Detection and Whole-Genome Sequencing of Carbapenemase-Producing Aeromonas hydrophila Isolates from Routine Perirectal Surveillance Culture

Abstract

Perirectal surveillance cultures and a stool culture grew Aeromonas species from three patients over a 6-week period and were without epidemiological links. Detection of the bla_KPC-2 gene in one isolate prompted inclusion of non-Enterobacteriaceae in our surveillance culture workup. Whole-genome sequencing confirmed that the isolates were unrelated and provided data for Aeromonas reference genomes.

TEXT

Aeromonas species are Gram-negative environmental bacteria that can cause infections in both immunocompetent and immunocompromised persons (1, 2) and have been isolated from a wide range of aquatic environments, including marine ecosystems, freshwater, hospital effluent, and drinking water (1, 3–6). The global rise of multidrug-resistant bacteria raises concern that a ubiquitous organism, such as an Aeromonas sp., could acquire antibiotic resistance genes derived from human isolates (3, 4, 7, 8). A variety of antibiotic resistance phenotypes, including carbapenem resistance (3, 4, 9–15), have been found in both clinical and environmental Aeromonas isolates. Although rare, clinical Aeromonas isolates containing bla_KPC genes have been reported (16, 17).

Active microbial surveillance for multidrug-resistant Gram-negative organisms is integral to inpatient care at the NIH Clinical Center. Perirectal swabs are collected upon hospital admission, twice weekly in the intensive care unit, and monthly from all inpatients (excluding those in behavioral health wards). Perirectal swabs were inoculated onto HardyCHROM CRE (Hardy Diagnostics, Santa Maria, CA) and incubated at 35°C ambient air for 18 to 24 h. Isolates were identified using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS), and isolates belonging to Enterobacteriaceae underwent PCR for detection of bla_KPC and bla_NDM-1. Stool specimens were plated onto sheep blood, MacConkey, and xylose lysine deoxycholate agars with incubation at 35°C and 5% CO₂ for 48 h, and on Campy CVA agar with incubation at 42°C under microaerophilic conditions. Results from the surveillance cultures that suggest possible nosocomial acquisition of carbapenemase-producing bacteria trigger epidemiological investigations, including environmental sampling.

Perirectal surveillance cultures from two hospitalized adult patients and a stool culture from a third patient grew Aeromonas spp. within a 6-week period. The stool culture isolate (from patient A) was resistant to ertapenem and susceptible to imipenem and meropenem; the perirectal isolates (from patients B and C) were resistant to all three carbapenems. MALDI-TOF MS could not reliably identify the isolates to the species level. On average, clinical Aeromonas spp. had been isolated from clinical cultures in only four to five patients per year in our institution since 2010; this was the first time that Aeromonas had been isolated from perirectal surveillance cultures since they were introduced in 2011. This apparent cluster of Aeromonas colonization prompted an investigation. A patient trace (TheraDoc) revealed that patients A and B had each received care in the intensive care unit for 13 days but were never housed in the same room. Both were subsequently transferred to different rooms on another medical ward; patient B was discharged from ward room 1 five days later, and patient C was admitted to ward room 1 seven weeks after patient B had vacated the room (Fig. 1). An extensive chart review identified no other epidemiological links among the three patients, and extensive environmental samples, including sink water and swabs of faucet aerators for culture on multiple media, were negative for Aeromonas spp. and carbapenem-resistant organisms.

FIG 1.

Patient trace.

To investigate possible genetic relatedness, whole-genome sequencing was performed using Illumina MiSeq and assembled with MIRA for all three Aeromonas isolates (18). 16S rRNA sequences indicated that patient B's isolate was related to Aeromonas hydrophila, while isolates from patients A and C were most closely related to Aeromonas veronii, thus narrowing the epidemiological investigation.

