Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Feb 6;40(7):3453–3462. doi: 10.1021/acs.langmuir.3c02934

On Chip Sorting of Stem Cell-Derived β Cell Clusters Using Traveling Surface Acoustic Waves

Nikhil Sethia †, Joseph Sushil Rao ‡,§, Zenith Khashim ∥, Anna Marie R Schornack ∥, Michael L Etheridge ⊥, Quinn P Peterson ∥,#, Erik B Finger ‡, John C Bischof ⊥,7, Cari S Dutcher †,⊥,*
PMCID: PMC10883307  PMID: 38318799

Abstract

graphic file with name la3c02934_0006.jpg

There is a critical need for sorting complex materials, such as pancreatic islets of Langerhans, exocrine acinar tissues, and embryoid bodies. These materials are cell clusters, which have highly heterogeneous physical properties (such as size, shape, morphology, and deformability). Selecting such materials on the basis of specific properties can improve clinical outcomes and help advance biomedical research. In this work, we focused on sorting one such complex material, human stem cell-derived β cell clusters (SC-β cell clusters), by size. For this purpose, we developed a microfluidic device in which an image detection system was coupled to an actuation mechanism based on traveling surface acoustic waves (TSAWs). SC-β cell clusters of varying size (∼100–500 μm in diameter) were passed through the sorting device. Inside the device, the size of each cluster was estimated from their bright-field images. After size identification, larger clusters, relative to the cutoff size for separation, were selectively actuated using TSAW pulses. As a result of this selective actuation, smaller and larger clusters exited the device from different outlets. At the current sample dilutions, the experimental sorting efficiency ranged between 78% and 90% for a separation cutoff size of 250 μm, yielding sorting throughputs of up to 0.2 SC-β cell clusters/s using our proof-of-concept design. The biocompatibility of this sorting technique was also established, as no difference in SC-β cell cluster viability due to TSAW pulse usage was found. We conclude the proof-of-concept sorting work by discussing a few ways to optimize sorting of SC-β cell clusters for potentially higher sorting efficiency and throughput. This sorting technique can potentially help in achieving a better distribution of islets for clinical islet transplantation (a potential cure for type 1 diabetes). Additionally, the use of this technique for sorting islets can help in characterizing islet biophysical properties by size and selecting suitable islets for improved islet cryopreservation.

Introduction

Sorting complex materials such as pancreatic islets of Langerhans, exocrine acinar tissues, and embryoid bodies is critical for applications in biomedical research and clinical settings.1−3 These materials are cell clusters and therefore relatively large (>50 μm). These materials broadly vary in size, shape, morphology, and deformability, which makes sorting them challenging. Pancreatic islets are one important example of complex biomaterials for which size-based sorting would play a role in patient outcomes. Pancreatic islets are insulin-secreting cell clusters that broadly range in diameter between 50 and 400 μm in humans.1,4,5 In the past few decades, pancreatic islet transplantation has emerged as a potential cure to type 1 diabetes (T1D),6 though the availability of human islets required for transplantation can be a challenge. Recent developments in generating and characterizing human stem cell-derived β cell clusters (SC-β cell clusters) have shown great potential for SC-β cell clusters’ use as a potential alternative to human islets.7 However, SC-β cell clusters, like human islets, are still significantly heterogeneous in size.

During the early post-transplant period, the transplanted islets’ nutrient and oxygen needs are solely met through a diffusion process.1 Because smaller islets have a relatively shorter diffusion distance, they have a better supply of nutrients and oxygen than larger islets.1 A better supply of oxygen and nutrients can contribute to a higher level of insulin production and a reduced rate of islet cell death, per unit islet volume, observed among smaller islets.1,8 Sorting the islets could potentially characterize islets on the basis of size and aid in transplanting islets of different sizes in their respective ideal transplant sites (which are being investigated9). Furthermore, sorting islets can enhance the distribution of islets, help reduce the graft volume, and improve graft survival, which can eventually improve transplant outcomes. Additionally, there has been a great interest in developing cryopreservation protocols for long-term storage of various biological systems, including islets.10−15 Sorting islets by size can help in characterizing the biophysical properties of islets, which can be dependent on size,16 and selecting the optimum range of islet sizes for cryopreservation. While islet’s size is a crucial parameter influencing the outcome of islet transplantation, other parameters such as the islet’s viability, functionality, and biomechanical property are also highly important to consider. The sorting technique developed in this work can be used alongside other islet characterization techniques to select the most suitable islets for transplantation.17−21

Traditional flow cytometers have been mostly limited to sorting smaller biological systems (<50 μm in size) such as cells. The COPAS FP platform, developed by Union Biometrica, is an exception, as it been previously used for sorting human islets.22 The platform could achieve a high sorting throughput of 10–30 islets/s with sorting efficiencies of >95%. However, the COPAS FP platform can be expensive and bulky, which can limit its everyday research and clinical usage.

The field of microfluidics has advanced greatly over the past few decades largely due to key advantages offered by miniaturized device designs.23 Microfluidic devices are usually more cost-effective to fabricate, easier to transport, and capable of integrating complex operations (such as micromixing) more effectively than their large-scale counterparts. Though extensive research efforts toward cell size-based sorting have been made,24 efforts for sorting islets and other cell clusters by size have been relatively more limited.25−29

Nam et al. developed a multilayered microfluidic device with constriction channels for size-based sorting of mouse pancreatic islets.25 The device can sort small batches of islets (∼100–200 islets) but would fail in clinical settings, where hundreds of thousands of islets need to be sorted, due to blockage of the constriction channel by larger islets. The device also requires the manual retrieval of the sorted islets, which can limit automation of the sorting process. Microfluidic-based devices have also been used in sorting embryoid bodies (EBs). Lillehoj et al. used a microfluidic device with strategically positioned pillars to separate 100–300 μm EBs by size.27 The device’s EB sorting efficiency was limited by clogging events in the device and the heterogeneity in the shape of the EBs. Both of these factors will lead to the islet sorting failing in clinical settings. Additionally, both of the devices described above were passive in nature, where the cutoff size for separation was dependent on the dimensions of the microchannel features. Redesigning and refabricating a new device with a different cutoff size according to the clinical need can be laborious and expensive. To address this concern, Buschke et al. devised an active image detection-based microfluidic device for sorting EBs (100–400 μm) on the basis of size and fluorescence intensity.28,29 The sorting throughput of this device was limited due to the low scan speed (∼5 frames/s), long actuation time (0.1–1 s), and necessary time delay (1–3 s) in the system due to use of an actuation mechanism based on solenoid valves.

