The dominoes fall after long-term antibiotic exposure

Abstract

Broad-spectrum antibiotics should prevent disease, right? In this issue of Cell Host & Microbe, Drummond et al. turn logic on its head and show they actually drive more deadly invasive fungal-bacterial systemic co-infection. Prophylactic antibiotics increase susceptibility to these infections by targeting the commensal microbes required for gut-derived IL-17-mediated immunity.

Since their discovery, antibiotics have saved billions of lives from lethal bacterial infections, marking them as one of the most successful weapons in the battle against microbial pathogens. However, accumulating evidence from both clinical observations and experimental animal models suggests there is collateral damage from long-term broad-spectrum antibiotic exposure, largely due to the disruption of gastrointestinal (GI) commensal microbiome. Commensal bacteria suppress intestinal pathogenic microbes directly by secreting inhibitory compounds and indirectly by inducing antimicrobial peptides from the host (Fan et al., 2015; Kim et al., 2019). Some metabolites from commensals, such as short chain fatty acids, are vital in the maintenance of barrier integrity and the development of certain types of immune cells (Dupraz et al., 2021; Litvak et al., 2018). Therefore, inappropriate long-term administration of broad-spectrum antibiotics triggers the dominoes of microbial dysbiosis and host cell malfunction to fall, further exacerbating complicated medical conditions such as organ transplantation and systemic infection (Zhang et al., 2022).

In the context of fungal infections, especially invasive candidiasis, antibiotic exposure has long been recognized as a risk factor. The constant administration of anti-anaerobic antibiotics is required for specific-pathogen-free (SPF) mice to maintain high levels of Candida albicans intestinal colonization and is associated with intestinal expansion of Candida species (Koh et al., 2008; Zhai et al., 2020). The antibiotics deplete anaerobes, such as Lachnospiraceae and Bacteroidetes species, which normally drive the production of anti-Candida peptides by intestinal epithelial cells (Fan et al., 2015). This creates an intestinal niche for Candida to expand. Together with the loss of barrier integrity and defects in the immune system, intestinal-colonizing Candida translocate to the bloodstream to cause lethal systemic disease. Interestingly, in invasive candidiasis derived from non-GI sources, such as skin-colonized Candida, clinical observations also suggested that administration of antibiotics is associated with worse outcomes, with unknown underlying mechanisms (Pappas et al., 2018).

To bridge this knowledge gap, Drummond and colleagues employed a murine model of disseminated candidiasis to mimic human skin-derived candidiasis (Drummond et al., 2022). They treated mice with ampicillin, metronidazole, neomycin, and vancomycin (AMNV), a broad-spectrum antibiotic cocktail commonly used in laboratory research, followed by intravenous infections of C. albicans. The mice with AMNV exposure displayed worse survival compared with those without antibiotics. Strangely, the survival defect was not associated with the increased fungal burden in kidney or other sterile sites typically seen with candidemia-associated mortality. Instead, the authors found an increase of fungal burden in the GI-tract of the antibiotic treated animals, suggesting an impaired site-specific antifungal immunity. The authors further dissected the cellular compartments of the intestinal mucosa site and identified a dysfunction of Th17 immunity, marked by the reduction of the key cytokines IL17A, IL22, and granulocyte-macrophage colony-stimulating factor (GM-CSF). Adoptive transfer of lung and gut CD4⁺ and innate lymphoid cells from non-treated mice partially rescued this failure of Th17 immunity. The antibiotic regimen also led to dysbiosis of the GI commensal flora and induced the expansion of bacterial commensals and/or opportunistic pathogens, such as Enterobacteriaceae species. Together with the impaired local immunity, the expansion of intestinal bacteria eventually translocated to the bloodstream, which led to bacterial and fungal co-infections in the host. It is likely that these disseminated bacteria aggravate the fungal infection either directly or indirectly through immunopathology, leading to more and more rapid mortality of the mice after antibiotic exposure. The exacerbated mortality by a secondary bacteremia could be ameliorated by treatment with IL-17A or GM-CSF, further linking deficiencies of these cytokines in the GI track to mortality.

Interestingly, these AMNV cocktail-induced phenotypes can be largely replicated with oral vancomycin alone. Vancomycin is usually administered intravenously for treatment or prophylaxis purposes. The only scenario for oral administration of vancomycin is to treat Clostridium difficile infections. Clinical observations have shown that C. albicans and C. difficile co-infection exhibits substantial mortality (Russo et al., 2015). However, current clinical data of C. albicans and C. difficile co-infections usually lacks the sequential information of the two infections and information on whether the co-infection includes other bacterial pathogens. Drummond et al. mined the Cerner Health Facts database to test if there is a real-world link between antibiotic exposure and co-infections in invasive candidiasis. They found that systemic broad-spectrum antibiotic administration in patients with invasive candidiasis increases the risk of bacterial co-infection and overall mortality. Given that only oral (and not intraperitoneal) vancomycin treatment exacerbated candidiasis in mice, it is possible that oral antibiotic treatment of patients may have an even greater effect than the systemic antibiotics received by these patients. These data suggest that a more comprehensive and retrospective review of the clinical data from patients with C. albicans, C. difficile, and other bacterial infections is an important and useful tool to validate discoveries from animal models.

In this study, Drummond et al. elegantly dissected the sequential impacts of antibiotics in a murine model of candidiasis and identified gut-derived bacteremia as a potential cause for the increased mortality seen in candidiasis. Their work highlights striking gaps in our basic knowledge about how antibiotics alter human commensals and thereby both GI and systemic immunity. It also highlights a need to develop therapeutics to combat the negative immunological effects of gut dysbiosis. From a clinical perspective, these results reinforce the importance of antibiotic stewardship in high-risk patients, raising important next questions about route and choice of antibiotics for prophylaxis.