Rate of FKS Mutations among Consecutive Candida Isolates Causing Bloodstream Infection

Abstract

Precise FKS mutation rates among Candida species are undefined because studies have not systematically screened consecutive, disease-causing isolates. The Sensititre YeastOne (SYO) assay measures echinocandin MICs against Candida with less variability than reference broth microdilution methods. However, clinical breakpoint MICs may overstate caspofungin nonsusceptibility compared to other agents. Our objectives were to determine Candida FKS mutation rates by studying consecutive bloodstream isolates and to determine if discrepant susceptibility results were associated with FKS mutations. FKS hot spots were sequenced in echinocandin-intermediate and -resistant isolates and those from patients with breakthrough candidemia or ≥3 days of prior echinocandin exposure. Overall, 453 isolates from 384 patients underwent susceptibility testing; 16% were echinocandin intermediate or resistant. Intermediate susceptibility rates were higher for Candida glabrata than for other species (P < 0.0001) and higher for caspofungin than for other agents (P < 0.0001). Resistance rates were similar between agents. FKS mutations were detected in 5% of sequenced isolates and 2% of isolates overall. Corresponding rates among C. glabrata isolates were 8% and 4%, respectively. Among Candida albicans isolates, rates were 5% and <1%, respectively. Mutations occurred exclusively with prior echinocandin exposure and were not detected in other species. Isolates with discrepant susceptibility results did not harbor FKS mutations. Mutation rates among isolates resistant to ≥2, 1, and 0 agents were 75%, 13%, and 0%, respectively. In conclusion, FKS mutations were uncommon among non-C. glabrata species, even with prior echinocandin exposure. Discrepancies in echinocandin susceptibility by SYO testing were not driven by mutations and likely reflect imprecise caspofungin clinical breakpoints.

INTRODUCTION

Echinocandin resistance is emerging among clinical Candida isolates, particularly those of the haploid species Candida glabrata (1–3). Resistance is mediated through point mutations in hot spot regions of the FKS1 and FKS2 genes, which encode the echinocandin target enzyme β-1,3-d-glucan synthase. FKS mutations are associated with echinocandin treatment failures and high mortality rates among patients with invasive candidiasis (1–5). Exposure to the echinocandins almost always precedes the emergence of FKS mutations and development of resistance (3, 6, 7). Up to 32% of C. glabrata isolates from patients with prior echinocandin exposure harbor FKS mutations; risk is greatest with breakthrough infections during echinocandin prophylaxis or treatment (6). Overall rates of FKS mutant Candida are imprecisely defined. Rates of 8 to 18% have been reported among C. glabrata isolates from patients at high-risk centers (1, 2); however, these data may overstate mutation rates, since the studies were limited by incomplete access to medical records and a lack of systematic testing of consecutive isolates (1). The other major Candida species (Candida albicans, C. parapsilosis, C. tropicalis, C. krusei, and C. guilliermondii) account for 60 to 80% of invasive candidiasis (8), but FKS mutations have been described only in case series and reports (9). To date, no study has systematically screened sequential Candida isolates for the presence of FKS mutations.

In the clinical microbiology laboratory, resistance is typically assessed by measuring drug MICs and comparing results to reference breakpoints. Reference broth microdilution testing methods and clinical breakpoint MICs for echinocandins against Candida species have been developed by the Clinical and Laboratory Standards Institute (CLSI) (10) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) (11). Unfortunately, several features of the broth microdilution reference methods have limited their utility in clinical practice. First, there is significant interlaboratory variability in caspofungin MICs (12), which has prevented EUCAST but not CLSI from proposing interpretive criteria for caspofungin. Second, echinocandin MICs have not been shown to correlate consistently with outcomes among patients with invasive candidiasis who are treated with these agents (13). Third, the reference methods are not used in most clinical microbiology laboratories (14), which instead employ commercial assays such as Sensititre YeastOne (SYO; Trek Diagnostics) and Etest (bioMérieux) or automated systems like the Vitek 2 (bioMérieux) antifungal testing instrument.

