage isolates and overall in vitro resistance to voriconazole (based on C. albicans breakpoints [≤0.125 mg/liter for susceptible, ≥1.0 mg/liter for resistant]) being seen in over 90% of isolates of the South African lineage (Table 2). Although variable levels of resistance to the other triazole antifungals and to amphotericin B, anidulafungin, terbinafine, and nystatin were observed, no significant clade-specific differences in susceptibility were evident (Table 2). Of note here was the apparently bimodal distribution of flucytosine MICs seen specifically with Southern Asian clade isolates. However, 11 of 17 isolates with elevated flucytosine MICs were obtained from patients with urinary tract (n = 6) or bloodstream (n = 5) C. auris infections, whereas 4/33 isolates from these same sites (1 bloodstream, 3 urinary tract) had lower flucytosine MICs (<4 mg/liter). Thus, it is possible that this bimodal MIC distribution results from acquired resistance to flucytosine in patients being treated actively for C. auris infection. Unfortunately, insufficient clinical information pertaining to antifungal treatment was supplied from the referring laboratories to fully investigate this possibility for the majority of isolates included in the current study.

TABLE 2.

MIC distributions of isolates of the Southern Asian (Indian) and African lineages of C. auris determined by CLSI broth microdilution^a

Drug (clade, no. of isolates)	No. of isolates with the following MIC (mg/liter):														% R	Overall % R
	0.015	0.03	0.06	0.125	0.25	0.5	1	2	4	8	16	32	64	>64
Amphotericin B (Southern Asian, n = 74)					1	5	54	14							18.9	14.6
Amphotericin B (South African, n = 49)				1	1	21	22	3	1						6.1
Amphotericin B (Japanese, n = 1)							1								0
Fluconazole (Southern Asian, n = 77)										6	27	12	9	23	100	100
Fluconazole (South African, n = 50)											1			49	100*
Fluconazole (Japanese/Korean, n = 1)														1	100
Itraconazole (Southern Asian, n = 34)	20	2	6	1	1	2					1				0	4.3
Itraconazole (South African, n = 12)					7	4	1								8.3
Itraconazole (Japanese/Korean, n = 1)					1										0
Voriconazole (Southern Asian, n = 71)			12	20	15	13	7	4			2				18.3	48.3
Voriconazole (South African, n = 49)					1	3	17	15	11	1	1				91.8*
Voriconazole (Japanese/Korean, n = 1)							1								100
Posaconazole (Southern Asian, n = 27)	18	2	3		1			2			1				14.8	12.8
Posaconazole (South African, n = 12)		1	8	2	1										8.3
Posaconazole (Japanese/Korean, n = 1)			1												0
Isavuconazole (Southern Asian, n = 22)	20					1				1					9.1	5.9
Isavuconazole (South African, n = 12)	6	6													0
Isavuconazole (Japanese/Korean, n = 1)				1											0
Anidulafungin (Southern Asian, n = 72)	3	3	20	21	12	7	2	2	2						8.3	6.8
Anidulafungin (South African, n = 46)	1		15	20	6	2	1	1							3.3
Anidulafungin (Japanese/Korean, n = 1)		1													0
Flucytosine (Southern Asian, n = 49)			27	2	3		1		3	4	2	2	1	5	16.3	13.1
Flucytosine (South African, n = 12)			7		3	1	1								0
Flucytosine (Japanese/Korean, n = 1)				1											0
Terbinafine (Southern Asian, n = 22)							7	10	2		1	2			13.6	11.5
Terbinafine (South African, n = 4)							1	2	1						0
Terbinafine (Japanese/Korean, n = 1)							1								0
Nystatin (Southern Asian, n = 24)								1	19	3	1				4.2	3.6
Nystatin (South African, n = 4)									1	3					0
Nystatin (Japanese/Korean, n = 1)						1									0

^aMIC values for the single UK isolate of Japanese/Korean lineage are included for comparison. Numbers in bold represent MICs that exceed the current CLSI breakpoints/epidemiological cutoff values for C. albicans. % R, total proportion of isolates that would be considered resistant using C. albicans interpretive breakpoints. *, P < 0.0001, 2-sample t test of geometric means.

