Strains and target antimicrobials
Three strains each of S. pneumoniae and H. influenzae were used: one antimicrobial-susceptible strain and two resistant strains were prepared. Viable cell counting was performed using S. pneumoniae ATCC 49619 and H. influenzae ATCC 49766, which are used as control strains in CLSI testing 19.
To validate the early detection of resistance by FCM, the following antimicrobials were selected: 0.125 μg/mL penicillin G and 1–2 μg/mL cefotaxime (a cephalosporin); moreover, they were tested against S. pneumoniae ATCC 49619, ATCC 6303, and ATCC 700677 19, 46, 47. H. influenzae ATCC 49766, ATCC 33533, and ATCC 49247 were tested with 1–2 μg/mL ampicillin and a combination of 2 μg/mL ampicillin and 1 μg/mL sulbactam, which served as an ampicillin and beta-lactamase inhibitor 48, 49. S. pneumoniae ATCC 49619 and ATCC 6303 were purchased from Kanto Kagaku (Tokyo, Japan). Other strains were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA).
Confirmation of antimicrobial resistance
The antimicrobial resistance and MIC profiles of the strains were determined by the manual method, according to the BMD described in the CLSI criteria 30. Cation-adjusted Mueller-Hinton broth (Becton Dickinson, MD, USA), which had lysed horse blood added to a final concentration of 5%, was used as a medium for S. pneumoniae. The Haemophilus test medium for H. influenzae was used as described in the CLSI criteria. This medium was prepared using cation-adjusted Mueller-Hinton broth, nicotinamide adenine dinucleotide, and yeast extract. The production of H. influenzae beta-lactamase was confirmed by PCR using previously reported primers synthesized by Eurofins Genomics (Tokyo, Japan) 50.
Quantitative viable cell counting
S. pneumoniae strains were precultured in 5% sheep blood agar medium (Eiken Chemical, Tochigi, Japan) for 18 h at 35 °C. H. influenzae was precultured in chocolate agar medium (Eiken Chemical, Tochigi, Japan) in 5% CO2 at 35 °C. The obtained colonies were suspended in sterile saline (0.35% NaCl) and adjusted to MacFarland 1.0 by measuring absorbance (620 nm wavelength). The bacterial suspensions were serially diluted 4-fold with phosphate-buffered saline (PBS; 20 mM phosphate buffer, 130 mM NaCl, pH 7.4) using a glass tube. Diluted bacterial suspensions (5 μL) were further diluted with PBS in a 96-well Nunc-Immuno Module plate microplate (Thermo Fisher Scientific, Waltham, MA, USA). Using this microplate, FCM measurement was performed. For fluorescent staining, the samples were dispensed in a microplate containing PBS with 5 μM SYTO9 and 15 μM PI. Samples were subsequently incubated at room temperature in the dark for 5 min. FCM measurement was performed after 5-fold dilution with PBS to suppress background fluorescence. All fluorescent staining reagents were purchased from Thermo Fisher Scientific (Waltham, MA, USA). FCM measurement was performed using RF-500 (Sysmex, Kobe, Japan) equipped with a blue semiconductor laser (488 nm wavelength). The built-in front scattering (FSC) and side scattering (SSC) detectors were used to detect scattered light. An FL1 filter (527 +/- 15 nm, SYTO9) and FL3 filter (695 +/- 25 nm, PI) were used for fluorescence detection. The measurement volume was set to 20 μL and the flow rate was 1.85 μL/s.
Flowcytometric viable cell counting
For viable cell monitoring by FCM, a 5 μL sample was collected from 100 μL medium containing the bacteria. The sample was taken from the same microplate well of BMD and mixed with PBS containing SYTO9 (5 μM) and PI (15 μM). Next, each sample was incubated at room temperature in the dark for 5 min. FCM measurement was performed after 5-fold dilution with PBS to suppress background fluorescence. The microplate with inoculated wells was immediately returned to the incubator to restart the culture.
The measurement results were confirmed using RF-500 software (Sysmex, Kobe, Japan). Bacterial counting was performed using FCS Express 6 RUO Edition (De Novo Software, Los Angeles, CA, USA).
The detection limits of the viable cell counting were analyzed by Dunnett's multiple comparison test. This test was used to compare the measured value at each time point and the negative control value using StatFlex Ver7.0 (Artec Inc., Osaka, Japan). In these analyses, a p-value less than 0.05 was considered significant. Regression analysis of the viable cell counts was performed with GraphPad Prism 6 software (GraphPad Software Inc., San Diego, CA, USA).