(541) Effect of therapeutic antibiotic exposure on the oropharyngeal and fecal microbiota in infants with cystic fibrosis

Background: Infants with cystic fibrosis (IwCF) frequently exhibit respiratory symptoms, which are treated with antibiotics that likely affect their gastrointestinal (GI) and oropharyngeal (OP) microbiota. Each of these is important: GI microbes moderate immune system development and nutrient acquisition, and OP samples are increasingly used as proxies for CF lower respiratory tract (LRT) microbiology. Post-antibiotic microbes at each site could also serve as reservoirs for antibiotic resistance determinants in their “resistomes”, with the potential for transmission to the LRT. The effects of antibiotic exposure on these microbial communities and their resistomes are undefined. Given that IwCF are relatively naïve to antibiotics, we studied the effects of one class of antibiotic – β-lactams – on the GI and OP microbiota in IwCF treated for respiratory symptoms.

Methods: We collected OP and fecal samples from 14 IwCF before, during and after β-lactam exposure, and from 4 control IwCF unexposed to antibiotics. Extraction of DNA from OP swabs included selective lysis of eukaryotic cells and endonuclease digestion to minimize human and extracellular bacterial DNA. We measured total bacterial load by 16S rDNA qPCR. We defined microbiota using shotgun sequencing with metagenomic phylogenetic analysis (MetaPhlAn3) and resistomes by sequence read mapping to the Comprehensive Antibiotic Resistance gene Database (CARD). We compared microbial taxonomic and gene abundances from infants ever versus never exposed to antibiotics by permutational multivariate analysis of variance (PERMANOVA) and defined microbiota dissimilarity using Aitchison distance.

Results: OP swab and fecal samples yielded averages of 12 million and 21 million microbial sequence reads, respectively. Total bacterial loads were significantly lower for fecal samples from infants ever versus never exposed to antibiotics (p=0.003), but there were no differences between OP swabs. Overall taxonomic and CARD gene abundances differed significantly by antibiotic exposure in both fecal (p=0.001 and p=0.01, respectively) and OP (p=0.001 and p=0.007, respectively) sample sets. Despite these differences, very few of the resistance genes that differed significantly between groups encoded canonical β-lactam resistance determinants (2/12 fecal, 0/19 OP).

Conclusions: Using methods targeting viable bacterial cells, OP swab samples, which are heavily contaminated by human DNA, generated sufficient sequence reads for metagenomic analysis. We found that the fecal, but not OP, total bacterial loads decreased in IwCF exposed to β-lactams, suggesting unintentional effects in the GI tract. We identified significant differences in both taxonomic constituency and antibiotic resistance gene complement among IwCF exposed to antibiotics for both fecal and OP samples. However, few of the resistance genes with significantly different abundances encoded β-lactam resistance determinants, suggesting that resistome variations with antibiotic treatment reflected taxonomic changes, rather than selection for β-lactam resistance. Our findings indicate that antibiotics alter the microbiota in both the upper respiratory and GI tracts of IwCF, and future studies of antibiotics in infants/young children can and should include samples from both body sites.

Acknowledgements: This work was supported by the NIH (K24HL141669, DK089507) and CF Foundation (SANDER18A0-1, SINGH19R0)

References: