Breakthrough Invasive Candida glabrata in Patients on Micafungin: a Novel FKS Gene Conversion Correlated with Sequential Elevation of MIC

Abstract

Candida glabrata strains sequentially isolated from blood developed resistance to micafungin (MICs from <0.015 to 4 μg/ml). A novel mutation identified in micafungin-resistant strains at bp 262 of FKS2 (containing a deletion of F659 [F659del]) was inserted into the homologous region in FKS1.

CASE REPORT

A 93-year-old man was admitted to our hospital with a diagnosis of pulmonary tuberculosis (day 1). The patient had been receiving treatment for essential hypertension with chronic renal failure for a decade. On admission, the patient's vital signs were normal; however, serum laboratory data showed a marked elevation of creatinine (Cr) (4.7 mg/dl) and blood urea nitrogen (BUN) (74.5 mg/dl). After initiation of antituberculous therapy with oral isoniazid (300 mg/day) plus rifampin (450 mg/day), renal failure progressed (Cr, 7.0 mg/dl) due to drug-induced myoglobinemia (1,000 ng/ml) with uremic symptoms. Although an urgent flexible double lumen (FDL) catheter was introduced into the internal jugular vein, readministration of isoniazid (day 18) caused severe rhabdomyolysis (myoglobinemia, 25,650 ng/ml), with a recurrence of the uremic symptoms.

On day 27, the patient suddenly went into a state of shock with high fever and was empirically treated with intravenous meropenem (0.5 g/day), vancomycin (0.5 g, every 48 h [q48h]), and fluconazole (200 mg/day) based on a tentative diagnosis of aspiration pneumonia or catheter-related bloodstream infection complicated by sepsis. On the same day, two sets of blood cultures and serum endotoxin antigen were negative except for an elevation of β-d-glucan (133 pg/ml). On day 32, the patient's serum value of β-d-glucan rose to 530 pg/ml, and he had a positive result for serum galactomannan (Aspergillus antigen) of 4.5, thrombocytopenia (6.4 × 10³ platelets/μl), and leukocytopenia (2.0 × 10³ leukocytes/μl). Therefore, the fluconazole was changed to voriconazole (6 mg/kg of body weight/day, q12h) with the intent of targeting Aspergillus spp. However, on day 35, a blood culture collected on day 32 (strain NO1) was identified as Candida glabrata; therefore, voriconazole was changed to intravenous micafungin (100 mg/day) according to the Infectious Diseases Society of America (IDSA) 2009 guidelines (1). A blood culture taken on day 34 (NO2) also was positive for C. glabrata; however, after initiation of treatment with micafungin, the persistent fever subsided, and a blood culture taken at day 37 was negative for the yeast. Both strain NO1 and strain NO2 were susceptible to micafungin (MIC, <0.015 μg/ml) but susceptible-dose dependent to fluconazole (MIC <8 μg/ml) by Clinical and Laboratory Standards Institute (CLSI) broth microdilution (BMD) methods (CLSI document M27-S4). With regard to voriconazole, no breakpoint was determined for C. glabrata. In spite of two rounds of replacement of the FDL catheter, the serum value of β-d-glucan remained high (>600 pg/ml), and blood cultures taken on day 48 (NO3) and day 51 (NO4) again yielded C. glabrata. Based on the suspicion of a micafungin-resistant strain, micafungin treatment was changed to intravenous liposomal amphotericin B (3 mg/kg/day) on day 53. However, C. glabrata was still isolated from a blood culture taken on day 56 (NO5), and the patient died of septic shock on day 59. Following the patient's death, the NO3 strain was shown to be susceptible to micafungin (MIC, <0.015 μg/ml), while the NO4 and NO5 strains showed resistance to micafungin (based on CLSI document M27-S4), the MICs of which were 2 and 4 μg/ml, respectively (Table 1).

TABLE 1.