Techniques for sorting islets and other cell clusters can be further advanced to achieve higher throughput on the order of 10 cell clusters/s to match throughput of commercial flow cytometers. Sorting techniques that can process large sample volumes (hundreds of thousands of cell clusters), allow one to more easily change the cutoff size for separation (in the range of ∼50–500 μm) and can be automated are highly desirable. In this work, we propose a device based on traveling surface acoustic waves (TSAWs) that can be potentially used to achieve these goals. Acoustic-based actuators can be more durable (due to not having moving parts such as valves) and have short response times, on the order of milliseconds.30−32 They provide additional ways to control the sorting process rather than relying on the microchannel geometry alone for separation. Such devices can be operated continuously, which is crucial for processing large volumes of islet samples, and can perform separation in a contact-free manner, thereby overcoming clogging challenges associated with previous passive sorting techniques. TSAW-based devices are energy efficient, miniaturized, and usually fabricated using polydimethylsiloxane (PDMS), which increase their portability and integration with other microfluidic applications.33

Acoustic-based devices have been previously used for cell sorting.31−42 They have been effective in achieving the desired throughput (>10 cells/s) and sorting efficiency (>85%). Sorting larger biological systems (>50 μm) using acoustic-based devices is still a developing field.43,44 Cell clusters have a much broader distribution of shapes and sizes as well as different structures (such as the presence of intercellular regions) compared to single cells.20,26 Sorting cell clusters can further exhibit the robustness of the acoustic-based sorting platforms. During the process of sorting, cell clusters will be exposed to both an acoustic and a shear field, and therefore, demonstrating the biocompatibility of the sorting technique is also crucial for the end applications in islet transplantation. In this work, a TSAW-based actuator is combined with an image detection-based technique to sort SC-β cell clusters by size. SC-β cell cluster size is estimated in a label-free manner using bright-field imaging of SC-β cell clusters while in flow inside the device. Such a device does not require refabrication to separate with different size cutoffs. This work demonstrates a novel way of sorting SC-β cell clusters that has the potential to sort large sample volumes (hundreds of thousands of SC-β cell clusters) at higher throughputs (∼20 SC-β cell clusters/s) in an automated manner. To the best of our knowledge, this is the first time a TSAW-based sorting technique has been used for separating cell clusters (>100 μm) by size.

Experimental Section

TSAW-Based Actuation

The application of an acoustic field for manipulating a particle-laden fluid has been of great research interest.45−49 An acoustic field can be established using TSAWs, which are generated over a piezoelectric substrate by supplying an alternating current to the substrate-bound interdigital transducers (IDTs). The spacing between periodic IDT fingers determines the wavelength (λ) of the generated TSAWs. The frequency (f) of the TSAWs can be found using the relationship f = cs/λ, where cs (≈3992 m/s) is the speed of sound in the substrate. The TSAWs originate from the IDTs and travel to the microfluidic device bonded to the substrate. The TSAWs then leak into the particle-laden fluid and propagate further through the fluid.

Nonlinear propagation of the TSAWs and attenuation of acoustic energy in the fluid result in the time-averaged body force throughout the fluid volume.50 This body force, in turn, results in time-averaged flow in the fluid, commonly termed as acoustic streaming flow (ASF). Forces arising from TSAW attenuation in the fluid can also displace the suspended particles or biological samples. A suspension (particle/biological system in a liquid) can be translated in the direction of TSAW propagation due to acoustic radiation force (ARF).45,51 ARF depends on the suspension’s size, the energy density of the TSAWs, and a parameter usually termed as the acoustic radiation force factor (ARFF).52 ARFF is a complex function of the suspension size, TSAW frequency, and material properties of the carrier fluid and the suspension. If the suspension has an insignificant value of ARFF, it is not usually translated using ARF. Instead, such suspension can be translated using the drag force induced from ASF.45,47 ARF- and ASF-driven motions of biological samples using TSAW have been previously reported.31,43 In this work, TSAWs were used to drive SC-β cell cluster motion in the direction of TSAW propagation to achieve the final goal of sorting.

Device Fabrication

The sorting device consisted of PDMS-based microchannels over a piezoelectric substrate. The microchannels used in this study were fabricated using soft-lithography protocols. A 500 μm thick epoxy dry film photoresist sheet (SUEX, DJ MicroLaminates, Sudbury, MA) was first bonded over the clean surface of a silicon wafer using hot roll lamination. Using a photoresist sheet helps in circumventing challenges such as edge beads and air entrapment associated with spin coating photoresists. The wafer was then photopatterned with microchannel designs and used as a mold for subsequent device generation. The wafer was exposed to silane vapors [(tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane, Gelest, Morrisville, PA] before PDMS (SYLGARD 184 Silicone Elastomer Kit, Dow, Midland, MI) was poured over the wafer. After overnight incubation and curing at 70 °C, the PDMS, now containing microchannels, was peeled off the wafer. A 1.5 mm biopsy punch (Integra LifeSciences, Princeton, NJ) was used to punch inlet and outlet holes into the microchannels.

A 128° Y-cut X-propagation lithium niobate (LiNbO3) wafer (Precision Micro-Optics, Burlington, MA) was used as the piezoelectric substrate for TSAW generation. The LiNbO3 wafer is well-known for its high electromechanical coupling coefficient and biocompatibility (due to the absence of lead).33,53 The IDTs consisted of two metal layers (Cr/Au: 50 Å/800 Å) and were patterned over the substrate using e-beam evaporation. In total, IDTs had 37 concentric pairs of curved electrodes. The proximal end of the IDTs had ∼180 μm aperture, and it subtended 6° angle at the geometric center. Each electrode was 50 μm wide and 50 μm apart from adjacent electrodes to correspond to the generation of 200 μm wavelength TSAWs. The bonding surfaces of the substrate and the microchannels were activated using oxygen plasma (18 W, 1 min; Harrick Plasma, Ithaca, NY). Microchannels were bonded over the substrate with the help of alignment markers present on each bonding surface. The bonded assembly was incubated at 70 °C for 2 h to enhance bonding between the bonding surfaces. Finally, silver epoxy paste (8330D-19G, MG Chemicals, Burlington, ON) was used to bond wires to IDT pads. The assembled device was exposed to oxygen plasma (18 W, 40 s) before the experiments to make the microchannels hydrophilic.

