Reconfigurable microfluidic circuits for isolating and retrieving cells of interest

Abstract

Microfluidic devices are widely used in many fields of biology, but a key limitation is that cells are typically surrounded by solid walls, making it hard to access those that exhibit a specific phenotype for further study. Here, we provide a general and flexible solution to this problem that exploits the remarkable properties of microfluidic circuits with fluid walls - transparent interfaces between culture media and an immiscible fluorocarbon that are easily pierced with pipets. We provide two proofs-of-concept in which specific cell sub-populations are isolated and recovered: i) murine macrophages chemotaxing towards complement component 5a, and ii) bacteria (Pseudomonas aeruginosa) in developing biofilms that migrate towards antibiotics. We build circuits in minutes on standard Petri dishes, add cells, pump in laminar streams so molecular diffusion creates attractant gradients, acquire time-lapse images, and isolate desired sub-populations in real-time by building fluid walls around migrating cells with an accuracy of tens of micrometres using 3D-printed adaptors that convert conventional microscopes into wall-building machines. Our method allows live cells of interest to be easily extracted from microfluidic devices for downstream analyses.

Suicidal chemotaxis in bacteria

“…Unlike the BioFlux assays, where cells are fully surrounded by PDMS and glass walls, these open microfluidic devices allowed us to reconfigure channels mid-experiment to isolate and extract cells as they undergo chemotaxis. These experimental methods are explained in detail in a companion publication (37). Briefly, fluid-walled channels were fabricated using Dulbecco's Modified Eagle Medium (DMEM) + 10% Fetal Bovine Serum (FBS) on untreated, 50 mm glass-bottomed culture dishes (MatTek) using a custom-designed 'printer' (iotaSciences Ltd, Oxford, UK).…”

“…TB growth medium was then infused through the device to replace the DMEM + FBS used for fabrication (36). Antibiotic gradients were generated using 'm-shaped' devices where the flow from two inlet arms merges together in a central channel (37). Using the printer, a 1:1 co-culture of YFP-labelled WT and unlabelled ΔpilG cells was infused into the device at 0.4 µl min -1 and the cells were then left to attach to the bottom glass surface of the channel for 10 min in the absence of flow.…”

“…The needles where connected to 500 µl glass syringes (Hamilton) that were mounted on separate syringe pumps (PhD Ultra, Harvard Apparatus) using 1 m of tubing (PTFE, Adhesive Dispensing Ltd). Ciprofloxacin (100X MIC) was added to one of the syringes to generate a steady gradient in the channel (37). The growth media in both syringes was supplemented with 10% spent medium in order to elicit the chemotactic response more rapidly (see Fig.…”

“…After a sufficient number of cells had migrated towards ciprofloxacin (≈24 h), we re-configured the central channel of the device using a PTFE (polytetrafluoroethylene) stylus (38) mounted to the microscope condenser to isolate different populations of migrating cells within a separate fluid chamber (Fig. 3A-D; see (37) for details). Immediately prior to this re-configuration, we reduced the concentration of ciprofloxacin flowing through one of the inlet arms from 100X MIC to 3X MIC using a fourth syringe (Plastipak, 1 mL, Becton Dickinson) mounted on a separate syringe pump (PhD Ultra, Harvard Apparatus).…”

“…In order to overcome this challenge, we moved to a novel assay that uses open "fluid-walled" microfluidics (36), where twitching cells can be studied in a chemical gradient and then isolated after they have migrated, (Materials and Methods; for a detailed description of our approach, see (37)).…”