Abstract:
The human gut microbiome consists of bacteria, archaea, fungi, and viruses that
live in the gastrointestinal tract. It plays an important role in host physiology, and its
perturbation has been associated with various diseases. One of its key functions is to
protect from colonization and infection by incoming pathogenic bacteria, a process
termed colonization resistance. Sequencing-based microbiome research has identified
medication as one of the major factors that influences the gut microbiome composition.
While antibiotics have long been recognized for their collateral damage to the gut
microbiome, recent research has revealed that non-antibiotic drugs for human use
(human-targeted drugs) can similarly inhibit gut commensals.
In this thesis, I investigated the interaction of both antibiotics and non-antibiotics
with the human gut microbiota, focusing on their influence on colonization resistance
and implications for microbiome-focused therapies. I standardized and optimized a
high-throughput anaerobic screening protocol capable of assessing approximately
5,000 drug-microbe interactions within five days. Using synthetic gut microbial
communities and stool-derived communities, I showed that non-antibiotic drugs from
different classes can break colonization resistance against several pathogens:
Gammaproteobacteria, C. difficile, and vancomycin-resistant Enterococcus faecium
(VRE). Community biomass, composition, and the presence of direct niche competitors
were major factors influencing colonization resistance against all pathogens.
Importantly, I found that direct drug-microbe interactions were not the sole
contributors to altered colonization resistance. Specifically, I demonstrated that the
frequently reported association between proton pump inhibitors (PPIs) and C. difficile
infection appears to be driven primarily by drug-induced increases in gastrointestinal
pH rather than direct drug-mediated microbial inhibition. Further, using the highthroughput screening platform, I discovered that clinical isolates of VRE exhibited
increased sensitivity to non-antibiotics despite their high resistance to antibiotics.
Notably, the antimetabolite gemcitabine demonstrated high efficacy and selectivity in
depleting Enterococci, including clinical VRE isolates, from synthetic and stool-derived
communities, highlighting opportunities for drug repurposing.
Regarding the collateral damage of antibiotics on gut commensals, three drugs
̶ dicumarol, benzbromarone, and tolfenamic acid ̶ were identified as antagonists of
macrolide activity against Bacteroidales species. I showed that this mitigation of the VII
collateral damage of antibiotics is selective, retaining antibiotic activity against
pathogens, and is expandable from monocultures to synthetic and stool-derived
communities. I uncovered that these human-targeted drugs act by upregulating a
resistance-nodulation-division (RND)-type efflux pump in Bacteroidales, and that this
is strain-specific due to differences in drug susceptibility and efflux capabilities. While
this strategy suggests new opportunities for microbiome-protective therapies, it also
raises concerns about the unintended promotion of antimicrobial resistance through
the induction of microbial defense mechanisms by non-antibiotic drugs.
In conclusion, this thesis advances our understanding of how non-antibiotic
drugs influence the human gut microbiome, revealing overlooked risks for pathogen
colonization and antimicrobial resistance alongside new opportunities for drug
repurposing to protect or restore microbiome health.
Infection risk