Experimental Setup

A schematic of the sorting device and the overall sorting process is shown in Figure 1. The SC-β cell cluster sample, consisting of SC-β cell clusters of different sizes, was introduced into the device through inlet I2. The sample was laterally sandwiched between two asymmetrical sheath flows, which entered the device through inlets I1 and I3. The sheath flows were adjusted such that the clusters would exit the device through outlet O1, if not actuated using an acoustic field. The device was mounted over an inverted microscope (IX73, Olympus Corp., Tokyo, Japan) having a 2× magnification objective. A camera (acA2040-120um, Basler, Ahrensburg, Germany) was used to capture images of clusters in flow at a resolution of 2048 pixels × 830 pixels. The images were live-analyzed using a custom-written Python-based code. The code detected clusters from the device background and calculated the radius of the clusters using the projected cluster area. Smaller clusters, relative to the user-chosen cutoff size, traversed the device uninterrupted and exited the device through outlet O1. In response to detection of a larger cluster, the code signaled a connected waveform generator (33519B, Keysight Technologies, Santa Rosa, CA) to output an alternating current of the desired frequency, power, and duration. The current was amplified by a power amplifier (LZY-22+, Mini-Circuits, Brooklyn, NY) and finally delivered to the IDTs. This delivery of current resulted in the generation of TSAWs over the substrate, which then propagated toward the microchannels. The larger clusters, as a result, were pushed in the direction of wave propagation, which led to larger clusters exiting from the device through outlet O2.

Figure 1.

Figure 1

Open in a new tab

Schematic illustration of the microfluidic device used for sorting human stem cell-derived β cell clusters (SC-β cell clusters) by size. The device base consisted of a piezoelectric substrate. The substrate was patterned with interdigital transducers (IDTs), which were used to generate traveling surface acoustic waves (TSAWs). The device top comprised a PDMS-based microchannel used for the flow and size-based separation of the SC-β cell cluster sample. The sample included SC-β cell clusters between 100 and 500 μm diameter, which were introduced into the device through inlet I2. The sample was sandwiched between two sheath flows entering through inlets I1 and I3, which were used to focus the flow of clusters. While in flow inside the device, clusters were imaged, and their sizes were estimated. The size of each cluster was compared against a cutoff size chosen by the device user for the separation. Clusters smaller than the cutoff size continued their lateral flow inside the device uninterrupted and exited the device through outlet O1. Clusters larger than the cutoff size were pushed transversally into outlet O2 using a TSAW pulse. Both cluster detection and cluster actuation were automated and controlled remotely using a custom-written Python-based code. Figure not shown to scale.

Delivery of SC-β Cell Clusters to the Sorting Device

SC-β cell clusters were generated from a human embryonic stem cell line (HUES-8) by growth factor stimulation through six stages of differentiation.11,54 SC-β cell clusters were carefully loaded into a 500 μL syringe, where they were sandwiched vertically between equal volumes of culture media and Percoll solution (Sigma-Aldrich, St. Louis, MO). Percoll is a chemically inert solution that is routinely used in pancreatic islet research.55,56 Because the density of clusters was higher than the density of the culture media and lower than that of the Percoll solution, clusters did not settle inside the syringe. An interface between the Percoll solution and culture media inside the syringe containing clusters was visible. This syringe was held in a horizontal position to align clusters with the syringe’s orifice. The sheath flow solution loaded into 10 mL syringes consisted of equal volumes of culture media and Percoll solution. No clear interface between the Percoll solution and culture media inside the syringe containing the sheath solution was visible. Adding Percoll to the sheath flow solutions increases the viscosity of the sheath flows, which helps in prefocusing of the sample. Flow rates from the syringes were controlled using syringe pumps (Harvard Apparatus, Holliston, MA).

SC-β Cell Cluster Detection and Monitoring

Videos of experiments were recorded and later analyzed manually with the help of Python code. The resolution of the recorded images (512 pixels × 207 pixels) was lower than that of live-analyzed images to allow the code to record longer experiments. The code was used to detect SC-β cell clusters from the device background by successive transformation of the recorded images (as demonstrated in Figure S1). The code first binarized the recorded images. Binarization was followed by detection of edges in the images using the Canny edge detection algorithm57 and morphological operations (dilution and erosion). Relevant contours representing cluster boundaries were identified from the resultant images. Cluster size was estimated by equating the contour area to the projected area of an equivalent spherical cluster. Coordinates of the clusters’ center were identified and were used to track the position of the clusters. This method for cluster size estimation is a simplification and in line with the previous studies that also relied on two-dimensional area measurements to estimate islet size.16,58−60 Further optimization of the cluster size estimation, especially for less spherical clusters, would be needed for end applications requiring more accurate cluster volume measurements. For example, since clusters can rotate while in flow, the size estimation can be improved by averaging the cluster size over successive frames. Additionally, the edge detection of clusters can also be enhanced further by using an adaptive thresholding method to account for the variation in brightness in the device background.

Biological Assessment of SC-β Cell Clusters

The viability of intact SC-β cell clusters was qualitatively assessed using acridine orange/propidium iodide (AO/PI) stain. Intact SC-β cell clusters were stained with 8 ng/mL AO and 20 ng/mL PI (Millipore Sigma, Burlington, MA) for 2 min at room temperature. Stained SC-β cell clusters were coverslipped and imaged using an inverted confocal immunofluorescence microscope (Fluoview 3000, Olympus Corp.) with 502/525 nm filters for AO and 494/636 nm filters for PI. The images of SC-β cell clusters were captured using a 10× magnification objective at a resolution of 4020 pixels × 4020 pixels. Note that the SC-β cell cluster diameters in the captured confocal images are enlarged due to coverslip compression. Such coverslip compression was used to increase the effective depth of imaging. Quantitative measurements of the viability were performed on dissociated SC-β cell clusters. SC-β cell clusters were dissociated into single-cell suspensions in TrypLE Express (12605010, Thermo Fisher Scientific, Waltham, MA), quenched with culture media containing fetal bovine serum, and stained with 8 ng/mL AO and 20 ng/mL PI. After incubation for 15 s, 10 μL of the suspension was pipetted onto Countess Cell Counting Chamber Slides (C10228, Thermo Fisher Scientific) to quantify viability using a Countess II FL cell counter (AMQAF1000, Invitrogen by Thermo Fisher Scientific). The dissociated quantitative viability measurement technique was validated for accuracy. This validation was done by comparing the viability values to those obtained by image analysis of three-dimensional reconstructions of the confocal images of the AO/PI-stained intact SC-β cell clusters.