Drug susceptibilities of isolated Candida glabrata strains

Strain	MIC (μg/ml) ofa:
	MCFG (S, R)	AMPH-B (S)	5-FC (S)	FLCZ (SDD)	ITCZ (S, SDD)	VRCZ	PSCZ
NO1	<0.015	1	<0.12	8	1	0.5	2
NO2	<0.015	1	<0.12	8	1	0.5	0.5
NO3	<0.015	0.5	<0.12	4	0.5	0.25	0.5
NO4	2	1	<0.12	4	1	0.25	0.5
NO5	4	0.5	<0.12	8	1	1	2

^aAMPH-B, amphotericin B; 5-FC, 5-flucytosine; FLCZ, fluconazole; ITCZ, itraconazole; MCFG, micafungin; VRCZ, voriconazole; PSCZ, posaconazole; R, isolates were resistant; S, isolates were susceptible; SDD, isolates were susceptible, depending on the dose.

We performed morphological and genetic analyses for isolates NO1 to NO5. Interestingly, colonies observed in the isolation step for the NO1 and NO2 strains (Fig. S1A in the supplemental material) were of a size typical of C. glabrata (“normal”; they were light purple); in contrast, colonies for the NO3 and NO4 strains (Fig. S1B in the supplemental material) were heterogeneous, consisting of a mixture of small dark-purple and normal light-purple colonies, and the isolation step for NO5 (Fig. S1C in the supplemental material) yielded colonies of uniform size that were consistently small and darkly colored on CHROMagar Candida medium (Becton, Dickinson and Company). The change in colony size was thought to be due to mitochondrial deficiency (petite mutant) or FKS mutations, as described below. These results also suggest that the blood culture from which NO3 and NO4 were isolated contain several C. glabrata clones with heterogeneous growth rates. To examine whether the five clinical isolates originated from a single strain, randomly amplified polymorphic DNA (RAPD) and multilocus sequence typing (MLST) analyses were performed. Briefly, the template genomic DNA was extracted from C. glabrata cells, and a series of PCR and DNA sequencing reactions were performed using the primers indicated in Table S1 in the supplemental material. In RAPD assays (performed per previously reported methods [2]), all five of the tested strains yielded identical amplification patterns (data not shown). Furthermore, MLST analysis revealed that the five strains demonstrate a shared sequence type, ST22 (Table 2) (http://cglabrata.mlst.net/). These results suggest that the strains were probably derived from a single parental strain; however, more-detailed genetic analyses would be necessary for identification of the source of the strains.

TABLE 2.

Genetic characterization of isolated Candida glabrata strains^b

Strain (MIC [μg/ml] of MCFG)	Characterisitic(s) of gene:
	FKS1	FKS2
NO1 (<0.015)	Wild type	Wild type
NO2 (<0.015)	Wild type	Wild type
NO3 (<0.015)	Wild type	F659Δ L1767Δ
NO4 (2)	Gene conversiona	F659Δ
NO5 (4)	Gene conversiona	F659Δ

^aThe gene with the insertion is predicted to encode an Fks1 protein with the following changes: M555T, V558I, L563V, V568I, T583S, H600Q, A620S, and Y623Δ.

^bMCFG, micafungin. The sequence type for every strain was 22.

Cases harboring Candida spp. with reduced susceptibility to echinocandins are still uncommon, and strains of C. glabrata ranked as nonsusceptible to echinocandin have been reported only rarely in Japan (3, 4, 5). In previous reports, reduced susceptibility to echinocandins was related primarily to single point mutations within the FKS genes, which code for β-1,3-glucan-synthases. A total of 48 distinct mutations have been reported in 5 different yeast species, with 44 lesions occurring in hot spot 1 and 4 lesions occurring in hot spot 2 (6, 7, 8). The majority of the FKS1 mutations are predicted to result in an amino acid substitution at S629P (9), Phe625 (F625S or F625I), or Asp632 (D632G, D632E, or D632Y); FKS2 mutations are predicted to result in a substitution (F659V, F659S, or F659Y) (10) or a deletion (a deletion of F659 [F659del]) (11) in Phe659, as well as multiple mutations in residues 662 to 667, especially S663P (12, 13).

