Implications of the EUCAST Trailing Phenomenon in Candida tropicalis for the In Vivo Susceptibility in Invertebrate and Murine Models

Abstract

Candida tropicalis isolates often display reduced but persistent growth (trailing) over a broad fluconazole concentration range during EUCAST susceptibility testing. Whereas weak trailing (<25% of the positive growth control) is common and found not to impair fluconazole efficacy, we investigated if more pronounced trailing impacted treatment efficacy. Fluconazole efficacy against two weakly (≤25% growth), two moderately (26% to 50% growth), and one heavily (>70% growth) trailing resistant isolate and one resistant (100% growth) isolate were investigated in vitro and in vivo (in a Galleria mellonella survival model and two nonlethal murine models). CDR1 expression levels and ERG11 sequences were characterized. The survival in fluconazole-treated G. mellonella was inversely correlated with the degree of trailing (71% to 9% survival in treatment groups). In mice, resistant and heavily trailing isolates responded poorly to fluconazole treatment. CDR1 expression was significantly higher in trailing and resistant isolates than in wild-type isolates (1.4-fold to 10-fold higher). All isolates exhibited ERG11 wild-type alleles. Heavily trailing isolates were less responsive to fluconazole in all in vivo models, indicating an impact on fluconazole efficacy. CDR1 upregulation may have contributed to the observed differences. Moderately trailing isolates responded less well to fluconazole in larvae only. This confirms clinical data suggesting fluconazole is effective against infections with such isolates in less severely ill patients and supports the current 50% growth endpoint for susceptibility testing. However, it is still unclear if the gradual loss of efficacy observed for moderately trailing isolates in the larva model may be a reason for concern in selected vulnerable patient populations.

Keywords: EUCAST; Galleria mellonella; azoles; fluconazole; in vivo; trailing.

FIG 1

Overview of the EUCAST microdilution enzyme-linked immunosorbent assay (ELISA) readouts of the phenotypes of the six included isolates. Means (solid black lines) and ranges (dotted black lines) of seven repetitions as well as the 50% and 100% (antifungal-free control) growth lines (gray dotted lines) are shown. The background OD of the ELISA reader has been subtracted. On the y axes, growth in individual wells is presented as percentages of growth compared to growth in the antifungal-free growth control wells. For the ranges, the repetitions with the most and the least degrees of trailing were chosen.

FIG 2

Expression levels of CDR1 in C. tropicalis. Expression levels were calculated as described in Materials and Methods and were normalized to ACT1 expression relative to merged data of CT-WT1 and CT-WT2 (CT-WT). ANOVA one-way statistics using merged data for CDR1 of both CT-WT1 and CT-WT2 (CT-WT) were employed as a reference. NS, not significant; *, P < 0.05; ***, P < 0.001; ****, P < 0.0001.

FIG 3

Kaplan-Meier curves of larval survival over 5 days depending on treatment for the five included isolates. PBS, solid gray line; FLC 1 mg/kg, dotted line; FLC 5 mg/kg, dashed line; FLC 20 mg/kg, solid black line. The P values are results from the log rank test of PBS versus treatment groups (the P values on top are for PBS versus FLC 20 mg/kg and values on the bottom are for PBS versus FLC 1 mg/kg).

FIG 4

Treatment efficacy over 96 h of various doses of fluconazole on C. tropicalis infection with trailing and nontrailing isolates in an immunocompetent mouse model. Horizontal lines indicate the median values. (A) CFU counts on day 4 after inoculation and daily treatment (starting at t = 2 h) with one dose of either vehicle control or fluconazole 35 mg/kg. (B) For isolate CT-TR-R, no reduction in CFU counts was found. For the remaining four isolates, lower dosages of fluconazole 1, 5, and 15 mg/kg were also subsequently tested.

FIG 5

Treatment efficacy over 24 h of various doses of fluconazole on C. tropicalis infection with trailing and nontrailing isolates in a neutropenic mouse model. Horizontal lines indicate the median values. One pair of kidneys was lost during transport between facilities prior to CFU determination (CT-WT2, FLC 15 mg/kg) and thus was not available for analysis.