Results and Discussion

SC-β Cell Cluster Actuation Using TSAW Pulse

TSAW fields have been extensively used to sort cells,31,32,35−41 while their application for sorting large 3D cell clusters such as SC-β cell clusters has been less common.43 Large biological systems, like SC-β cell clusters, have much higher inertia (∝diameter3) compared to that of cells under similar flow conditions. Intercepting the motion of such biological systems for sorting would require much stronger forces, which needs to be tested when using TSAW fields. In the sorting device presented in this work, TSAW pulses are applied normal to the SC-β cell cluster flow direction. The SC-β cell cluster, selectively actuated using TSAW fields, needs to be displaced over long distance (>100 μm) to effectively separate it from rest of the SC-β cell clusters. This distance is much larger (∝diameter) than required for sorting cells. One way to sustain forces over a long distance is by using TSAWs of a relatively longer wavelength (λ). TSAWs attenuate along the fluid/substrate interface in the direction of wave propagation.61 The attenuation coefficient is given by the equation α = ρfcf/ρscsλ, where ρf, ρs, cf, and cs are the densities and speeds of sound in the fluid and substrate, respectively.62 The TSAW amplitude decreases by a factor of 1/e over a characteristic length of α–1 ∝ λ. We used 200 μm wavelength TSAW pulses to push the SC-β cell clusters. This wavelength was approximately 1 order of magnitude higher than the TSAW wavelengths used previously for cell sorting, which was chosen to accommodate our length scales of channels, cell clusters, and required displacements. The use of this wavelength allowed us to prefocus clusters at sufficient distance from the wall prior to actuation and actuate them over long distances. This wavelength corresponded to a TSAW frequency of ∼20 MHz.

A critical step toward sorting is to establish the actuation of SC-β cell clusters using TSAW pulses in our device. For this purpose, SC-β cell clusters were introduced into the device using inlet I2, sandwiched by two sheath flows entering through inlets I1 and I3. The inlet flow rates through I1–I3 were kept constant at 88, 20, and 150 μL/min, respectively. While each cluster was in flow, its motion was tracked, and each cluster’s size was estimated using the code. All of the clusters, irrespective of size, were actuated using a TSAW pulse of a given duration. Clusters traversed linearly in the device before being pushed by the TSAW pulse (Figure 2a,b). TSAW pulses pushed the clusters in the direction of wave propagation, leading to motion of the clusters transverse to the main flow direction (Figure 2b–d).

Figure 2.

Figure 2

Open in a new tab

Actuation of SC-β cell clusters under varying durations of TSAW pulses. (a and b) SC-β cell clusters of a wide range of sizes (∼100–500 μm in diameter) were introduced into the device. Clusters traversed linearly inside the device before being acted upon by the TSAW pulse near the microchannel exit. Arrows on the images represent cluster velocity. (c) Upon application of the TSAW field in the device, clusters were pushed transverse to the main flow. (d) Depending on the duration of the TSAW pulse, clusters exited from outlet O1 or O2. (e) Clusters, irrespective of size, were actuated using a TSAW pulse. Cluster displacement, defined as the center-to-center distance traveled by clusters in the transverse direction, was estimated for each discrete cluster in the flow. Experiments were performed with three different durations of TSAW pulses (18, 36, and 54 ms). Clusters varied in morphology and shape, which could affect their displacement when actuated by using TSAWs. (f) The average displacement of clusters increased with the duration of the TSAW pulse. Average displacement calculated over 22, 46, and 30 clusters that were actuated using TSAW pulses with durations of 18, 36, and 54 ms, respectively. Error bars represent the standard deviation in cluster displacement for each duration of the TSAW pulse. Inlet flow rates for inlets I1–I3 were 88, 20, and 150 μL/min, respectively. Clusters were prefocused prior to actuation by the flow at ∼110 μm from the microchannel center. The TSAW input power was 36.5 dBm. The scale bar is 250 μm.

Cluster displacement, defined as the center-to-center distance traveled by the cluster in the direction transverse to main flow, was estimated for the clusters in flow. Displacements of closely spaced clusters that were inadvertently deflected due to a previous cluster actuation were not measured. Clusters were actuated using three different durations of TSAW pulses (18, 36, and 54 ms) at constant input power (36.5 dBm) in separate experiments. The cluster displacement as a function of their size for each TSAW pulse duration is shown in Figure 2e. Clusters under investigation were within the diameter size range of ∼100–500 μm. The results for displacement as a function of cluster size for the 36 and 54 ms TSAW pulse show noticeable amounts of variability, though no obvious relationship is observed (Spearman’s ρ < 0.1). The scatter in the data could have several sources, both before and after actuation. Variability in cluster displacement can arise due to the differential position of clusters within the microchannel cross section before actuation. Clusters closer to the top microchannel wall can be expected to experience more drag from the wall when actuated, relative to the clusters farther from the top microchannel wall. Similarly, clusters positioned closer to the focal point of IDTs before actuations can be expected to experience more concentrated acoustic energy when actuated, which can further displace clusters. After actuation, clusters can be expected to undergo 3D translations and rotations when exposed to the flow field generated by TSAWs.46,51 With an increase in pulse duration, the differential interaction of the clusters with the flow field was longer, and therefore, more variation in displacements was observed. There was broad variability in cluster morphology and shape, which could also lead to variability in the cluster’s displacement at each TSAW pulse duration. Additionally, some of what were assumed to be small clusters could also be tissue debris, and their morphologies can be different from those of intact SC-β cell clusters. Tissue debris need not be differentiated from intact SC-β cell clusters as they are composed of the same cellular content. To the best of our knowledge, this is the first reported study of cluster actuation in flow using varying durations of TSAW pulses. The exact mechanism of cluster motion (ASF vs ARF) requires further investigation, though the goal of cluster actuation over long distances (>100 μm) using TSAWs was achieved. The average cluster displacement over each duration of the TSAW pulse is summarized in Figure 2f. With an increase in the duration of the TSAW pulse, clusters were on average pushed farther in the transverse direction. Clusters were displaced 132 ± 24, 297 ± 53, and 384 ± 52 μm using 18, 36, and 54 ms TSAW pulses, respectively. A 54 ms TSAW pulse duration was chosen for subsequent sorting experiments to ensure a high degree of separation of clusters.