We recently showed that the SYO assay, as employed by clinical labs in routine practice, may reduce interlaboratory variability in caspofungin MICs (14). However, echinocandin MIC clinical breakpoints are not validated for commercial methods, and results may overstate nonsusceptibility. We demonstrated that application of CLSI breakpoints results in disproportionately high rates of caspofungin nonsusceptibility among C. glabrata and C. krusei compared to other agents (14). For example, 18% and 19% of C. glabrata isolates in our study were identified as intermediate or resistant to caspofungin but susceptible to anidulafungin and micafungin, respectively (14). Indeed, categorical discrepancies occurred most frequently among C. glabrata and C. krusei isolates classified as caspofungin intermediate, anidulafungin susceptible, and micafungin susceptible (14). The significance of discrepant susceptibility results is unknown, and it is unclear if categorical discrepancies are driven biologically by agent-specific FKS mutations (15) or if they are an artifact of imprecise clinical breakpoints.

The objectives of this study were to determine FKS mutation rates across Candida species by systematic sequencing of at-risk isolates and to determine if discrepant echinocandin susceptibility results were associated with agent-specific FKS mutations.

MATERIALS AND METHODS

Consecutive cases of candidemia at the University of Pittsburgh Medical Center Presbyterian Hospital from October 2009 to December 2014 were evaluated. A unique case of candidemia was defined as a blood culture yielding Candida that was more than 30 days after any prior positive blood culture with Candida. For candidemia caused by more than one Candida species, each species was considered a separate case for analysis. Antifungal susceptibility testing was performed with SYO panels according to the manufacturer's recommendations (TREK Diagnostic Systems, Cleveland, OH, USA). C. krusei ATCC 6258 and C. parapsilosis ATCC 22019 were used as quality controls. Results were included only when both control isolates were within acceptable CLSI MIC ranges for all agents (10). MICs were interpreted in accordance with recently published CLSI M27-S4 clinical breakpoints (10). Ten cases of candidemia due to uncommon species (4 Candida lusitaniae, 3 C. dubliniensis, 2 C. kefyr, and 1 C. famata isolate) were excluded from the study because CLSI breakpoints have not been established.

We employed a targeted, systematic screening approach to identify FKS mutations, which were detected using previously described methods (3, 6, 7). In short, DNA was extracted, hot spots of FKS1 (all species) and FKS2 (C. glabrata only) were amplified by PCR, and purified DNA was sequenced for any isolate meeting any of the following criteria: (i) isolation from a patient with ≥3 days of prior echinocandin exposure, (ii) isolation from a patient receiving ≥3 days of echinocandin therapy at the time of positive blood culture (i.e., breakthrough candidemia), or (iii) an echinocandin MIC classified as intermediate or resistant by CLSI breakpoints (10). Rates of FKS mutations and resistance were compared by chi-squared or Fisher's exact tests, as appropriate. Significance was defined as a two-tailed P value of <0.05.

RESULTS

Characteristics of Candida isolates and antifungal susceptibilities.

A total of 453 Candida isolates from 384 unique patients with candidemia were included in the analysis. More than 1 isolate was included from patients with candidemia due to multiple species (n = 11), relapsing candidemia occurring >30 days after a previous episode (n = 14), or both (n = 19). C. albicans and C. glabrata (37% each) were the most common species encountered, followed by C. parapsilosis (16%), C. tropicalis (8%), C. krusei (1%), and C. guilliermondii (<1%).

Sixteen percent (71/453) of isolates were classified as intermediate or resistant to an echinocandin by CLSI breakpoints (Table 1). Rates of intermediate susceptibility were higher for caspofungin (12%, 53/453) than anidulafungin (1%, 3/453; P < 0.0001) or micafungin (1%, 3/453; P < 0.0001). Intermediate susceptibility was more common among isolates of C. glabrata (26%, 44/167) than other species (3%, 10/286; P < 0.0001). Ninety-eight percent (43/44) and 95% (42/44) of caspofungin-intermediate C. glabrata isolates were susceptible to anidulafungin and micafungin, respectively.

TABLE 1.

