Rapid Bacterial Detection and Gram-Identification Using Bacterially Activated, Macrophage-Membrane-Coated Nanowired-Si Surfaces in a Microfluidic Device

Bacterially induced sepsis requires rapid bacterial detection and identification. Hours count for critically ill septic patients, while current culture-based detection requires at least 10 h up to several days. Here, we apply a microfluidic device equipped with a bacterially activated, macrophage-membrane-coating on nanowired-Si adsorbent surfaces for rapid, bacterial detection and Gram-identification in bacterially contaminated blood. Perfusion of suspensions of Gram-negative or Gram-positive bacteria through a microfluidic device equipped with membrane-coated adsorbent surfaces detected low (<10 CFU/mL) bacterial levels. Subsequent, in situ fluorescence-staining yielded Gram-identification for guiding antibiotic selection. In mixed Escherichia coli and Staphylococcus aureus suspensions, Gram-negative and Gram-positive bacteria were detected in the same ratios as those fixed in suspension. Results were validated with a 100% correct score by blinded evaluation (two observers) of 15 human blood samples, spiked with widely different bacterial strains or combinations of strains, demonstrating the potential of the platform for rapid (1.5 h in total) diagnosis of bacterial sepsis.


Bacterial culture conditions and harvesting:
Reagent Co. LTD, China).One colony was inoculated in 10 mL of the appropriate liquid growth medium (Tryptone Soy Broth or Luria-Bertani) and incubated for 24 h at 37°C.This pre-culture was inoculated (1:200) into 200 mL fresh growth medium and incubated for 16 h.Bacteria were harvested by centrifugation (5000 ×g, 5 min) and washed twice with phosphate buffered saline (PBS; NaCl 0.137 M, KCl 0.0027 M, Na 2 HPO 4 0.01 M, KH 2 PO 4 0.0018 M, pH 7.4).Bacterial suspensions were diluted in PBS or growth medium to the concentrations required for the specific experiments.The final concentrations were determined using plate counting.

Fabrication of a microfluidic device:
The microfluidic device was fabricated as previously reported, 1,2 by affixing a polydimethylsiloxane (PDMS; SYLGARDTM 184 Silicon Elastomer Kit, DOW Co., Midland, USA) chip to a patterned nanowired Si wafer (SiNWs).The PDMS chip featured eight parallel straight microfluidic channels (length × width × height: 51 mm × 2 mm × 50 µm) with herring-bone shaped microstructures on the top wall of the channel to generate a turbulent flow to improve contact between the substances in the blood and the SiNW wafer constituting the bottom plate.

Macrophage membrane coating in the microfluidic device:
E. coli-activated, J774A.1 macrophage membranes were prepared as described before 1 by freeze-thawing and used within one day after preparation while being stored at 4°C.
For macrophage membrane coating, the SiNW wafer was first coated in the microfluidic device with (3-aminopropyl) triethoxysilane (APTES, Sigma-Aldrich, Saint Louis, USA) after which bacterially-activated macrophage membrane fragments suspended in PBS were perfused through the devices.

Bacterial capture and staining:
For evaluating bacterial capture by macrophage membrane coated SiNW adsorbent surfaces in the microfluidic device, 1 mL PBS with different bacterial concentrations or with different ratios of combinations of E. coli and S. aureus were perfused through the device at a flow rate of 1 mL/h.After perfusion, the device was filled with 35 μL of a fluorescent Gram-stain (LIVE Baclight TM Bacterial Gram Stain Kit, Thermo Fisher Scientific Inc., Waltham, USA) dissolved in water (8 μM SYTO 9 and 11 μM hexidium iodide) and left for 10 min under static conditions.After staining, the adsorbent surface with captured bacteria was removed from the device, washed with PBS and examined using Confocal Laser Scanning Microscopy (Zeiss CLSM 800, Jena, Germany), operated at a magnification of 400x, yielding a field of view equal to 0.018 mm 2 .Images were analyzed using Image J (National Institutes of Health, Wisconsin, USA) to calculate the number of captured Gram-stain bacteria.

Validation of bacterial absence or presence and Gram-type in bacterially-spiked human blood spiked with bacteria:
For validating the use of bacterially-activated macrophage coated SiNW adsorbent surfaces in a microfluidic device, human whole blood was obtained from healthy donors in the Soochow University Hospital with permission (January 14, 2022) from the Medical Ethics Committee (Soochow University, Suzhou, China) and the written informed consent of the donors.Blood was stored in an anticoagulant tube with ethylenediamine tetra-acetic acid (BD, New Jersey, USA) and spiked with different Gram-negative or Gram-positive bacterial strains or their combinations in concentrations ranging from above the detection limit established up to 200 CFU/mL, representing the range of clinically observed bacterial concentrations in blood of septic patients.Random bacterial concentrations and compositions of spiked blood (15 samples) were computer-generated (ChatGPT, Open AI, USA). 1 mL of each sample was perfused through a microfluidic device, as described above and CLSM images were analysed by two blinded observers, unaware of the composition of the sample, for the absence or presence of bacteria and their Gram-type.

Figure S1 .
Figure S1.Scanning electron micrograph of the architecture of nanowired-Si surfaces.