A Clinical Study Provides the First Direct Evidence That Interindividual Variations in Fecal β-Lactamase Activity Affect the Gut Mycobiota Dynamics in Response to β-Lactam Antibiotics

Abstract

Antibiotics disturb the intestinal bacterial microbiota, leading to gut dysbiosis and an increased risk for the overgrowth of opportunistic pathogens. It is not fully understood to what extent antibiotics affect the fungal fraction of the intestinal microbiota, the mycobiota. There is no report of the direct role of antibiotics in the overgrowth in healthy humans of the opportunistic pathogenic yeast Candida albicans. Here, we have explored the gut mycobiota of 22 healthy subjects before, during, and up to 6 months after a 3-day regimen of third-generation cephalosporins (3GCs). Using ITS1-targeted metagenomics, we highlighted the strong intra- and interindividual diversity of the healthy gut mycobiota. With a specific quantitative approach, we showed that C. albicans prevalence was much higher than previously reported, with all subjects but one being carriers of C. albicans, although with highly variable burdens. 3GCs significantly altered the mycobiota composition and the fungal load was increased both at short and long term. Both C. albicans relative and absolute abundances were increased but 3GCs did not reduce intersubject variability. Variations in C. albicans burden in response to 3GC treatment could be partly explained by changes in the levels of endogenous fecal β-lactamase activity, with subjects characterized by a high increase of β-lactamase activity displaying a lower increase of C. albicans levels. A same antibiotic treatment might thus affect differentially the gut mycobiota and C. albicans carriage, depending on the treated subject, suggesting a need to adjust the current risk factors for C. albicans overgrowth after a β-lactam treatment. IMPORTANCE Fungal infections are redoubtable healthcare-associated complications in immunocompromised patients. Particularly, the commensal intestinal yeast Candida albicans causes invasive infections in intensive care patients and is, therefore, associated with high mortality. These infections are preceded by an intestinal expansion of C. albicans before its translocation into the bloodstream. Antibiotics are a well-known risk factor for C. albicans overgrowth but the impact of antibiotic-induced dysbiosis on the human gut mycobiota-the fungal microbiota-and the understanding of the mechanisms involved in C. albicans overgrowth in humans are very limited. Our study shows that antibiotics increase the fungal proportion in the gut and disturb the fungal composition, especially C. albicans, in a subject-dependent manner. Indeed, variations across subjects in C. albicans burden in response to β-lactam treatment could be partly explained by changes in the levels of endogenous fecal β-lactamase activity. This highlighted a potential new key factor for C. albicans overgrowth. Thus, the significance of our research is in providing a better understanding of the factors behind C. albicans intestinal overgrowth, which might lead to new means to prevent life-threatening secondary infections.

Keywords: Candida albicans; antibiotics; beta-lactamases; gut mycobiota; healthy individuals.

FIG 1

Dynamic of the mycobiota characteristics in 22 healthy individuals during a 2-week period. (A) Fungal species relative abundances at 1-week apart time points for 22 healthy subjects. For each subject, barplot are ordered by time (−D15, −D7, −D1 before antibiotics). Represented species reached a mean relative abundance across subjects above 1%. (B) Alpha diversity: violin plot of the number of OTUs and of the Shannon index values at 1-week apart time points for 22 healthy subjects. (C) Beta diversity: Bray-Curtis dissimilarity values between samples donated by different subjects (between subjects) and between samples donated by the same subjects (within subjects) for ITS1 sequencing data. Values range from 0 to 1, with 0 being the least dissimilar and 1 being the most dissimilar. (D) Violin and boxplots of the C. albicans DNA levels at 1-week apart time points for 22 healthy subjects. Each dot represents a sample. For all panels, the upper whiskers extend from the hinge to the largest value below 1.5× the interquartile range, and the lower whiskers extend from the hinge to the smallest value above 1.5× the interquartile range.

FIG 2

Impact of 3-day cefotaxime and ceftriaxone IV treatment on the gut mycobiota of 22 healthy subjects followed for a 6-month period. (A) Study design. (B) Log₁₀ (foldchange [FC]) of the fungal load following ceftriaxone and cefotaxime treatment (gray area). Thin lines represent the subjects, thicker lines represent the medians at each day for each treatment group (blue and green) and for all subjects (black). (C) Main fungal species distribution following ceftriaxone and cefotaxime treatment. (D) Distribution of the relative abundance log₁₀ (FC) AUCs for Candida albicans, Debaryomyces hansenii, Penicillium roqueforti, and Saccharomyces cerevisiae, highlighting the duration and the amplitude of the perturbations *q value < 0.05, Wilcoxon t-test, false-discovery rate correction. Upper whiskers extend from the hinge to the largest value below 1.5× the interquartile range, and the lower whiskers extend from the hinge to the smallest value above 1.5× the interquartile range. (E) Log₁₀ (FC) of C. albicans DNA levels following ceftriaxone and cefotaxime treatment (gray area). Thin lines represent the subjects, thicker lines represent the medians at each day for each treatment group (blue and green) and for all subjects (black).

FIG 3

Change in β-lactamases activity levels as a key parameter for C. albicans proliferation in the gut after third-generation cephalosporin administration. (A) β-lactamases activity before, during, and after the antibiotic treatment. Thin lines represent the subjects; thicker lines represent the medians at each day for each treatment group (ceftriaxone, blue; cefotaxime, green) and for all subjects (black). (B) Distribution of the D0 to D10 AUCs of the change of β-lactamases activity in 22 healthy subjects. The density distribution is represented by the green curve and the number of subjects for each range of AUC values are represented by the white histogram. The group “high” (orange) regroups the subjects with a D0 to D10 AUC above the median (black dashed line) and the group “low” (blue) regroups the subjects with a D0 to D10 AUC below the median. (C) Boxplots of C. albicans DNA levels log₁₀ (FC) following ceftriaxone and cefotaxime treatment. Orange boxplots indicates values for the subjects from the group “high” of and blue boxplots represents the values for the subjects from the group “low.” Upper whiskers extend from the hinge to the largest value below 1.5× the interquartile range, and the lower whiskers extend from the hinge to the smallest value above 1.5× the interquartile range. (D) Correlation plot of the log₁₀ (FC) AUCs of C. albicans DNA levels and β-lactamases activity for the D0 to D10 period. Regression is represented by a blue line and the confidence interval by the gray area. Subjects are designated by their ID number.