Rapid Detection of Polymyxin Resistance in Enterobacteriaceae

Patrice Nordmann; Aurélie Jayol; Laurent Poirel


Emerging Infectious Diseases. 2016;22(6):1038-1043. 

In This Article

Materials and Methods

Isolate Collection

To evaluate the performance of the rapid polymyxin NP test, we used 200 isolates collected from clinical samples worldwide. This collection included 135 Enterobacteriaceae isolates resistant to polymyxin: 5 isolates of intrinsically polymyxin-resistant species (Morganella morganii, Proteus mirabilis, Proteus vulgaris, Providencia stuartii, and Serratia marcescens) and 130 isolates of various enterobacterial species (Klebsiella spp., E. coli, Enterobacter spp., and Hafnia alvei) with acquired resistance to polymyxins (online Technical Appendix), including a previously reported heteroresistant Klebsiella pneumoniae isolate for which MIC for colistin was high.[12] The other 65 enterobacterial isolates belonged to various species and were susceptible to polymyxins (online Technical Appendix).

MIC Determination

To determine MICs for polymyxins, we used the broth microdilution method in cation-adjusted Mueller-Hinton broth (MHB-CA, reference 69444; Bio-Rad, Marnes-La-Coquette, France) as recommended by Clinical Laboratory Standard Institute (CLSI) guidelines.[13–15] We considered this method to be the standard for comparison with the rapid polymyxin NP results. Polymyxin B and colistin sulfate (Sigma-Aldrich, St. Louis, MO, USA) were tested over a range of dilutions (0.12–128 μg/mL). All experiments were repeated in triplicate in separate experiments. As recommended by CLSI, microdilution was performed without addition of Tween 80,[15] and E. coli ATCC 25922 was used as a control strain.

Because no breakpoint is available for polymyxins for Enterobacteriaceae according to CLSI guidelines,[14] we used the breakpoints of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) for reference.[16] Enterobacterial isolates with colistin or polymyxin B MICs <2 μg/mL were categorized as susceptible; those with MICs >2 μg/mL were categorized as resistant.

PCR Amplification and Sequencing

We recovered the chromosomal DNA of the isolates by using a commercially available kit (QIAquick; QIAGEN, Courtaboeuf, France) according to the manufacturer's instructions. We sequenced the pmrA, pmrB, phoP, phoQ, and mgrB genes possibly involved in colistin resistance in K. pneumoniae and K. oxytoca, as described previously.[12,17–19] We performed PCR amplification for detection of the plasmid-mediated mcr-1 gene as described.[7]

Isolate Genotyping by Pulsed-field Gel Electrophoresis

We assessed the genetic relatedness of the colistin-resistant isolates with identical molecular mechanisms of colistin resistance. We used pulsed-field gel electrophoresis with XbaI-digested genomic DNA as described previously.[20]

Rapid Polymyxin NP Test

Reagents and Solutions. The rapid polymyxin NP test uses 2 reagents and solutions: stock solutions of polymyxins and rapid polymyxin NP solution. Each is described below.

For stock solutions of polymyxins, colistin sulfate and polymyxin B powders (Sigma Aldrich) were diluted into MHB-CA medium in glass tubes to obtain a concentration of 0.2 mg/mL. These powders can be stored at 4°C before use, and the diluted polymyxin solutions can be kept at −20°C for 1 year. Of note, polymyxin-containing batches from commercial origin can be used, but the colistimethate sulfate powder, a therapeutic prodrug of colistin, cannot be used.

To prepare 250 mL of the rapid polymyxin NP solution, we mixed the culture medium and the pH indicator in a glass bottle as follows: 6.25 g of MHB-CA powder, 0.0125 g of phenol red (Sigma Aldrich), and 225 mL of distilled water. The pH of the solution was adjusted to 6.7 by adding drops of 1 mol/L HCL. This solution was then autoclaved at 121°C for 15 min. After cooling the solution to room temperature, we added 25 mL of 10% anhydrous d(+)-glucose (Roth, Karlsruhe, Germany) sterilized by filtration. The final concentrations in the rapid polymyxin NP solution were consequently 2.5% of MHB-CA powder, 0.005% of phenol red indicator, and 1% of d(+)-glucose. This rapid polymyxin NP solution can be kept at 4°C for 1 week or at −20°C for 1 year. This solution must be prewarmed at 37°C before use to prevent growth delay and therefore a delayed color change.

Just before performing the experiment, we added colistin to the rapid polymyxin NP solution and mixed it into sterile glass tubes to obtain a rapid polymyxin NP solution containing a final colistin concentration of 5 μg/mL. For example, we added 25 μL of colistin stock solution at 0.2 mg/mL to 1 mL of rapid polymyxin NP test solution for the testing of 1 isolate and respective negative and positive controls.