Susceptibility profile of three echinocandins based on MICs determined by SYO

Species	n	No. (%) of isolatesa
		Intermediateb			Resistantc			Intermediate or resistant to any EC	MIC > SYO-specific ECVd
		ANF	CSP	MCF	ANF	CSP	MCF		ANF	CSP	MCF
C. albicans	169	2 (1)	6 (4)	0 (0)	0 (0)	1 (0.6)	1 (0.6)	8 (5)	4 (2)	7 (4)	9 (5)
C. glabrata	167	1 (0.6)	44 (26)	2 (1)	7 (4)	13 (8)	5 (3)	58 (35)	9 (5)	13 (8)	13 (8)
C. parapsilosis	71	0 (0)	0 (0)	1 (1)	0 (0)	0 (0)	0 (0)	1 (1)	0 (0)	0 (0)	0 (0)
C. tropicalis	38	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	2 (5)
C. krusei	6	0 (0)	3 (50)	0 (0)	0 (0)	1 (17)	0 (0)	4 (67)	0 (0)	0 (0)	0 (0)
C. guilliermondii	2	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Total	453	3 (1)	53 (12)	3 (1)	7 (2)	15 (3)	6 (1)	71 (16)	13 (3)	20 (4)	24 (5)

^aANF, anidulafungin; CSP, caspofungin; EC, echinocandin; ECV, epidemiological cutoff value; MCF, micafungin.

^bIntermediate susceptibility was adapted from CLSI criteria. For C. albicans, C. tropicalis, and C. krusei, MICs were 0.5 μg/ml; for C. parapsilosisi and C. guilliermondii, MICs were 4 μg/ml; and for C. glabrata, MICs were 0.25 μg/ml for anidulafungin and caspofungin and 0.12 μg/ml for micafungin.

^cResistance was adapted from CLSI criteria. For C. albicans, C. tropicalis, and C. krusei, MICs were ≥1 μg/ml; for C. parapsilosisi and C. guilliermondii, MICs were ≥8 μg/ml; and for C. glabrata, MICs were ≥0.5 μg/ml for anidulafungin and caspofungin and ≥0.25 μg/ml for micafungin.

^dEpidemiologic cutoff values were obtained from reference 16. The ECVs for anidulafungin MICs against C. albicans, C. glabrata, C. parapsilosis, C. tropicalis, C. krusei, and C. guillermondii were 0.12, 0.12, 4, 0.5, 0.25, and 4 μg/ml, respectively. The corresponding values for caspofungin were 0.25, 0.25, 2, 0.25, 1, and 2 μg/ml, respectively. The corresponding values for micafungin were 0.06, 0.03, 4, 0.06, 0.25, and 2 μg/ml, respectively.

Rates of resistance did not differ significantly for caspofungin (3%, 15/453), anidulafungin (2%, 7/453; P = 0.12), and micafungin (1%, 6/453; P = 0.07) (Table 1). Caspofungin resistance was identified among 17% (1/6), 8% (13/167), and 0.6% (1/169) of C. krusei, C. glabrata, and C. albicans isolates, respectively. Resistance to anidulafungin or micafungin was not detected among C. kruseii isolates. Anidulafungin and micafungin resistance was identified among 4% (7/167) and 3% (5/167) of C. glabrata isolates, respectively, and 0% (0/169) and 0.6% (1/169) of C. albicans isolates, respectively.

Three percent (13/453), 4% (20/453), and 5% (24/453) of isolates demonstrated MICs above recently proposed SYO-specific epidemiologic cutoff values (ECVs) for anidulafungin, caspofungin, and micafungin, respectively (16) (Table 1). Rates of MICs above the ECV were comparable between echinocandin agents and ranged from 5 to 8% and 2 to 5% among C. glabrata and C. albicans isolates, respectively. Micafungin MICs were above the ECV for 2 C. tropicalis isolates; otherwise, none of the isolates from other species exhibited an echinocandin MIC above the ECV.

The overall rate of fluconazole resistance (excluding C. krusei) was 10% (46/447). Rates of fluconazole resistance were 19% (31/167), 13% (5/38), 6% (4/71), 4% (6/169), and 0% (0/2) among C. glabrata, C. tropicalis, C. parapsilosis, C. albicans, and C. guilliermondii isolates, respectively. Twenty-six percent (15/58) and 25% (2/8) of echinocandin-intermediate or -resistant C. glabrata and C. albicans isolates were resistant to fluconazole, respectively.