We determined the entire coding sequences of FKS1 and FKS2 in the five isolates (NO1 to NO5). Briefly, the entire FKS genes (7,332 bp and 7,941 bp for FKS1 and FKS2, respectively) were amplified by PCR from genomic DNA extracted from each strain, and DNA sequences of the PCR fragments were determined as described elsewhere but with different set of primers (see the primer list in Table S1 in the supplemental material) (5). The FKS sequences of the C. glabrata strains were reconfirmed with an independent analysis. In comparison to database sequences (the GenBank [https://www.ncbi.nlm.nih.gov/genbank/] accession numbers for FKS1 and FKS2 are HM366440 and HM366442, respectively), both NO1 and NO2 do not have any mutations which cause amino acid substitutions in both FKS1 and FKS2. The subsequently isolated strain NO3 harbored two deletion mutations (F659del and L1767del) in FKS2 (Table 2 and Fig. 1A). It was curious that NO3 was still susceptible to micafungin (MIC <0.015 μg/ml); however, F659del in FKS2 was thought to confer echinocandin resistance, as mentioned above (11, 12). It was also demonstrated that the laboratory-constructed C. glabrata mutant which had F659del in FKS2 was susceptible to echinocandins (8). These observations all together suggest that F659del in FKS2 alone does not necessarily confer echinocandin resistance to C. glabrata. It is also conceivable that L1767del in FKS2 or an unknown genetic modification(s) other than those in FKS genes suppress the effect of F659del; however, we have no evidence supporting these hypotheses.

FIG 1.

(A) Proposed model for the introduction of mutations into FKS genes. The flow chart of FKS1 and FKS2 modification is schematically rendered. The FKS1-type sequence and FKS2-type sequence are black and white, respectively. The representative point mutations which resulted in an amino acid substitution are designated with asterisks. FKS2 in NO3 carries two mutations, resulting in a predicted protein harboring F659del and L1767del (bottom left). In contrast, FKS2 in strains NO4 and NO5, both isolated from the patient subsequent to NO3's isolation, carries the F659del-encoding mutation alone (bottom right). This discrepancy suggests that NO3 and NO4 originated from a common ancestor (center middle) that was presumably susceptible to micafungin. The blood culture from which NO3 and NO4 were isolated also may have contained this hypothetical strain. (B) Alignment of flanking sequences around the region of the FKS gene substitution. FKS1 in NO3 (top), FKS1 in NO4 (middle, bold), and FKS2 in NO4 (bottom) are aligned for comparison. The numbering represents positions within the respective open reading frame, and the conserved nucleotides are highlighted by gray shading. The FKS1 sequence in NO4 is identical to a homologous FKS2 sequence in NO4 from bp 1635 to 1899 (numbering for FKS1 is above the alignment), whereas the FKS1 sequence in NO4 is identical to parental FKS1 in NO3 both upstream of bp 1635 and downstream of bp 1899. The F659del-encoding mutation in FKS2 also was inserted into FKS1 in NO4 (arrow). The resulting eight amino acid substitutions from the gene conversion are designated with single letters above and below the alignment (above for parental Fks1p amino acid residues and below for Fks2p; asterisks indicate amino acid residues conserved between Fks1p and Fks2p). The hot spot region related to echinocandin resistance is underlined. WT, wild type.

In micafungin-resistant NO4 and NO5, bp 262 of the FKS2 sequence (containing F659del) was substituted for the homologous region of FKS1, resulting in a predicted protein harboring multiple amino acid mutations (M555T, V558I, L563V, V568I, T583S, H600Q, A620S, Y623del) compared to the parental sequence (Table 2 and Fig. 1B). One possible cause for the micafungin resistance in NO4 and NO5 is the multiple mutations in FKS1 resulting from the genetic substitution from FKS2 to FKS1; however, both NO4 and NO5 still keep F659del in FKS2, presumably related to echinocandin resistance (Fig. 1A and B). It is ambiguous which FKS mutation(s) conferred micafungin resistance to NO4 or NO5; therefore, each FKS gene in NO4 and NO5 should be separately expressed in a C. glabrata laboratory strain and functionally characterized in future work.

Also of note was the fact that the mutation encoding L1767del in FKS2 (observed in NO3) was not observed in NO4 and NO5. The loss of this mutation indicates that NO3 and NO4 developed from a common ancestor whose FKS2 gene harbored the F659del-encoding mutation alone (Fig. 1A). Thus, the heterogeneous colony sizes observed in NO3 and NO4 may reflect the existence of a population of C. glabrata organisms carrying heterogeneous FKS gene sequences in NO3 and NO4.

To our knowledge, this is the first report suggesting that a genetic addition from FKS2 to FKS1 can mediate micafungin resistance in C. glabrata. This case also suggests that morphological colony phenotypes may be associated with changes in micafungin susceptibility in C. glabrata isolates.