Oral Concurrent 3 Thursday, February 14 • 1:15 PM - 4:00 PM • Augustus Ballroom 5-6 • Caesars Palace| Volume 220, ISSUE 1, SUPPLEMENT , S30-S31, January 01, 2019

39: Amniotic fluid contains detectable microbial DNA that significantly differs from appropriate contamination controls


      A growing body of literature, including data from multiple labs, is emerging that suggests the intrauterine environment, including amniotic fluid, may contain unique microbial communities. However, despite use of appropriate controls, this work is criticized as being confounded by contaminants. Using sterile techniques and negative controls, we sought to assess whether a microbial signature exists in a large prospective cohort of subjects amniotic fluid obtained during amniocentesis during early-to-mid second trimester.

      Study Design

      As part of a feasibility study we performed targeted amplicon analysis of the V4 region of the 16S rRNA gene on amniotic fluid collected from gravidae at the time of genetic amniocentesis (gestational ages 16- 22wks; feasibility n=95, total cohort n=731). Microbial DNA was extracted from samples and negative controls (processed in parallel or archival, n=32), and submitted for Illumina MiSeq high-throughput sequencing. Sequences were processed with DADA2 to generate an amplicon sequence variant (ASV) table that contains highly resolved taxonomic classifications of microbes. Prevalence-based filtering of putative DNA contaminant sequences using decontam was performed.


      We initially identified 439 genera present in amniotic fluid. Multidimensional scaling (Bray-Curtis) revealed amniotic fluid samples to discretely separate from negative controls (PERMANOVA, p<0.001)(Fig 1A). Using prevalence-based filtering we identified up to 44 genera that represent putative contaminants. Removal of contaminant ASVs resulted in minimal alterations in the abundance of the top twenty-five most abundant taxa or overall microbial community structure (Fig 1B). Post-filtering, when compared to DNA extraction kit negatives, ten of the twenty-five most abundant taxa were classified as indicator taxa (p<0.05) including Staphylococcus, Streptococcus, Lactobacillus, and Bifidobacterium spp. (Fig 1C). Differential abundance testing revealed 9 ASVs that were significantly enriched based on gestational timepoint at which the fetal urine begins contributing to the amniotic fluid (Fig 1D).


      Amniotic fluid contains quantifiable levels of microbial DNA and clusters separately from negative controls. Furthermore, we identified highly prevalent ASVs belonging to biologically relevant taxa implicated as important in the postnatal gut health of the newborn, including Lactobacillus and Bifidobacterium spp.
      Figure thumbnail fx1