High-throughput production of sub-10 micron droplets for ultrasensitive extracellular vesicle profiling

Introduction: : Extracellular vesicles (EVs) are a heterogeneous population (30-1000nm) of membrane-bound particles containing molecular cargo (DNA, RNA, and proteins) that facilitate intercellular communication, emerging as promising diagnostic biomarkers.1,2 Technologies for ultrasensitive single EV analysis are critical to resolve its heterogeneity and map its composition to disease signatures. Droplet microfluidics-based digital assays have enabled high-throughput analysis of single targets through microscale compartmentalization.3 Droplets are typically formed using T-junction or flow-focusing designs, in which highly monodisperse, micrometer-size droplets are formed through the controlled pinch-off of at least two immiscible fluids. To achieve highly sensitive target detection, studies have explored generating sub-micron droplets, as the local concentration within droplets increases with decreasing volume, scaling with the cube of the droplet radius. However, current methods for sub-micron droplet generation rely on single-channel designs due to challenges in controlling fluid dynamics and fabricating microfluidic channels with sub-micron precision4,5, limiting the throughput necessary for detecting low abundance targets and increasing susceptibility to single-point failures such as clogging.
Here, we present a highly parallelized step emulsification device for generating sub-micron droplets using standard soft lithography techniques. As a proof-of-concept, a device with 20 nozzles was fabricated to generate ~4 µm droplets, achieving high throughput production of 500kHz, with the ultimate goal of incorporating >100 nozzles to generate < 1 µm droplets. To demonstrate the utility of the droplets, we encapsulated CRISPR-Cas12a reagents for rapid and sensitive detection of microRNA (miRNA). We envision this device to enable rapid, ultrasensitive single extracellular vesicle (EV) profiling for rare biomarker discovery and detection.

Materials and

Methods: : Microfluidic device fabrication and droplet formation: The device is designed using AutoCAD and a master mold is fabricated using standard lithography processes. To fabricate the two-layer mold, 10 µm negative tone photoresist (HARE SQ 10) is spin-coated onto a silicon wafer. The continuous phase channels are patterned into the resist coating using a UV mask aligner and developed. Positive tone resist (S1813) is then spin-coated to 1 µm. The dispersion phase nozzles are patterned using UV exposure and developed. The master mold is then silanized with trichloro(1H,1H,2H,2H-perfluorooctyl)silane. To make the devices, polydimethylsiloxane (PDMS; Corning, USA) was prepared at a 10:1 elastomer:curing agent ratio, poured over the master mold, degassed, and cured at room temperature for 48 hours. Fluidic ports were created using 1mm biopsy punches. PDMS devices were then cleaned by sonication in 70% ethanol and bonded to glass under air plasma. Continuous phase channels are rendered fluorophilic using 1% silane and dried under vacuum prior to use. To form droplets, HFE-7500 with 2% (w/w) fluorosurfactant (RAN Biotechnologies, USA) was injected as a continuous phase at 300 µL/hr. The dispersed phase containing the Cas12a master mix is injected with a flow rate of 300 µL/hr.
Cas12a assay: A master mix containing Alt-R A.s. Cas12a, crRNA, and FAM-BHQ are prepared in equimolar ratios in NEBuffer r2.1 and nuclease-free water. Varying concentrations of synthetic miRNA are incubated with the master mix upon dropletization and incubated at 37°C for 15 minutes off-chip. Fluorescent images of the droplets are compared before and after incubation.

Results, Conclusions, and Discussions:: We developed a high-throughput platform for generating monodisperse sub-micron droplets using conventional soft lithography techniques, enabling a broad range of applications such as ultrafast droplet-based assays and sensitive profiling of single extracellular vesicle cargo (Fig. 1A-C). The device utilizes 20 nozzles at this time for parallelization (Fig. 2A,B), with the opportunity to scale up to hundreds of nozzles at a time.6 Profiling of the PDMS device showed proper nozzle formation (Fig. 2C). The device produces monodisperse (CV = 7.01%) droplets (Fig. 3A-C). To demonstrate increased sensitivity as a result of decreased droplet size, a Cas12a fluorescent assay using synthetic miR-16 at a target concentration of 100pM was run in larger (60 µm) and smaller (15 µm) droplets utilizing a standard T-junction droplet device, with the latter showing a 320-fold increase in percentage of positive droplets. Images were blank subtracted, and brightfield images were overlaid with the fluorescent images (Fig. 3D). While the PDMS-based device currently forms 4 µm droplets, we plan to further decrease droplet size and % CV by fabricating a silicon-based step emulsification device, enabling the formation of smaller channels that can withstand higher pressures. Moving forward, we plan to utilize this device to combine the sub-micron droplet maker with the CRISPR assay for high-throughput single EV biomarker screening.

Acknowledgements and/or References (Optional):: 1. Kalluri R. et al. Science 367, eaau6977 (2020).
2. Morales R.T. et. al. ACS Nano 16, 11619 (2022).
3. Park J. et. al. ACS Nano 19, 17884 (2025).
4. Toprakcioglu Z. et al. Sci. Adv. 6, eaay7952 (2020).
5. Ohho R. et al. Sensors and Actuators B: Chemical 421, 136537 (2024).
6. Amstad E. et al. Lab Chip 16, 416 (2016).