Sorting of SC-β Cell Clusters by Size

Having established SC-β cell cluster displacement using a TSAW pulse, we then used the device for sorting SC-β cell clusters by size. SC-β cell clusters of various sizes were introduced into the devices through inlet I2. The inlet flow rates for inlets I1–I3 were set to 66, 40, and 150 μL/min, respectively (for a total flow rate similar to that of Figure 2). At these flow rates, clusters were prefocused prior to actuation at a sufficient distance (∼160 μm) from the microchannel center to allow for efficient separation of small and large clusters. A cutoff size of 250 μm was chosen for separation to demonstrate the proof of concept. TSAW pulses were used to selectively actuate clusters on the basis of optical detection of cluster size. Clusters larger than 250 μm were pushed into outlet O2 using the 54 ms TSAW pulse. No actuation signal in response to clusters detected to be smaller than 250 μm was generated. Snippets of the experiment, where smaller and larger clusters flowing consecutively were separated, are shown in Figure 3 (also see Movie S1). The smaller cluster measured ∼220 μm in diameter and continued to flow laterally to exit through outlet O1 (Figure 3a–c). The larger cluster measured ∼300 μm in diameter and was actuated using a 54 ms TSAW (Figure 3d,e). The larger cluster finally exited through outlet O2 (Figure 3f).

Figure 3.

Figure 3

Open in a new tab

Sorting SC-β cell clusters by size. (a) The SC-β cell cluster (∼220 μm in diameter), highlighted with a dashed red circle, traversed inside the sorting device. The cluster was smaller than the cutoff size (250 μm) for sorting. (b) The smaller cluster traversed in a straight line toward the channel exit. No TSAW pulse was generated to actuate the smaller cluster. Another SC-β cell cluster (∼300 μm in diameter), highlighted with a dashed green circle, also entered the device. This newly entered cluster was larger than the size cutoff for sorting. (c) The smaller cluster exited the microchannel through outlet O1. (d) A 54 ms duration TSAW pulse selectively acted upon the larger cluster. (e) The larger cluster was pushed transverse to the main flow direction. (f) The larger cluster exited the microchannel through outlet O2. Arrows on the images represent the velocity of each cluster. Inlet flow rates for inlets I1–I3 were 66, 40, and 150 μL/min, respectively. The TSAW input power was 36.5 dBm. The scale bar is 250 μm.

Presence of Clumps of SC-β Cell Clusters during SC-β Cell Cluster Size-Based Sorting

Experiments sorting SC-β cell clusters were performed under the same conditions that were used in Figure 3. The initial SC-β cell cluster sample consisted of a significant number of clumps of SC-β cell clusters alongside discrete SC-β cell clusters. Clumping of clusters can be understood as an effective increase in the cluster size, as clumping limits the diffusion of oxygen in the clusters.4 These clumps of clusters were detected and counted as single clusters for the purpose of estimating the sorting performance. As shown in Figure 4a (also see Movie S2), a small discrete cluster (∼240 μm in diameter) and a large clump of clusters (≫250 μm in diameter) were introduced into the device. No TSAW pulse was generated to actuate the smaller cluster (Figure 4b), and the cluster exited through outlet O1, meant for smaller cluster collection. On the contrary, the clump of clusters was actuated using a 54 ms TSAW pulse (Figure 4c). As shown in Figure 4d, the clump of clusters exited through outlet O2, which was meant for larger cluster collection. The sorting efficiency was defined as the total percentage of clusters successfully sorted into their desired outlets (smaller clusters into outlet O1 and larger clusters into outlet O2). Over the 12 sorting experiments performed (with 20–105 clusters each), the sorting efficiency ranged between 78% and 90% (as summarized in Table S1) with sorting throughputs of up to 0.2 SC-β cell clusters/s achieved. Cumulatively over about 700 sorted clusters, a sorting efficiency of 87% was achieved (as summarized in Table S2). Note that this efficiency was achieved with a relatively unoptimized geometry; the majority of the failed sorting events could be successfully performed with wider and taller channels. Most of the missorted events were related to the sorting of clumps of clusters. The missorted clumps of clusters were on average >500 μm in diameter, which is larger than the height of the current microchannel. With successful sorting of these larger SC-β cell clusters shown, we now consider the biocompatibility of the method for biological systems, as they are subject to both shear and acoustic fields.

Figure 4.

Figure 4

Open in a new tab

Sorting of a clump of SC-β cell clusters from a discrete SC-β cell cluster. (a) The SC-β cell cluster sample had discrete SC-β cell clusters (highlighted by a dashed red circle) as well as clumps of SC-β cell clusters (highlighted by a dashed green circle). (b) The discrete cluster (∼240 μm) was smaller than the cutoff size of 250 μm and therefore not actuated. (c) The clump of clusters was identified as a single cluster that was >250 μm in size and therefore pushed transverse to the main flow direction. The smaller discrete cluster exited the microchannel through outlet O1. (d) The clump of clusters exited the microchannel through outlet O2. The scale bar is 250 μm.

Biocompatibility of TSAW Pulses for SC-β Cell Cluster Sorting

While SC-β cell clusters have been actuated and sorted in this work, it was also critical to demonstrate that the flow and the streaming wave would not result in mechanical damage to or a change in the viability of the SC-β cell clusters. To test the biocompatibility of the method, SC-β cell clusters were introduced into the device at the same inlet flow rates that were used in Figure 3. The cutoff size for sorting was set at 50 μm to ensure actuation of each cluster in the flow. Approximately 50 clusters were actuated using 54 ms TSAW pulses, which is the longest duration of TSAW pulses used in this work (as used in Figure 3). After the recovery of clusters from the device outlets, the clusters were transported in SC-β cell cluster culture media for viability assessments. Qualitatively, TSAW-actuated clusters (Figure 5d–f) appeared to be similar to controls (Figure 5a–c) with no obvious increase in the number of necrotic/dead cells (PI-stained red cells). The imaging results correlated to a quantified viability of 89% for both control and actuated clusters, measured through dissociation and single-cell viability assessment.

Figure 5.

Figure 5

Open in a new tab

Biocompatibility of the TSAW actuation. Qualitative viability assessments of (a–c) control and (d–f) TSAW-actuated SC-β cell clusters, using acridine orange (cyan for all cells) and propidium iodide (red for dead/necrotic cells). All SC-β cell clusters were captured at 10× magnification using an Olympus Fluoview 3000 single-photon confocal immunofluorescence microscope. Clusters imaged using confocal microscopy are qualitative representations of their respective groups (control/TSAW-actuated). Additional images can be found in Figure S2. The scale bar is 150 μm.

