Fig 1.
Timeline of current bacteraemia diagnostics.
Diagram showing major stages in current bacteraemia diagnostics featuring three main test streams with a minimum of 16 hours and maximum of 2 or more days for definitive bacterial identification.
Table 1.
NxT AFC parameters.
Table 2.
Gating methods flow chart.
Fig 2.
Bacterial detection by acoustic flow cytometry after direct staining with SYTO 9 in pure bacterial culture and inoculated blood culture.
A: SYTO 9 gating for bacteria grown in TSB as pure culture: the population of interest (POI) was gated based on the FSC-H/SSC-H scatter plot (i) while single bacterial cell events were identified in the FSC-H/FSC-A plot (ii). Events that were SYTO 9 positive were gated in the BL1-H channel (iii) and back-gated to FSC-H/SSC-H scatter plot to identify bacterial cells labelled with SYTO 9, denoted as derived POI (dPOI) (iv) since SYTO 9 binds to other particles containing nucleic acids. B: Bacterial detection in pure culture (i) and blood culture (ii) by SYTO 9 staining. Uninfected BC was used as a negative control (iii). Both pure culture (i) and blood culture (ii) were inoculated with identical concentrations of bacteria at approximately 8.00 x 104 CFU/mL. The pure bacterial culture was stained directly with SYTO 9, while both spiked and uninfected blood cultures were lysed with 10% Triton X-100 before SYTO 9 staining. Note the strong background noise produced during the BC lysis in both infected and uninfected BC samples overshadowing dPOI (ii and iii).
Fig 3.
PNA-FISH assay optimisation flowchart using pure bacterial culture.
Four optimisation protocols were performed one of each for formamide concentration, hybridization temperature, probe concentration, and hybridization duration. In the first two protocols (top rows), SYTO 9 was used to determine changes in the overall bacterial population subjected to variations in formamide concentrations and temperature. Initially, varying concentrations of formamide were added to bacterial suspensions containing equal amounts of bacteria followed by incubation at 40°C for 15 minutes prior to staining with SYTO 9. The highest bacterial recovery of 75% was achieved with 30% formamide (v/v, thick frame) compared to other treated and untreated samples. This formamide concentration (30%) was selected for subsequent experiment in which bacterial cells were exposed to varied hybridization temperatures from 30°C to 60°C (second row). The best result with 52% bacterial cell recovery was obtained at 40°C (thick frame). For the last two optimisation steps, PNA-FISH probe was used in hybridization buffer containing 30% formamide at 40°C (third and fourth rows). In the third row, three different probe concentrations from 100 to 300 nM were tested by PNA-FISH. The probe concentration of 200 nM turned out to be optimal (thick frame) since no significant increase in hybridisation signal was observed when the probe concentration was elevated to 300 nM. Eventually, hybridization duration of 15 minutes was determined as optimal since the majority of the hybridization positive events were detected in the first 15 minutes. The squares with thick frames stacked vertically indicate optimised hybridization conditions applied to all subsequent experiments.
Table 3.
Optimized hybridization condition* for all bacterial species studied.
Fig 4.
Gating of PNA-FISH hybridized BC samples.
Bacteria-specific events were separated from the background noise of lysed blood based on a difference between the PNA probe-specific signal and autofluorescence produced by blood (i). This was followed by identification of single cells on the FSC-H/FSC-A scatter plot within the bacterial population (ii). Probe specific signal was gated as AlexaFluor 488 + on the histogram (iii), and all events within this gate were backgated and shown on the FSC-H/SSC-H plot as dPOI (iv).
Fig 5.
Early bacterial detection in BC with PNA-FISH-AFC.
A: Detection of E. coli by PNA-FISH enhanced AFC in spiked BC following incubation on the BacTEC FX system. A 9-time increase in number of events was observed in the bacteria-specific gate from 5 to 7 hours of incubation indicating actively growing bacterial population. B: Growth curves of bacteria in spiked BC on the BacTEC system. Two fast growing bacterial species, E. coli and K. pneumoniae, and one slow grower, P. aeruginosa, were incubated on the BacTEC FX system. The 2-hourly determined plate counts (y-axis, CFU/mL) were used to produce growth curves, while the appearance of bacteria-specific events by PNA-FISH enhanced AFC was monitored hourly for the first 12 hours of incubation. The overall positivity on the BacTEC system was also monitored in parallel. Bacteria-specific signal by PNA-FISH-AFC was first observed at 5 and 10 hours of incubation for the fast (E. coli, K. pneumoniae) and slower growing bacterial species (P. aeruginosa), respectively. This was confirmed by three independent experiments. The positivity on the BacTEC system was reached at approximately 13 hours and 21 hours of incubation for the fast and slow growers, respectively, also confirmed by three independent experiments with slight variations between experiments. The horizontal broken lines indicate the sensitivity of the PNA-FISH-AFC at 103−104 CFU/mL, and for the BacTEC system at 108−109 CFU/mL. The time gained by the PNA-FISH assay when compared to BacTEC positivity was 8 hours for fast growers, as indicated by blue arrows, and 10 hours for slow grower, as indicated by red arrows, respectively.
Table 4.
Bacterial enumeration using PNA-FISH and plate counts at the time of detection.
Table 5.
Earliest time to detection using different diagnostic assays.