Twenty-one percent (96/453) of Candida isolates were recovered from patients with prior echinocandin exposure; 3% (15/453) of isolates were classified as breakthrough (Table 2 ). Thirty-nine percent (28/71) of C. parapsilosis isolates were associated with prior echinocandin exposure, compared to 25% of C. glabrata isolates (41/166; P = 0.03) and 12% of C. albicans isolates (20/169; P < 0.0001). Prior exposure was more common among C. glabrata isolates than C. albicans isolates (P = 0.003). Twenty-four percent (23/96) and 13% (48/357) of isolates collected from patients with and without prior echinocandin exposure were classified as intermediate or resistant to an echinocandin, respectively (P = 0.02).

TABLE 2.

Candida isolates at risk for FKS gene mutations

Species	No. (%) of isolates			Median (range) duration of exposure in days	No. (%) of isolates
	Total	Intermediate or resistant to any ECa	With prior EC exposure		Breakthrough	At riskb	Harboring FKS mutations
C. albicans	169	8 (5)	20 (12)	30 (3–190)	2 (1)	27 (16)	1 (0.6)
C. glabrata	167	58 (35)	41 (25)	18 (4–450)	6 (4)	77 (46)	6 (4)
C. parapsilosis	71	1 (1)	28 (39)	67 (8–211)	4 (6)	29 (41)	0 (0)
C. tropicalis	38	0 (0)	7 (18)	22 (4–400)	2 (5)	7 (18)	0 (0)
C. krusei	6	4 (67)	0 (0)		0 (0)	4 (67)	0 (0)
C. guilliermondii	2	0 (0)	0 (0)		0 (0)	0 (0)	0 (0)
Total	453	71 (16)	96 (21)	27 (3–450)	15 (3)	144 (32)	7 (2)

^aAs defined in Table 1. EC, echinocandin.

^bAt-risk isolates include those from patients with prior or ongoing (i.e., breakthrough) echinocandin exposure or echinocandin-intermediate or -resistant isolates.

FKS mutations.

FKS hot spots were sequenced in 144 “at-risk” isolates: 71 isolates were either intermediate or resistant to an echinocandin by CLSI criteria (23 were also recovered from patients with prior echinocandin exposure), and 73 isolates were recovered from patients with prior echinocandin exposure but were susceptible to all echinocandins (Table 2). Five percent (7/144), 7% (7/96), and 10% (7/71) of at-risk isolates, isolates from patients with prior echinocandin exposure, and isolates that were intermediate or resistant to ≥1 echinocandin harbored FKS hot spot mutations, respectively (Table 2). Thirty percent (7/23) of isolates intermediate or resistant to an echinocandin from patients with prior exposure were FKS mutants. Mutations were not identified in isolates associated with prior echinocandin exposure but susceptible to all three echinocandins. Mutations were identified exclusively among C. albicans and C. glabrata isolates from patients with prior echinocandin exposure (Table 3). Among such isolates, the C. albicans and C. glabrata species-specific rates of FKS mutations were 5% (1/20) and 15% (6/41), respectively. Fifty percent (1/2) of breakthrough C. albicans and 67% (4/6) of breakthrough C. glabrata isolates harbored mutations.

TABLE 3.

Characteristics of FKS mutant Candida isolates

Species	FKS mutation	MIC (μg/ml)a				No. of days of:		Echinocandin breakthrough
		ANF	CSP	MCF	FLUC	Prior echinocandin exposure	Prior azole exposure
C. albicans	S645P	0.5 (I)	>8 (R)	2 (R)	>256 (R)	68	103	Yes
C. glabrata	D632Y	0.5 (R)	2 (R)	0.25 (R)	2 (S-DD)	239	None	Yes
	F659del	2 (R)	>8 (R)	4 (R)	128 (R)	46	139	No
	S663P	2 (R)	>8 (R)	8 (R)	256 (R)	155	94	Yes
	F659S	0.5 (R)	1 (R)	0.06 (S)	>128 (R)	7	22	Yes
	F659L	0.06 (S)	1 (R)	0.06 (S)	8 (S-DD)	117	108	No
	R636S	0.5 (R)	4 (R)	0.12 (I)	1 (S-DD)	450	227	Yes

^aANF, anidulafungin; CSP, caspofungin; FLUC, fluconazole; MCF, micafungin. The CLSI interpretation of the MIC is in parentheses. I, intermediate; R, resistant; S, susceptible; S-DD, susceptible, dose dependent.

