Rapid detection of bacteria and antimicrobial resistant genes in intraamniotic infection using nanopore adaptive sampling

Intraamniotic infection, a major cause of the preterm labor syndrome, is a risk factor for neonatal morbidity and mortality and for long-term complications.1Romero R. Dey S.K. Fisher S.J. Preterm labor: one syndrome, many causes.Science. 2014; 345: 760-765Crossref PubMed Scopus (1328) Google Scholar Rapid identification of the microorganisms involved in intraamniotic infection is crucial because detection could influence clinical decisions, such as the administration of tocolysis and antimicrobial agents and neonatal care. However, the gold standard for the diagnosis of infection in clinical medicine still relies on cultivation, and such methods require days to isolate bacteria, which deems culture unsuitable to inform clinical decisions in obstetrics. In addition, some bacteria cannot be easily cultured in the laboratory, which may lead to false-negative diagnoses. There is an urgent need for rapid identification of microorganisms in time for clinical decision-making and for selection of antimicrobial agents. We recently reported on the rapid diagnosis of intraamniotic infection with full-length 16S ribosomal RNA (rRNA) gene sequencing (Oxford Nanopore Technologies [ONT], Oxford, United Kingdom). However, this method does not provide information about antimicrobial susceptibility patterns of bacterial species because only 1 conserved gene is used for species identification.2Chaemsaithong P. Romero R. Pongchaikul P. et al.Rapid diagnosis of intra-amniotic infection using nanopore-based sequencing.J Perinat Med. 2023; 51: 769-774Crossref PubMed Scopus (2) Google Scholar A recent technological advance—adaptive sampling, a bioinformatics-controlled function specific to the ONT platform—allows direct target reduction of undesirable host DNA molecules during sequencing and enables enrichment of any DNA sequence of the pathogen's genome, including antimicrobial resistance genes located on bacterial chromosomes or plasmids.3Martin S. Heavens D. Lan Y. Horsfield S. Clark M.D. Leggett R.M. Nanopore adaptive sampling: a tool for enrichment of low abundance species in metagenomic samples.Genome Biol. 2022; 23: 11Crossref PubMed Scopus (40) Google Scholar In this study, we report that adaptive sampling, also known as selective sequencing, can identify the microorganisms in silent intraamniotic infection and the characteristics of antimicrobial susceptibility for bacteria. A 38-year-old gravida 4, para 0030 at 30+5 weeks of gestation presented with preterm prelabor rupture of the membranes (PROM) at the delivery unit. Ultrasound examination showed normal fetal anatomy with anhydramnios and partial placenta previa. The patient had an elevated white blood cell count (26,110 cells/μL; normal range, 5600–15,000 cells/μL) and an elevated C-reactive protein (CRP) level, (69.91 mg/L; normal range, 0.15–1.5 mg/L). Antenatal glucocorticosteroid treatment was administered to promote fetal lung maturity, and antibiotics (ampicillin and erythromycin) were given to prolong the latency period. At 31+1 weeks of gestation, the patient developed regular uterine contractions every 5 minutes. Placenta previa prompted a cesarean delivery. A female neonate was delivered with Apgar scores of 7 and 9 at 1 and 5 minutes after delivery, respectively; the birthweight was 1820 grams. During surgery, malodorous, yellowish pus was found in the fetal nostrils (Figure). The microbiologic work-up of the pus included a Gram stain, culture, and imaging by scanning electron microscopy. Swabs of the uterine cavity and the space between the chorioamniotic membranes were obtained and a placental histologic examination was performed. DNA extracted directly from the pus sample was used to prepare a DNA library without polymerase chain reaction amplification; the library (SQK-LSK114; ONT) was loaded into the nanopore flow cell and inserted into a GridION device (ONT) to run the adaptive sampling mode for human depletion, which is selective sequencing that targets microbial DNA and ignores human DNA. Afterward, we stopped the adaptive sampling and sequenced the same library without depletion for 12 hours (Figure). The Gram stain of the pus specimen revealed short Gram-negative rods and coccobacilli forms with multiple neutrophils, and the scanning electron microscopy showed multiple rod-shaped bacteria (Figure). The intrauterine swab, the nasal pus specimen, and the culture of the chorioamniotic membranes showed no bacterial growth. The mother recovered from surgery without complications. The neonate was at risk for neonatal sepsis. However, because of the maternal clinical risk factor, the neonate's white blood cell count (12,350 cells/μL; normal range, 9000–30,000 cells/μL) and CRP level (1.24 mg/L; normal range, 1.5–20 mg/L) were within the normal range, and the blood culture showed no bacterial growth. Ampicillin and gentamicin were administered for 7 days. The neonate had symptomatic hypoglycemia, retinal hemorrhage, and jaundice of prematurity. She remained in the neonatal intensive care unit for 35 days. Placental histopathology revealed acute histologic chorioamnionitis (stage 3, grade 3) with acute funisitis (Supplemental Figures 1 and 2). Within 15 minutes of nanopore sequencing with adaptive sampling, the top bacteria identified were Prevotella species, namely Prevotella histicola, Prevotella intermedia, and Prevotella jejuni. After 2 hours of adaptive sequencing, we identified Prevotella bivia at around 20× genome coverage. Subsequently, after 12 hours of sequencing, we obtained 80× coverage of the P. bivia genome, which contained cfxA3 (100% identity) or β-lactamase antimicrobial resistance genes. By contrast, after 2 hours of sequencing without adaptive sampling, we obtained reads of about 5× depth coverage, which were not sufficient for good-quality genome assembly (Figure). In this study, we report that nanopore adaptive sampling can be used to sequence DNA isolated from pus, without amplification, to rapidly identify bacteria in intraamniotic infection. In addition, we found that pathogen enrichment by adaptive sampling combined with nanopore sequencing detected antimicrobial resistance genes such as cfxA3. These findings are important because knowledge of the presence of specific bacterial species and antimicrobial resistance genes can guide decision-making to deliver or to treat intraamniotic infection with a particular antibiotic. In addition, the precise identification of a microorganism and its antimicrobial resistance gene profile could be helpful to obstetricians and neonatologists to tailor antimicrobial agents appropriate for each newborn. In this case, analysis of amniotic fluid was not possible because of anhydramnios. Therefore, the diagnosis of intraamniotic infection was made based on the presence of bacteria and acute histologic chorioamnionitis and funisitis, which reflect a maternal and a fetal inflammatory response, respectively.4Kim C.J. Romero R. Chaemsaithong P. Chaiyasit N. Yoon B.H. Kim Y.M. Acute chorioamnionitis and funisitis: definition, pathologic features, and clinical significance.Am J Obstet Gynecol. 2015; 213: S29-S52Abstract Full Text Full Text PDF PubMed Scopus (600) Google Scholar