To further investigate the observed clade-specific differences in antifungal susceptibility, endpoint plating of a selection of the isolates from the MIC assays was performed in order to establish minimum fungicidal concentrations (MFCs). With the exception of amphotericin B, none of the antifungals exhibited significant fungicidal activity against C. auris isolates belonging to either the Southern Asian or the South African lineage. However, geometric mean MFC values for itraconazole, posaconazole, isavuconazole, anidulafungin, and flucytosine all showed a trend toward being higher against South African clade isolates than against their Southern Asian clade counterparts, with the difference being statistically significant with posaconazole (Table 3).

TABLE 3.

MFCs of isolates of the Southern Asian and South African clades, determined by endpoint plating of broth microdilution susceptibility tests

Clade	Range (GM) MFC or MFC (mg/liter)a
	AMB	FLC	ITC	VRC	PSC*	ISAV	AND	5FC	NYT	TRB
South African (n = 8)	1 to 4 (1.68)	>64 (128)	8 to >16 (22.6)	>16 (32)	8 to >16 (22.6)	16 to >16 (29.3)	4 to >8 (11.3)	1 to 32 (16)	4 to 16 (6.6)	>16 (32)
Southern Asian (n = 17)	1 to 16 (1.84)	>64 (128)	8 to >16 (15.36)	>16 (32)	2 to 4 (3.40)	8 to >16 (15.36)	2 to >8 (4.71)	1 to >64 (3.40)	4 to 32 (5.32)	>16 (32)
Japanese/Korean (n = 1)	1	>64	1	16	0.5	2	0.5	16	4	>16

^aGM, geometric mean. *, P < 0.001 (2 sample t test) for GM of South African versus Southern Asian clade isolates. Antifungal drug abbreviations: AMB, amphotericin B; FLC, fluconazole; ITC, itraconazole; VRC, voriconazole; PSC, posaconazole; ISAV, isavuconazole; AND, anidulafungin; 5FC, flucytosine; NYT, nystatin; TRB, terbinafine.

Exposure to triazole and echinocandin antifungal agents, but not to amphotericin B or flucytosine, induces reversible aggregate formation in C. auris isolates of the Southern Asian lineage.

Since C. auris isolates of the South African lineage form robust cellular aggregates in vitro and the triazole antifungal agents exhibit higher MICs and/or MFCs for isolates of this lineage, we postulated that cell aggregate formation might be linked to protection against triazoles. To explore a possible link between aggregation and drug exposure/protection, we examined the cellular morphology of nonaggregating Southern Asian clade isolates directly in the wells of the microtiter plates employed for antifungal susceptibility testing using the CLSI methodology (Fig. 2A). C. auris cells grown in drug-free control wells exhibited a typical single-cell morphology, whereas exposure to even low concentrations (0.03 or 0.06 mg/liter) of itraconazole caused pronounced aggregate formation. Similar aggregation was seen in equivalent experiments performed with fluconazole, voriconazole, and anidulafungin but not in experiments performed with flucytosine or amphotericin B (data not shown). To confirm these findings, we repeated antifungal susceptibility testing using Etest gradient antifungal strips and isolates of C. auris from the Southern Asian lineage (Fig. 2B) and compared the morphology of cells sampled from the perimeter of the plate to that of cells sampled directly adjacent to the zone of growth inhibition (Fig. 2B). Cell morphology was unaffected by exposure of the isolates to amphotericin B or flucytosine, as evidenced by growth as single cells both around the perimeter of the test plates and near the Etest gradient strip (Fig. 2, panels AMB [amphotericin B] and 5FC [flucytosine]), Conversely, isolates were induced to grow as aggregates when exposed to any of the triazole antifungals (fluconazole, itraconazole, voriconazole) (Fig. 2; panels FLC [fluconazole], ITC [itraconazole], and VRC [voriconazole]). Aggregate formation was substantially greater when the experiments were repeated with the echinocandin antifungal agents, caspofungin or anidulafungin (Fig. 2; panels CAS [caspofungin] and AND [anidulafungin]). In all cases, this induction of aggregation was reversible: when the aggregating population was harvested from one experiment and used to prepare the inoculum for a subsequent Etest, cells around the plate periphery had reverted to the single-cell phenotype but continued to aggregate near the inhibition zone on plates with Etest strips containing the triazole or echinocandin antifungals but not those containing amphotericin B or flucytosine (see Fig. 2B, panels CAS*, for illustration, and data not shown). This inducible, reversible aggregation was evidenced with many independent isolates of the Southern Asian lineage, but only with triazole or echinocandin antifungal agents (data not shown).

FIG 2.