Bacterial Inoculum Preparation. We prepared a standardized enterobacterial inoculum by using freshly obtained (overnight) bacterial colonies grown on Luria-Bertani or Mueller-Hinton plates. We resuspended the bacterial colonies into 10 mL of sterile NaCL (0.85%) to obtain a 3.0–3.5 McFarland optical density (≈109 CFU/mL), which corresponds to an ≈10-μL full loop of bacterial colonies diluted in 10 mL of NaCL. A bacterial suspension was prepared for each isolate to be tested and for the colistin-susceptible and -resistant isolates used as controls controls (isolates FR-01 and FR-136; respectively; online Technical Appendix). As recommended by the EUCAST guidelines for susceptibility testing, we used the bacterial suspensions within 15 min of preparation and for no longer than 60 min after preparation.[16]

Tray Inoculation. We performed testing in a 96-well polystyrene microtest plate (round base, with lid, sterile, reference 82.1582.001; Sarstedt, Nümbrecht, Germany). For each isolate, bacterial suspension was inoculated in parallel into 2 wells, with and without colistin. The following steps of the rapid polymyxin NP test were then performed (Figure):


Representative results of the rapid polymyxin NP [Nordmann/Poirel] test. Noninoculated wells are shown as controls (first column). The rapid polymyxin NP test was performed with a reference colistin-susceptible isolate (second column) and with a reference colistin-resistant isolate (third column) in a reaction medium without (upper row) and with (lower row) colistin. The tested isolate grew in the presence (and absence) of colistin (wells B4 and A4, respectively) and was therefore reported to be colistin-resistant.

  • Step 1: 150 μL of colistin-free solution was transferred to wells A1–A4.

  • Step 2: 150 μL of the rapid polymyxin NP solution containing colistin was transferred to wells B1–B4.

  • Step 3: 50 μL of NaCl 0.85% was added to wells A1 and B1.

  • Step 4: 50 μL of the colistin-susceptible isolate suspension used as negative control was added to wells A2 and B2.

  • Step 5: 50 μL of the colistin-resistant isolate suspension used as positive control was added to wells A3 and B3.

  • Step 6: 50 μL of the tested isolate suspension was added to wells A4 and B4.

We mixed the bacterial suspension with the reactive medium by pipetting up and down. The final concentration of bacteria was ≈108 CFU/mL in each well, and the final concentration of colistin sulfate was 3.75 μg/mL.

Tray Incubation. We incubated the inoculated tray for up to 4 h at 35 ± 2°C in ambient air, without being sealed and without agitation. We did not seal the tray because oxygen is required for carbohydrate metabolism.

Tray Reading. We visually inspected the tray (checked for no spontaneous color change) after 10 min and then every hour for 4 h. We considered the test result positive (polymyxin resistance) if the polymyxin-resistant isolate grew in presence of colistin and negative (polymyxin susceptibility) if the polymyxin-susceptible isolate did not grow in presence of colistin. We considered the test result interpretable if the following 4 conditions were met: 1) both wells with 0.85% NaCl without bacterial suspension (wells A1 and B1) remained orange (absence of medium contamination); 2) the wells with bacterial suspension and colistin-free (wells A2–A4) turned from orange to yellow, confirming the metabolism of glucose by the isolates; 3) the wells with the colistin-susceptible reference bacterial suspension (negative control) gave negative results (wells A2 and B2); and 4) the wells with the colistin-resistant reference bacterial suspension (positive control) gave positive results (wells A3 and B3). The test result was positive when the well containing colistin (well B3) and the isolate to be tested turned from orange to yellow, giving exactly the same color as the well without colistin (well A3), indicating glucose metabolism and growth in presence of colistin (i.e., colistin resistance) (Figure). The test result was negative when the well containing colistin (well B2) with the isolate to be tested remained orange (unchanged color) (Figure) or was more clear than the wells with 0.85% NaCl but not exactly the same color as the well without colistin (not shown). Results were interpreted by 2 technicians who did not know which isolates were colistin resistant and colistin susceptible.

Other Experimental Conditions Tested

Polymyxin B Instead of Colistin. We evaluated the possibility of adapting the test to susceptibility testing of polymyxin B in countries where polymyxin B is prescribed. To do so, we performed the rapid polymyxin NP test with 20 colistin/polymyxin B–susceptible isolates and 20 colistin/polymyxin B–resistant isolates with polymyxin B at the same concentrations of colistin sulfate.

Incubation Conditions. To determine effects of the incubation atmosphere on time to result, we incubated the tray with 20 colistin-susceptible isolates and 20 colistin-resistant isolates, in parallel, under 2 conditions: ambient air and atmosphere with 5% CO2. We also incubated the tray for 20 colistin-susceptible isolates and 20 colistin-resistant isolates in parallel with and without agitation.

Culture Media. To determine the potential effects of culture medium on the test results, we performed the test with 20 colistin-susceptible isolates and 20 colistin-resistant isolates cultured overnight on different agar plates. The following culture media were tested: 1) nonselective culture medium such as Columbia agar + 5% sheep blood (bioMérieux, La-Balme-Les-Grottes, France); 2) chocolate agar + PolyVitex (bioMérieux); 3) nonselective chromogenic medium UriSelect 4 (Bio-Rad); 4) Eosin methylene blue agar (Sigma Aldrich); 5) Drigalski agar (Bio-Rad); 6) MacConkey agar (VWR BDH Prolabo, Leuven, Belgium); and 7) bromocresol purple (bioMérieux).