Based on caspofungin MICs, 100% (1/1) and 46% (6/13) of resistant C. albicans and C. glabrata isolates, respectively, harbored mutations. No caspofungin-intermediate C. albicans (0/6; P = 0.14) or C. glabrata (0/44; P < 0.0001) isolates harbored mutations. All FKS mutants were caspofungin resistant. Rates of anidulafungin and micafungin resistance among FKS mutant Candida isolates were 71% (5/7) and 57% (4/7), respectively. Median caspofungin (3 versus 0.12 μg/ml), anidulafungin (0.5 versus 0.03 μg/ml), and micafungin (0.19 versus 0.015 μg/ml) MICs against FKS mutant C. glabrata isolates were higher than those against wild-type isolates (P < 0.0001 for each). Across species, FKS mutation rates were 75% (6/8), 13% (1/8), and 0% (0/437) among isolates resistant to ≥2, 1, and 0 echinocandins, respectively. Among isolates classified as intermediate or resistant to ≥2, 1, and 0 echinocandins, mutation rates were 69% (6/9), 2% (1/62), and 0% (0/382), respectively.

FKS mutations were present in 56% (5/9), 46% (6/13) and 46% (6/13) of C. glabrata isolates for which anidulafungin, caspofungin, and micafungin MICs were above the ECV, respectively. The corresponding rates for C. albicans isolates were 25% (1/4), 14% (1/7) and 11% (1/9). Eighty-seven percent (6/7), 100% (7/7), and 100% (7/7) of FKS mutant Candida isolates exhibited anidulafungin, caspofungin, and micafungin MICs above the ECV, respectively.

No FKS mutations were identified among an additional 40 isolates (27 C. glabrata, 8 C. albicans, 3 C. tropicalis, and 2 C. parapsilosis isolates) that did not meet sequencing criteria (negative controls). Assuming that none of the remaining isolates were FKS mutants, the overall mutation rate across all species was 2% (7/453). Overall FKS mutation rates were 4% (6/167) and 0.6% (1/169) among C. glabrata and C. albicans, respectively.

DISCUSSION

This is the first study to report the rates of FKS hot spot mutations across the major Candida species recovered sequentially from patients at a single center. Several findings are particularly noteworthy. First, echinocandin resistance and FKS gene mutations were uncommon during consecutive cases of candidemia and were encountered exclusively among C. glabrata and C. albicans isolates recovered from patients with prior echinocandin exposure. Second, we found no evidence to support agent-specific FKS mutations among isolates with discrepant echinocandin susceptibility results. Most notably, none of the isolates that tested as intermediate to caspofungin but susceptible to other agents by CLSI breakpoints harbored an FKS mutation. Finally, 75% of isolates that were classified as resistant to two or more echinocandins were FKS mutants, implicating hot spot mutations as a predominant, but not exclusive, mechanism of resistance. Taken together, our data provide new insights into echinocandin resistance and carry important implications for the use of these agents in clinical practice.

It is striking that only 2% (7/453) of all Candida isolates, and 5% (7/144) of isolates considered to be at risk for resistance, harbored FKS mutations. The corresponding rates for C. glabrata and C. albicans were 4 and 8% and <1 and 5%, respectively. The low frequency of gene mutations identified here is consistent with data from previous studies of isolates in an international repository (17, 18). Our C. glabrata rates are significantly lower than the recently reported rates of 8 to 18% at two major U.S. centers (1, 2), which may reflect institutional differences or the lack of systematic screening strategies in the earlier studies. On balance, the cumulative data indicate that FKS mutations and echinocandin resistance are important clinical problems, but the phenomena need to be placed in context. Indeed, selection for FKS gene mutations generally occurs in highly specific, echinocandin treatment-experienced patients (1–3, 6, 7). Isolates recovered from patients with breakthrough infections are at significantly greater risk than isolates from patients with more distant echinocandin exposure. In fact, only 4% of nonbreakthrough C. glabrata and C. albicans isolates that were associated with past echinocandin exposure were FKS mutants (2/35 and 0/18, respectively). Moreover, durations of prior exposure preceding resistance are typically quite extensive. It is likely that both the duration and timing of echinocandin exposure facilitate the emergence of echinocandin-resistant mutants (19).