Scope for Further Optimization of SC-β Cell Cluster Sorting

While the proof of concept for the novel acoustic-based sorting of SC-β cell clusters has been established, the sorting technique can be further optimized for higher sorting efficiency and throughput (summarized in Table 1). For improved culturing of clusters and limiting clumping of clusters, cluster samples are routinely cultured in spinning flasks in research settings. Cluster samples can be directly delivered into the sorting device from such culturing flasks with the help of a pressure-based flow. The device then can be used for continuous separation of large sample volumes of clusters (hundreds of thousands of SC-β cell clusters). More uniform spacing between the clusters in flow can possibly be achieved when the sample is well mixed. The initial cluster sample can also be diluted with the culture media, which can increase the spacing between successive clusters in the flow. Sufficient spacing between clusters would reduce the concurrent occurrence of clusters under the TSAW field. Sample dilution would also aid in reducing any inadvertent motion of incoming clusters due to the flow field from previous actuation. Additionally, an increased separation of clusters in the flow is possible through increased flow rates of the sheath. Though the device can be used for sorting large clumps of clusters, their flow into the device can be enhanced by using wider and taller channels. Such channels can reduce the level of a clump’s interaction with the microchannel walls and enable the fluid drag, arising from inlet sheath flows, to focus them on the desired streamlines. This focusing of a large clump of clusters was a challenge in this design. In addition, the desired sorting efficiency can be achieved by passing the sorted sample multiple times through the device. Though we did not observe differences in sorting efficiency among experiments performed over time, it might also be of potential interest to investigate the variation of the TSAW resonance frequency over time.63,64

Table 1. Future Optimization of the Sorting Technique to Achieve a Higher Sorting Efficiency and Throughput.

Higher Sorting Efficiency
intended measures expected outcomes
delivery of clusters from spinning flasks improved spacing between clusters
dilution of the initial cluster sample increased spacing between clusters
use of higher sheath flow rates increased spacing between clusters
use of wider and taller channels improved streamline flow of clusters
sample passed through the device multiple times increased sorting efficiency per pass
Higher Sorting Throughput
intended measures expected outcomes
use of higher sample flow rates more clusters processed by the device
use of a higher-power TSAW pulse reduced actuation time of clusters
concentration of the initial cluster sample more frequent cluster detection and actuation
sorting across multiple devices increased total throughput
Open in a new tab

While a downstream cluster is being actuated, an upstream cluster can be recognized. If the time spent per image (the recognition time) is shorter than the actuation time (54 ms), the throughput is not limited by the detection. The average time spent per image per sorting experiment ranged from ∼36 to ∼50 ms, meaning that the recognition process tends not to be the process-limiting step in our work. If needed, the code can be further optimized for a shorter recognition time (more details are available in the Supporting Information). Under the current operating parameters of the device (as used in Figure 3), a theoretical maximum of ∼20 SC-β cell clusters/s can be sorted on the basis of the 54 ms TSAW pulse duration. At that rate, ∼300 000 islets, the number required in a typical transplant,1 could be sorted within 5 h, which is an acceptable processing time within the current clinical islet manufacturing workflow.65 Actual rates would be somewhat reduced in the current device as it requires entry and exit of a new cluster under the acoustic field every 54 ms. However, a higher throughput could be attained by increasing the TSAW power, which would lead to a reduced duration of the TSAW pulse, and by using higher sample flow rates. Higher flow rates can themselves require the use of higher TSAW pulse powers, especially if the cluster actuation is driven by ASF.47 Concentrating the initial cluster sample with more clusters can increase the throughput by reducing the time between successive sorting events. However, concentrating the sample can also lead to a decreased sorting efficiency, therefore indicating the need for further optimization of the initial sample concentration. In addition, more than one of these miniaturized devices can be run in parallel to reduce the total sorting time.

Conclusion

A unique way to sort human stem cell-derived β cell clusters by size using traveling surface acoustic waves has been presented in this work. The sorting device was fabricated by bonding a PDMS-based microchannel over a piezoelectric substrate (LiNbO3). Because both the microchannel and the substrate were transparent, label-free imaging of clusters inside the device was possible. The size of individual clusters was therefore estimated from bright-field images of the clusters while in flow inside the device. First, the ability to push clusters over hundreds of micrometers transverse to the main flow direction using TSAW pulses was established. Then selective actuation of clusters larger than a user-chosen cutoff size was demonstrated. TSAW-based actuation of clusters was found to be biocompatible, as it maintained the viability of the clusters. We conclude the work by discussing strategies for further optimizing the sorting of the clusters to achieve a higher sorting efficiency and throughput. Overall, this work is a unique addition to the methods for label-free and contact-free sorting of cell clusters. This technique can be potentially used to sort large sample volumes of pancreatic islets at high throughput for applications in islet transplantation and islet cryopreservation.

Acknowledgments

Portions of this work were conducted in the Minnesota Nano Center, which is supported by the National Science Foundation through the National Nanotechnology Coordinated Infrastructure (NNCI) under Award Number ECCS-2025124. Figure 1 and the graphical abstract were created with BioRender.com. This work is supported by grants from the National Science Foundation (EEC 1941543, J.C.B. and E.B.F.) and the National Institutes of Health (R01DK131209, E.B.F. and J.C.B.). Investigators are supported by the Eunice L. Dwan Endowed Diabetes Research Chair (E.B.F.). Q.P.P. acknowledges the generosity of the J. W. Kieckhefer Foundation, the Stephen and Barbara Slaggie Family, and the Khalifa Bin Zayed Al Nahyan Foundation for supporting this work.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.langmuir.3c02934.

  • Overview of the bright-field image transformation process, representative confocal images of control and TSAW-actuated clusters, and summary of SC-β cell cluster sorting experiments (PDF)

  • Sorting discrete SC-β cell clusters by size (AVI)

  • Sorting of clump of SC-β cell clusters from discrete SC-β cell cluster (AVI)

The authors declare no competing financial interest.

Special Issue

Published as part of Langmuirvirtual special issue "2023 Pioneers in Applied and Fundamental Interfacial Chemistry: Janet A. W. Elliott".

