Retrospective analysis of multidrug-resistant clinical and environmental isolates for the presence of the colistin-resistance gene mcr-1

To the Editor—Carbapenem-resistant Enterobacterales (CRE) are a public health threat due to increased mortality, cost, and transmissibility of these infections. Although colistin is rarely considered as a last-resort antibiotic to treat CRE infections, increasing reports of plasmid-mediated colistin-resistant CRE isolates worldwide ¹ are concerning. Resistance to colistin is conferred by mcr genes and was first linked to the mcr-1 gene. ² Since the initial mcr-1 report from 2005, researchers have screened isolate collections for mcr-1 and have found plasmid-mediated colistin resistant strains in animal and human populations. ² At the time of this study, 4 clinical isolates containing the mcr-1 gene were identified in the United States, ³ which has since increased to 55 isolates in at least 21 states. ⁴ The overall prevalence, distribution, and impact of mcr-1 remain unclear.

In North Carolina, more than half of hospitals have reported CRE infections, ⁵ yet resistance to colistin has not been systematically examined. Given the active agriculture industry within the state, the potential to identify mcr-1 among clinical and environmental isolates may be higher with more opportunity for transmission between animal and human populations. To address this, we screened clinical and environmental isolates for mcr-1 to determine the prevalence and dissemination of colistin resistance among MDR organisms.

Methods

Study inclusion criteria included an isolate (1) identified as a member of Enterobacterales via biochemicals or mass spectrometry, (2) with antimicrobial susceptibility results, and (3) with drug resistance to ≥3 drug classes (ie, multidrug-resistant or MDR). Clinical isolates were obtained from the UNC Clinical Microbiology Laboratory and were determined to be MDR or CRE as part of routine patient care. Additional clinical isolates were obtained from archived MDR isolates collected from surveillance swabs of burn unit patients. Bacterial isolates were collected from December 2015 to May 2018.

To determine whether MDR strains containing mcr-1 were circulating in the environmental microbial population, we screened isolates collected from surface water of industrial hog operations in eastern North Carolina, and drug-resistant Escherichia coli isolates were collected from UNC Hospitals and Chapel Hill–Carrboro community sewage. Based on previous studies that screened clinical isolate collections, we expected the prevalence of mcr-1 to be ∼1%. ⁶ Thus, by screening clinical and environmental MDR isolates, we sought to increase the likelihood of identifying mcr-1–positive strains.

To verify individual isolates containing the mcr-1 gene could be detected when pooled and screened, we established the limit of detection (LOD) using 2 mcr-1–positive E. coli strains as controls. We made a dilution series, performed polymerase chain reaction (PCR), 100 µL of each dilution was plated onto sheep blood agar, and incubated overnight at 35°C to calculate colony-forming units (CFU/mL). We were able to reproducibly detect mcr-1 if present in an individual sample at ≥925 CFU/mL, which is considerably lower than the CFU present in pools of bacterial isolates.

Archived isolates were cultured at 35°C on MacConkey agar. To extract DNA, single colonies were placed in nuclease-free water, boiled for 10 minutes, and centrifuged; this extract was used as the PCR template. Real-time PCR was performed using the 2x Taq Man DNA Universal Mastermix kit (Applied Biosystems, Foster City, CA), and primers and probes (Biosearch Technologies, Petaluma, CA) previously published to detect DNA internal to the mcr-1 gene. ⁷ To screen a large number of isolates in a cost-effective manner, extracted DNA was pooled into groups of 20 isolates and tested for the presence of mcr-1. Positive pools were retested in individual PCR reactions to identify positive isolates. Two E. coli strains (0494 and 0495) containing the mcr-1 gene from the CDC and FDA Antibiotic Resistance Isolate Bank served as positive controls.

Results

In total, 711 clinical and environmental MDR isolates were screened, including 638 clinical patient isolates, 17 industrial hog operation and 1 control site MDR or confirmed β-lactamase–producing isolates, and 55 hospital and municipal sewage isolates. Among the 711 isolates, 677 isolates met the study inclusion criteria and grew in culture from frozen stock. Clinical isolates grew from a range of culture types including 389 isolates from urine, 44 from blood, 34 aerobic/anaerobic isolates, 21 respiratory isolates, 1 sterile fluid isolate, and 124 isolates from burn unit surveillance. All 677 isolates were tested by PCR, and none were positive for the mcr-1 gene.

Discussion

Colistin is one of few remaining antibiotics to treat MDR infections. Thus, understanding dissemination of mcr-1–mediated colistin resistance in microbial populations is vital. Although we did not identify any mcr-1–containing isolates in our screen, the lack of detectable colistin resistance still provides valuable information for patient and public health. This absence suggests a low prevalence of mcr-1–mediated colistin resistance in the microbial population circulating in North Carolina. In March 2019, the first clinical isolate positive for mcr-1 in North Carolina was detected, ⁸ further supporting our screen findings and low prevalence in the state.

Table 1. Summary of Clinical and Environmental Isolates Screened for mcr-1–Mediated Colistin Resistance

^a Met MDR criteria and grew in culture from frozen stock.

^b Resistant to all 5 drug classes tested, including (1) extended-spectrum cephalosporins (cefepime, ceftriaxone); (2) fluoroquinolones (ciprofloxacin, levofloxacin); (3) aminoglycosides (gentamicin, tobramycin, amikacin); (4) carbapenems (meropenem, ertapenem); (5) piperacillin/tazobactam (if ceftriaxone was intermediate or resistant, piperacillin/tazobactam was considered resistant).

^c Screened and isolated based on extended-spectrum β-lactamase (ESBL)– and Klebsiella pneumoniae carbapenemase (KPC)–mediated resistance, not evaluated for extensive resistance.

Very low prevalence of mcr-1 is consistent with findings in similar studies, such as a screen of 1,000 environmental Shiga toxin–producing E. coli isolates in the agricultural region of California, identified by PCR that did not detect mcr-1 or mcr-2. ⁹ Similarly, a screen of cecal contents in >2,000 food animals identified only 2 samples (<0.1%) positive for mcr-1. ¹⁰

Our report provides the first assessment of mcr-1 dissemination in North Carolina. The absence of mcr-1 positive isolates detected in clinical and environmental settings provides valuable data to inform clinicians, pharmacists, and epidemiologists of the risk associated with colistin in the treatment of MDR organisms. Transmission of colistin resistance via mcr-1 remains a significant threat worldwide and highlights the importance of continued surveillance to track the spread of MDR organisms.