NEW YORK (GenomeWeb) – A team in the Bacterial Diseases Branch of the Walter Reed Army Institute of Research recently became the first to detect a species of bacteria resistant to an antibiotic-of-last-resort called colistin in a US patient's specimen, potentially heralding an era of "superbugs."
The resistance was detected in a strain of Esherichia coli collected from a woman in Pennsylvania with a urinary tract infection, according to a report published in Antimicrobial Agents and Chemotherapy last month. The E. coli was found to contain the mcr-1 gene on a transmissible plasmid as well as genes conferring other types of antibiotic resistance.
It was not, however, a "superbug" itself, as it did not harbor resistance to several other classes of antibiotics. But if the mcr-1 gene is transferred and ends up on a resistance plasmid that contained genes encoding carbapenemases or 16s methylation, for example, it could be dire.
The particular plasmid type that contains the resistance elements is important to the epidemiology. Some plasmids are specialized to a particular species. Others are "very promiscuous [and] can be transferred between different species and at a much higher rate — they really move genes around quite quickly," said Erik Snesrud, a clinical research scientist in the Multidrug-resistant Organism Repository and Surveillance Network (MRSN) at Walter Reed and co-author on the study.
The E. coli in question wound up having two plasmids with different complements of other antibiotic resistance genes, including extended-spectrum beta-lactamase genes, but Snesrud emphasized that it was not a pan-resistant strain. The plasmid that contained the mcr-1, however, was of a type that is able to be transmitted readily.
Colistin was developed in the late 1950's but fell into disuse because of a potential to cause kidney damage. Resistant strains have existed for a long time, Snesrud said, but previously these were due to genetic mutations on the genome that could only be passed on to daughter cells. The mcr-1 on transferrable plasmids can instead be passed on to "other strains and other species that they come in contact with in the environment," he said.
The mcr-1 gene was first discovered in late 2015 in a Chinese pig, and the MRSN began routinely testing for it last month. Why it first showed up in a UTI in Pennsylvania is a mystery. "With the amount of interconnectedness in travel and the world food supply, these things move around the globe very fast and in very unexpected ways," Snesrud said.
Typically, bugs that are isolated from UTIs will also show up in surveillance swabs of the patient, Snesrud said, and they are often colonized in their gastrointestinal tract.
While the US Centers for Disease Control and Prevention covers antibiotic surveillance testing and general public health outside of the military healthcare system, the MRSN group covers all Army, Navy, and Air Force sites around the globe.
"Any antibiotic-resistant bacteria that show up in a clinical laboratory at any of the military's hospitals or clinics come to our repository for characterization," Snesrud said.
To date, the repository contains over 40,000 samples that have been phenotypically characterized as antibiotic resistant. About 3,000 isolates have been selectively sequenced for different projects, Snesrud said. The group also receives hundreds of samples per month from military hospitals that must be characterized and added to the collection.
Notably, beginning this fall the CDC will provide new infrastructure and lab capacity to detect and respond to resistant organisms recovered from human samples, according to a Health and Human Services statement on antibiotic resistance detection. More than 44,000 Salmonella and 9,000 E. coli/Shigella have been tested in roughly the past two months and none has shown the presence of the mcr-1 gene, the agency said.
The MRSN labs use a number of different technologies in their work. For antibiotic resistance testing, they employ all three of the major antibiotic susceptibility testing machines — the Becton Dickinson Phoenix, the Biomerieux Vitek, and the Beckman Coulter Microscan, Snesrud said.
"Different hospitals tend to pick different machines, so in order for our results to be universal and comparable to everybody else we are running each [sample] that comes in on all three of the machines," he said. These provide susceptibility profiles as well as species identification using non-molecular methods by essentially running a panel of biochemical tests to see what the strain can and cannot metabolize.
The lab also performs various other assays, such as a MALDI-based species identification, and the BioMérieux E-test to determine minimum inhibitory concentration of antibiotics that will kill the bacteria. The E-test was used to show collistin resistance in the mcr-1-containing E. coli, and presence of the gene was then confirmed quickly using a recently published PCR assay.
Sequencing is used to ID questionable cases, but also to provide greater detail on the location of the resistance genes. For whole-genome sequencing the lab deploys Illumina's MiSeq and NexSeq, as well as Pacific Biosciences' RS II system.
Plasmids present a tricky problem, however, because resistance genes tend to be located in resistance islands that are associated with insertion sequences and mobile genetic elements, Snesrud said.
These often have extensive repetition in the sequence in those regions. "The assemblies tend to get broken up in to a lot of little pieces, so what would typically happen is once you go and do the sequencing and try to assemble things, your gene of interest assembles as a contig, but you don't really have much idea about what's around it or what plasmid it's on," Snesrud said. On the other hand, the lab has to balance a need for longer reads with a need for higher throughput.
Specifically, the longest reads on the Illumina machines is in the 600 base-pair range, while the PacBio system can plow through 50 kilobases in a single read. However, that system can only handle eight to 16 samples per run, while the MiSeq and NextSeq process 36 and up to 350 samples, respectively, and the PacBio system is also somewhat more expensive per run. So, the choice of sequencer for a given sample depends on the nature of the project, Snesrud said.
"With the mcr-1 project it was important to understand the resistance plasmid that was carrying the gene," he said, so the group turned to the PacBio sequencer. For other projects, such as outbreak investigations, short-read sequencers are sufficient to tell strains apart.
Snesrud noted that MRSN has been keeping an eye on the Oxford Nanopore MinIon for ultra-long read sequencing, and has been waiting for a new upgrade for the system's sequencing chemistry to launch. "It looks like they have broken into the sequencing quality level that it might be on par with what PacBio has been doing for a while," Snesrud said. But at this point, the lab has made its investment in the RS II and it is "running all the time" and providing "great results."