Supplementary Material

la3c02934_si_001.pdf (1.1MB, pdf)
la3c02934_si_002.avi (224.8KB, avi)
la3c02934_si_003.avi (340KB, avi)

References

  1. Lehmann R.; Zuellig R. A.; Kugelmeier P.; Baenninger P. B.; Moritz W.; Perren A.; Clavien P. A.; Weber M.; Spinas G. A. Superiority of Small Islets in Human Islet Transplantation. Diabetes 2007, 56 (3), 594–603. 10.2337/db06-0779. [DOI] [PubMed] [Google Scholar]
  2. Ichii H.; Pileggi A.; Molano R. D.; Baidal D. A.; Khan A.; Kuroda Y.; Inverardi L.; Goss J. A.; Alejandro R.; Ricordi C. Rescue Purification Maximizes the Use of Human Islet Preparations for Transplantation. Am. J. Transplant. 2005, 5 (1), 21–30. 10.1111/j.1600-6143.2005.00698.x. [DOI] [PubMed] [Google Scholar]
  3. Valamehr B.; Jonas S. J.; Polleux J.; Qiao R.; Guo S.; Gschweng E. H.; Stiles B.; Kam K.; Luo T. J. M.; Witte O. N.; Liu X.; Dunn B.; Wu H. Hydrophobic Surfaces for Enhanced Differentiation of Embryonic Stem Cell-Derived Embryoid Bodies. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (38), 14459–14464. 10.1073/pnas.0807235105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Suszynski T. M.; Avgoustiniatos E. S.; Papas K. K. Intraportal Islet Oxygenation. J. Diabetes Sci. Technol. 2014, 8 (3), 575–580. 10.1177/1932296814525827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Korsgren O.; Nilsson B.; Berne C.; Felldin M.; Foss A.; Kallen R.; Lundgren T.; Salmela K.; Tibell A.; Tufveson G. Current Status of Clinical Islet Transplantation. Transplantation 2005, 79 (10), 1289–1293. 10.1097/01.TP.0000157273.60147.7C. [DOI] [PubMed] [Google Scholar]
  6. Shapiro A. M. J.; Pokrywczynska M.; Ricordi C. Clinical Pancreatic Islet Transplantation. Nat. Rev. Endocrinol. 2017, 13 (5), 268–277. 10.1038/nrendo.2016.178. [DOI] [PubMed] [Google Scholar]
  7. Hogrebe N. J.; Ishahak M.; Millman J. R. Developments in Stem Cell-Derived Islet Replacement Therapy for Treating Type 1 Diabetes. Cell Stem Cell 2023, 30 (5), 530–548. 10.1016/j.stem.2023.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Fujita Y.; Takita M.; Shimoda M.; Itoh T.; Sugimoto K.; Noguchi H.; Naziruddin B.; Levy M. F.; Matsumoto S. Large Human Islets Secrete Less Insulin per Islet Equivalent than Smaller Islets in Vitro. Islets 2011, 3 (1), 1–5. 10.4161/isl.3.1.14131. [DOI] [PubMed] [Google Scholar]
  9. Rajab A. Islet Transplantation: Alternative Sites. Curr. Diabetes Rep. 2010, 10 (5), 332–337. 10.1007/s11892-010-0130-6. [DOI] [PubMed] [Google Scholar]
  10. Kojayan G. G.; Alexander M.; Imagawa D. K.; Lakey J. R. T. Systematic Review of Islet Cryopreservation. Islets 2018, 10 (1), 40–49. 10.1080/19382014.2017.1405202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Zhan L.; Rao J. S.; Sethia N.; Slama M. Q.; Han Z.; Tobolt D.; Etheridge M.; Peterson Q. P.; Dutcher C. S.; Bischof J. C.; Finger E. B. Pancreatic Islet Cryopreservation by Vitrification Achieves High Viability, Function, Recovery and Clinical Scalability for Transplantation. Nat. Med. 2022, 28 (4), 798–808. 10.1038/s41591-022-01718-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Mohammed L.; Marquez-Curtis L. A.; Elliott J. A. W. Cryopreservation of Human Cerebral Microvascular Endothelial Cells with Glycerol. Cryobiology 2023, 113, 104551. 10.1016/j.cryobiol.2023.104551. [DOI] [PubMed] [Google Scholar]
  13. Marquez-Curtis L. A.; Dai X. Q.; Hang Y.; Lam J. Y.; Lyon J.; Manning Fox J. E.; McGann L. E.; MacDonald P. E.; Kim S. K.; Elliott J. A. W. Cryopreservation and Post-Thaw Characterization of Dissociated Human Islet Cells. PLoS One 2022, 17 (1), e0263005. 10.1371/journal.pone.0263005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Marquez-Curtis L. A.; Dorobantu L. S.; Sauvageau D.; Elliott J. A. W. Cryopreservation of Swine Colostrum-Derived Cells. Cryobiology 2020, 97, 168–178. 10.1016/j.cryobiol.2020.05.010. [DOI] [PubMed] [Google Scholar]
  15. Eskandari N.; Marquez-Curtis L. A.; McGann L. E.; Elliott J. A. W. Cryopreservation of Human Umbilical Vein and Porcine Corneal Endothelial Cell Monolayers. Cryobiology 2018, 85, 63–72. 10.1016/j.cryobiol.2018.10.001. [DOI] [PubMed] [Google Scholar]
  16. Woods E. J.; Liu J.; Zieger M. A. J.; Lakey J. R. T.; Critser J. K. Water and Cryoprotectant Permeability Characteristics of Isolated Human and Canine Pancreatic Islets. Cell Transplant. 1999, 8 (5), 549–559. 10.1177/096368979900800510. [DOI] [PubMed] [Google Scholar]
  17. Adewola A. F.; Lee D.; Harvat T.; Mohammed J.; Eddington D. T.; Oberholzer J.; Wang Y. Microfluidic Perifusion and Imaging Device for Multi-Parametric Islet Function Assessment. Biomed. Microdevices 2010, 12 (3), 409–417. 10.1007/s10544-010-9398-1. [DOI] [PubMed] [Google Scholar]
  18. Nourmohammadzadeh M.; Xing Y.; Lee J. W.; Bochenek M. A.; Mendoza-Elias J. E.; McGarrigle J. J.; Marchese E.; Chun-Chieh Y.; Eddington D. T.; Oberholzer J.; Wang Y. A Microfluidic Array for Real-Time Live-Cell Imaging of Human and Rodent Pancreatic Islets. Lab Chip 2016, 16 (8), 1466–1472. 10.1039/C5LC01173F. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Mamidi A.; Prawiro C.; Seymour P. A.; de Lichtenberg K. H.; Jackson A.; Serup P.; Semb H. Mechanosignalling via Integrins Directs Fate Decisions of Pancreatic Progenitors. Nature 2018, 564 (7734), 114–118. 10.1038/s41586-018-0762-2. [DOI] [PubMed] [Google Scholar]