The A. hydrophila isolate from patient B (AHNIH1) was sequenced on the PacBio RSII instrument to generate a fully contiguous genome. AHNIH1 has a 4.91-Mb chromosome and carries a 143-kb plasmid (pASP-135) (Fig. 2). Unexpectedly, this plasmid carries a bla_KPC-2 gene (retrospectively confirmed by bla_KPC PCR) and TEM β-lactamase and genes encoding resistance to aminoglycosides, chloramphenicol, fluoroquinolones, macrolides, and mercury. We speculate that patient B acquired the isolate from an aqueous source within the hospital environment; however, the origin of the bla_KPC-2 plasmid is unknown, as its sequence does not match either those of other bla_KPC-2 bacteria identified in our institution or other publically available references. The bla_KPC-2 gene is not carried on the commonly associated Tn4401 element and is instead flanked by IS26 elements, which are important factors in the mobility of resistance determinants (19). All three isolates contain chromosomally encoded homologs of CphA, a metallo-β-lactamase that is a frequent etiology of carbapenem resistance in Aeromonas species (20).

FIG 2.

The pASP-135 plasmid from the A. hydrophila isolate has short regions of identity to previously sequenced plasmids (pKOX-86d and pENT-d0d), largely demarcated by mobile elements (in brown). Genes are colored by function: conjugation, dark green; antibiotic resistance, yellow; metal resistance, blue; transposase/integrase/resolvase, brown; hypothetical, gray; other, light green. IS26 elements are highlighted on pASP-135 with red underlining. Aligned regions between plasmids are indicated by gray ribbons, with percent identities indicated.

To explore a possible epidemiological link between the A. veronii isolates from patients A and C, we undertook a genetic analysis. As no finished reference genome for A. veronii is available, we generated a fully contiguous reference genome on the PacBio RSII instrument. The A. veronii isolate from patient C (AVNIH1) has a 4.76-Mb chromosome and a 198-kb plasmid (pASP-a58) carrying genes encoding resistance to aminoglycosides, β-lactams, fluoroquinolones, macrolides, mercury, and rifampin, possibly explaining the recovery on selective surveillance media. The genome of the A. veronii-like isolate from patient A (AVNIH2) was assembled using AVNIH1 (from patient C) as the reference. The isolates from patients A and C are only 98.7% identical across aligned regions, excluding the possibility of transmission between these two patients. Neither isolate was found to contain a bla_KPC gene.

Although epidemiological data alone were unable to establish or refute a link among the three Aeromonas isolates, despite the temporal clustering, whole-genome microbial sequencing established that these isolates were unrelated to each other and did not represent a true nosocomial cluster. The ability of sequencing data to resolve this epidemiological uncertainty enabled us to avoid further investigation, which would likely have been fruitless, labor-intensive, and expensive.

Analysis of sequencing data demonstrated that the Aeromonas isolates in this report belong to two different species, and we unexpectedly found that one isolate carried the bla_KPC-2 gene on a plasmid. Few complete genome sequences exist for Aeromonas spp., making these a valuable addition to the reference database.

Hospital surveillance for bla_KPC bacteria should not be limited to Enterobacteriaceae, as the bla_KPC carbapenemase has been reported in numerous species belonging to a wide range of other bacteria, including non-Enterobacteriaceae (e.g., Pseudomonas, Aeromonas, and Acinetobacter) (21–23). As a result of this case, the standard workup of bacterial colonies growing on perirectal surveillance cultures at our institution now includes bla_KPC and bla_NDM screening for all Gram-negative bacilli, not limited to the Enterobacteriaceae.

In summary, the detection of bla_KPC in a surveillance Aeromonas isolate led to a change in our hospital's multidrug-resistant Gram-negative surveillance protocol. Whole-genome microbial sequencing suggested that nosocomial transmission did not occur, prevented a more laborious evaluation, and provided reference genome sequences that might be useful for future epidemiological investigations.

Nucleotide sequence accession number.

The whole-genome sequencing data can be retrieved at NCBI BioProject no. PRJNA279868.