Induction of aggregate formation by exposure to triazole and echinocandin antifungal drugs. (A) Microscopic appearance of Southern Asian C. auris cells removed from CLSI antifungal susceptibility test microtiter wells containing the indicated concentrations (in milligrams per liter) of itraconazole after 48 h of incubation. (B) Microscopic appearance of Southern Asian C. auris cells in aqueous suspensions prepared from the indicated areas of Etest gradient antifungal susceptibility assays. Suspensions were examined at ×400 magnification by light microscopy; the panels in the top row were prepared from cells adjoining the zone of growth inhibition; the panels in the bottom row were prepared from cells around the perimeter of the Etest plates (as indicated by arrows, right-hand side). Abbreviations: AMB, amphotericin B; 5FC, flucytosine; CAS, caspofungin; AND, anidulafungin; FLC, fluconazole; VRC, voriconazole; ITC, itraconazole. The panels labeled CAS* were prepared from the equivalent areas of a caspofungin Etest assay prepared using the cellular aggregates from the panel marked with an asterisk (i.e., with cells growing near the zone of inhibition of a previous caspofungin Etest assay) as the inoculum. Similar reversible inhibition of aggregation was observed with all of the triazole and echinocandin antifungal agents tested, but not with amphotericin B or flucytosine (data not shown). Bars = 20 μm.

DISCUSSION

Previous studies have confirmed the existence of distinct, geographically restrained, clonal lineages of Candida auris (9–21), with evidence that this novel pathogen has simultaneously emerged on three different continents (21). We have previously shown that three of the four lineages (the Southern Asian, South African, and Japanese/Korean lineages) have been introduced independently multiple times into the United Kingdom, with isolates of the Southern Asian and South African clades circulating in several health care trusts (24). Indeed, isolates from the Southern Asian (33) and South African (34) lineages have caused several large nosocomial outbreaks in the United Kingdom, as well as on their continents of origin (13, 16). Here, we have shown that isolates belonging to the Southern Asian and South African lineages differ significantly in their phenotypic behavior in vitro, including the ability to grow on actidione, the ability to produce pseudomycelium, and cellular appearance (single cells versus aggregates) on standard mycological media. The formation of large cellular aggregates, which we have previously shown is related to reduced virulence in Galleria mellonella wax moths (27), was seen predominantly in isolates of the South African lineage (and in one of the two Japanese/Korean clade isolates). Similarly, South African isolates, but not their Southern Asian counterparts, grew well on media containing actidione, which could be used as a simple laboratory test to differentiate between these two clades. While of limited clinical value, the ability to rapidly distinguish these clades has epidemiological and infection control utility, especially in centers where isolates from both clades might be circulating simultaneously (24). Conversely, pseudomycelium formation was seen only in isolates of southern Asian origin, in keeping with a recent study that reported phenotypic switching between a yeast form and filamentous form in a single isolate of this lineage upon passage through a mammalian host (35). To our knowledge, the current study provides the first direct demonstration that isolates from the different C. auris lineages might exhibit different phenotypic behaviors. It remains speculative as to whether such in vitro differences might influence the ability of the different clonal lineages to cause sustained outbreaks.

A limitation of the current study is the inclusion of only two isolates representing the Japanese/Korean C. auris lineage. Unfortunately, this is due to the lack of availability of such isolates worldwide. To date only 1 isolate from this lineage has been recovered in the United Kingdom, and an identical situation has been reported for the United States (36). Moreover, with the exception of the type strain, which was included in the current study, the majority of the original strains described from this region (9, 10) are not preserved in culture collections. In this respect, it will be important in future work to attempt to determine why large outbreaks have not been reported with the Japanese/Korean C. auris clade after its introduction into either the United Kingdom (24), the United States (36), or elsewhere worldwide. However, it is worth noting here that the Japanese type strain of C. auris behaved differently from the UK isolate of the Japanese clade and from most of the isolates of the South African and Southern Asian clades with respect to both phenotypic behavior and antifungal susceptibility (Tables 1 and 3), suggesting that it might not be very representative of the species as a whole.

