Impact of the Gram-Negative-Selective Inhibitor MAC13243 on In Vitro Simulated Gut Microbiota

Abstract

New Gram-negative-selective antimicrobials are desirable to avoid perturbations in the gut microbiota leading to antibiotic-induced dysbiosis. We investigated the impact of a prototype drug (MAC13243) interfering with the Gram-negative outer membrane protein LolA on the faecal microbiota. Faecal suspensions from two healthy human donors were exposed to MAC13243 (16, 32, or 64 mg/L) using an in vitro gut model (CoMiniGut). Samples collected 0, 4, and 8 h after exposure were subjected to viable cell counts, 16S rRNA gene quantification and V3-V4 sequencing using the Illumina MiSeq platform. MAC13243 exhibited concentration-dependent killing of coliforms in both donors after 8 h. Concentrations of ≤32 mg/L reduced the growth of aerobic bacteria without influencing the microbiota composition and diversity. An expansion of Firmicutes at the expense of Bacteroidetes and Actinobacteria was observed in the faecal microbiota from one donor following exposure to 64 mg/L of MAC13242. At all concentrations tested, MAC13243 did not lead to the proliferation of Escherichia coli nor a reduced abundance of beneficial bacteria, which are typical changes observed in antibiotic-induced dysbiosis. These results support our hypothesis that a drug interfering with a specific target in Gram-negative bacteria has a low impact on the commensal gut microbiota and, therefore, a low risk of inducing dysbiosis.

Keywords: Escherichia coli; LolA; antibiotic targets; microbiome; selective antimicrobial activity.

Figure 1

Viable cell counts of coliforms on MacConkey agar (A,C), and of total bacteria on Brain Heart Infusion agar (B,D) in faecal samples from two human donors after exposure to MAC13243. The limit of detection (200 CFU/mL) is highlighted with a dashed line.

Figure 2

Quantification of total bacteria populations using qPCR of the 16S rRNA gene.

Figure 3

Two-dimensional non-metric multidimensional scaling (nMDS) plot of the microbial community compositions in the donor samples. An nMDS plot based on the Bray–Curtis dissimilarity matrix was used to simultaneously visualise individual samples (dots) originating from donor 1 (D1, red) and donor 2 (D2, blue). Sample clustering of the two donors was significantly different (PERMANOVA, p = 0.001).

Figure 4

Effects of MAC13243 on microbiota composition and diversity in faecal samples from two human donors; (A) relative abundance at the phylum level; (B) alpha diversity; (C) bubble plots showing the proportion of the highest abundant genera across all samples.

Figure 5

Rooted tree based on the alignment of the LolA protein sequences of the main Gram-negative bacteria. Bootstrap values of >70% (1000 replicates) are illustrated by black, filled circles at the nodes. The scale bar indicates the expected number of substitutions per site.