Almost 26% of C. glabrata isolates were classified as caspofungin intermediate but susceptible to anidulafungin or micafungin using CLSI breakpoints (43/167 and 42/167, respectively). These discrepancy rates were slightly higher than the corresponding rates of 16% and 17%, respectively, that were reported in our earlier multicenter study (14). The number of C. krusei isolates was small, but 50% (3/6) were caspofungin intermediate and susceptible to the other agents. There is some evidence that certain FKS mutations may confer differential relative resistance to individual echinocandins (15, 20). However, the fact that none of our caspofungin-intermediate C. glabrata or C. krusei isolates had an FKS mutation indicates that categorical discrepancies in echinocandin susceptibility, in general, are not driven by such mutations but are more likely artifacts of imprecise caspofungin breakpoints.

Due to the interlaboratory variability in caspofungin MICs obtained with reference broth microdilution methods, recommendations have been made to use anidulafungin or micafungin MICs as a surrogate for the echinocandin class (21, 22). We found that anidulafungin or micafungin resistance was slightly more sensitive than caspofungin resistance for detecting FKS mutations (71% [5/7] and 67% [4/6], respectively, versus 47% [7/15]). However, 100% (7/7) of FKS mutants were caspofungin resistant, whereas only 71% (5/7) and 57% (4/7) were anidulafungin and micafungin resistant, respectively. Of note, the sensitivity and specificity of resistance to ≥2 agents for identifying FKS mutant Candida were 75% (6/8) and >99% (444/445), respectively, compared to 13% (1/8) and 99% (439/445) for resistance to one agent. Therefore, the best approach to identifying FKS mutations may be to consider MICs of all three echinocandins, rather than any single agent. In this regard, SYO panels and other commercial assays that provide results for each of the echinocandins may offer advantages for clinical microbiology laboratories.

Each of the mutations we identified has been linked with echinocandin resistance (20). At the same time, FKS mutations were not the sole determinants of diminished echinocandin susceptibility, as 25% of isolates resistant to ≥2 agents were not mutants. The mechanisms of resistance in these isolates are unclear, but they may involve modulation of membrane sphingolipids (23) and upregulation of cell wall chitin and/or other cell wall compensatory mechanisms (24). Furthermore, chromosomal instability during stress leads to increased genetic diversity, which enables isolates to develop rapid resistance to multiple antifungal drug classes (19). Along these lines, it is noteworthy that 57% of FKS mutant isolates reported here were also resistant to fluconazole; all four fluconazole-resistant isolates were recovered from patients with prior azole exposure (Table 3). Multidrug-resistant C. glabrata isolates, in particular, are a serious threat to the antifungal armamentarium, as at least 10% of fluconazole-resistant isolates are reported to harbor FKS mutations (25). Our data add to accumulating evidence that resistance to echinocandins is associated with an increased likelihood of azole resistance, and vice versa (2, 25, 26).

ECVs are designed to distinguish between a population of wild-type, drug-susceptible isolates and a population that includes non-wild-type isolates with acquired resistance mechanisms. A recent multicenter study assigned echinocandin ECVs against Candida species by using the SYO assay (16). In keeping with data from the multicenter study, we found that the species-specific ECVs correctly classified almost all of our FKS mutant C. glabrata and C. albicans isolates. Follow-up studies are needed to determine the value of MICs, clinical breakpoints, ECVs, and the presence of FKS mutations in predicting outcomes of echinocandin treatment among patients with invasive candidiasis.

In conclusion, this study provides important perspectives on echinocandin resistance among Candida species. These drugs are now the first choice for treatment of most cases of candidemia (27–30). Reports of the emergence of echinocandin-resistant and multidrug-resistant Candida isolates (in particular, C. glabrata) are concerning, but our data suggest that FKS mutations remain rare and are fairly difficult to induce in the clinic. Clinicians should maintain suspicion for resistance among patients with extensive prior echinocandin exposure, especially those with breakthrough infections or more recent treatment courses. In these settings, echinocandin MICs and screening for FKS mutations may help guide treatment decisions (3, 6, 21). Outside of these settings, however, resistance is extremely uncommon, and it is reasonable for clinicians to assume that each of the agents retains activity.