Since whole-genome sequencing of C. auris isolates from different clonal origins previously revealed high intraclonal genetic homogeneity but large genetic distances separating the clonal lineages (21), it is perhaps not surprising that isolates from the various lineages might exhibit slight variations in phenotypic behavior. Those studies also demonstrated that different clade-specific mutations in the ERG11 gene contributed to fluconazole resistance (21). Since various ERG11 mutations contribute differentially to fluconazole resistance in other pathogenic yeasts (37), we evaluated the antifungal susceptibility profiles of large numbers of isolates corresponding to the two major C. auris clades circulating in the United Kingdom. Here, we applied the CLSI breakpoints for C. albicans to interpret C. auris MICs, as suggested previously (38), rather than the higher breakpoints proposed more recently by the CDC (39) (an MIC of >32 mg/liter for resistance), in part due to reports of persistent or breakthrough infections despite fluconazole treatment caused by C. auris isolates with fluconazole MICs of 2 and 8 mg/liter (12). Based on CLSI C. albicans interpretive criteria, all C. auris isolates were resistant to fluconazole, and most isolates of the South African lineage were also resistant to voriconazole. Moreover, fluconazole and voriconazole had significantly higher geometric mean MICs, and posaconazole MFCs were also higher for isolates of the South African clade than for those of the Southern Asian clade. Although a previous report suggested possible clade-specific variations in antifungal susceptibility (40), albeit with low numbers of isolates from some of the lineages, we believe that the present study provides the first evidence that significant differences may exist.

Whether these differences are a direct consequence of the different resistance mutations found in isolates of the two clades or also involve factors such as ERG11 copy number variations or duplications of multidrug transporter genes, as reported for some isolates of the South African lineage (41), remains to be determined. A possible contributory explanation for these observed antifungal susceptibility differences might also reside in the ability of isolates of the South African lineage to form large cellular aggregates in vitro (Table 1) (24), which could confer a protective effect on the cell population by impeding antifungal drug penetration to the center of the aggregates. Indirect support for this idea was provided here by the observation that isolates of the Southern Asian clade could be reversibly induced to switch to the aggregative phenotype by exposure to triazole antifungal agents (Fig. 2). These observations suggest that aggregation may be an escape/stress response to perturbation of ergosterol synthesis in naturally nonaggregative isolates and/or that South African lineage isolates may have some natural defect in that synthesis pathway that naturally confers the aggregative, protective phenotype. Recently, transcriptome analyses of C. auris biofilms compared to planktonic cells revealed upregulation of a number of key resistance genes encoding efflux pumps, including ATP-binding cassette and major facilitator superfamily transporters (42). Thus, it is also plausible that aggregate formation typical of the South African lineage isolates might mimic biofilm formation and also activate some of those genes, further contributing to the elevated triazole antifungal MICs observed with such isolates in the current study. Upregulation of expression of efflux pump genes or subtle changes in ergosterol organization/membrane fluidity in South African clade isolates may also explain why those isolates, but not their Southern Asian counterparts, are resistant to cycloheximide (actidione) in vitro, since they have been shown to mediate such resistance in C. glabrata and Saccharomyces cerevisiae, respectively (43, 44).

The ability of the echinocandin class of antifungal agents to effectively penetrate Candida biofilms (45) might explain why the naturally aggregate-forming isolates of the South African clade were not more resistant than their Southern Asian counterparts to those antifungal agents in vitro (Table 2). However, the most impressive inducible aggregate formation with Southern Asian clade isolates was provoked by exposure to anidulafungin or caspofungin (Fig. 2), suggesting that perturbation of the synthesis of β-1,3-d-glucan as well as ergosterol can induce C. auris aggregation. Additional indirect support for the role of cell morphology and structure in the different, clade-specific C. auris phenotypes described here comes from recent studies involving deletion of the evolutionarily conserved stress-activated protein kinase Hog1. Perturbation of the hog1 gene in C. auris isolates induced the formation of large cellular aggregates/clumps; increased resistance to cell wall-damaging agents, including the echinocandins; an increased exposure of chitin on the cell surface; and notably, a significant reduction in virulence in vivo (46). Further studies will be required to evaluate whether C. auris isolates from the different clades exhibit differences in Hog1 activity or activation profiles or differential interactions between Hog1 and associated mitogen-activated protein kinase pathways.

In summary, here we provide evidence that C. auris isolates of the Southern Asian and South African lineages exhibit significant differences in phenotypic and antifungal susceptibility profiles in vitro. Further studies will aim to elucidate whether such differences have clinical significance and to attempt to establish why isolates of three of the four clonal lineages have been reported from large-scale nosocomial outbreaks but none to date have been attributed to isolates from the Japanese/Korean clade.