  20. Torres-Castro K.; Azimi M. S.; Varhue W. B.; Honrado C.; Peirce S. M.; Swami N. S. Biophysical Quantification of Reorganization Dynamics of Human Pancreatic Islets during Co-Culture with Adipose-Derived Stem Cells. Analyst 2022, 147 (12), 2731–2738. 10.1039/D2AN00222A. [DOI] [PubMed] [Google Scholar]
  21. Kragl M.; Schubert R.; Karsjens H.; Otter S.; Bartosinska B.; Jeruschke K.; Weiss J.; Chen C.; Alsteens D.; Kuss O.; Speier S.; Eberhard D.; Muller D. J.; Lammert E. The Biomechanical Properties of an Epithelial Tissue Determine the Location of Its Vasculature. Nat. Commun. 2016, 7 (1), 13560. 10.1038/ncomms13560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Steffen A.; Ludwig B.; Krautz C.; Bornstein S.; Solimena M. Functional Assessment of Automatically Sorted Pancreatic Islets Using Large Particle Flow Cytometry. Islets 2011, 3 (5), 267–270. 10.4161/isl.3.5.15939. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Aubry G.; Lee H. J.; Lu H. Advances in Microfluidics: Technical Innovations and Applications in Diagnostics and Therapeutics. Anal. Chem. 2023, 95 (1), 444–467. 10.1021/acs.analchem.2c04562. [DOI] [PubMed] [Google Scholar]
  24. Shen Y.; Yalikun Y.; Tanaka Y. Recent Advances in Microfluidic Cell Sorting Systems. Sens. Actuators B Chem. 2019, 282, 268–281. 10.1016/j.snb.2018.11.025. [DOI] [Google Scholar]
  25. Nam K. H.; Yong W.; Harvat T.; Adewola A.; Wang S.; Oberholzer J.; Eddington D. T. Size-Based Separation and Collection of Mouse Pancreatic Islets for Functional Analysis. Biomed. Microdevices 2010, 12 (5), 865–874. 10.1007/s10544-010-9441-2. [DOI] [PubMed] [Google Scholar]
  26. Varhue W. B.; Langman L.; Kelly-Goss M.; Lataillade M.; Brayman K. L.; Peirce-Cottler S.; Swami N. S. Deformability-Based Microfluidic Separation of Pancreatic Islets from Exocrine Acinar Tissue for Transplant Applications. Lab Chip 2017, 17 (21), 3682–3691. 10.1039/C7LC00890B. [DOI] [PubMed] [Google Scholar]
  27. Lillehoj P. B.; Tsutsui H.; Valamehr B.; Wu H.; Ho C. M. Continuous Sorting of Heterogeneous-Sized Embryoid Bodies. Lab Chip 2010, 10 (13), 1678–1682. 10.1039/c000163e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Buschke D. G.; Resto P.; Schumacher N.; Cox B.; Tallavajhula A.; Vivekanandan A.; Eliceiri K. W.; Williams J. C.; Ogle B. M. Microfluidic Sorting of Microtissues. Biomicrofluidics 2012, 6 (1), 014116. 10.1063/1.3692765. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Buschke D. G.; Vivekanandan A.; Squirrell J. M.; Rueden C. T.; Eliceiri K. W.; Ogle B. M. Large Particle Multiphoton Flow Cytometry to Purify Intact Embryoid Bodies Exhibiting Enhanced Potential for Cardiomyocyte Differentiation. Integr. Biol. 2013, 5 (7), 993–1003. 10.1039/c3ib20286k. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Collins D. J.; Neild A.; Ai Y. Highly Focused High-Frequency Travelling Surface Acoustic Waves (SAW) for Rapid Single-Particle Sorting. Lab Chip 2016, 16 (3), 471–479. 10.1039/C5LC01335F. [DOI] [PubMed] [Google Scholar]
  31. Ma Z.; Zhou Y.; Collins D. J.; Ai Y. Fluorescence Activated Cell Sorting via a Focused Traveling Surface Acoustic Beam. Lab Chip 2017, 17 (18), 3176–3185. 10.1039/C7LC00678K. [DOI] [PubMed] [Google Scholar]
  32. Nawaz A. A.; Soteriou D.; Xu C. K.; Goswami R.; Herbig M.; Guck J.; Girardo S. Image-Based Cell Sorting Using Focused Travelling Surface Acoustic Waves. Lab Chip 2023, 23 (2), 372–387. 10.1039/D2LC00636G. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ding X.; Li P.; Lin S. C. S.; Stratton Z. S.; Nama N.; Guo F.; Slotcavage D.; Mao X.; Shi J.; Costanzo F.; Huang T. J. Surface Acoustic Wave Microfluidics. Lab Chip 2013, 13 (18), 3626–3649. 10.1039/c3lc50361e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nawaz A. A.; Chen Y.; Nama N.; Nissly R. H.; Ren L.; Ozcelik A.; Wang L.; McCoy J. P.; Levine S. J.; Huang T. J. Acoustofluidic Fluorescence Activated Cell Sorter. Anal. Chem. 2015, 87 (24), 12051–12058. 10.1021/acs.analchem.5b02398. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Zhou Y.; Ma Z.; Ai Y. Hybrid Microfluidic Sorting of Rare Cells Based on High Throughput Inertial Focusing and High Accuracy Acoustic Manipulation. RSC Adv. 2019, 9 (53), 31186–31195. 10.1039/C9RA01792E. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Wang K.; Zhou W.; Lin Z.; Cai F.; Li F.; Wu J.; Meng L.; Niu L.; Zheng H. Sorting of Tumour Cells in a Microfluidic Device by Multi-Stage Surface Acoustic Waves. Sens. Actuators B Chem. 2018, 258, 1174–1183. 10.1016/j.snb.2017.12.013. [DOI] [Google Scholar]
  37. Li P.; Liang M.; Lu X.; Chow J. J. M.; Ramachandra C. J. A.; Ai Y. Sheathless Acoustic Fluorescence Activated Cell Sorting (AFACS) with High Cell Viability. Anal. Chem. 2019, 91 (24), 15425–15435. 10.1021/acs.analchem.9b03021. [DOI] [PubMed] [Google Scholar]
  38. Mutafopulos K.; Spink P.; Lofstrom C. D.; Lu P. J.; Lu H.; Sharpe J. C.; Franke T.; Weitz D. A. Traveling Surface Acoustic Wave (TSAW) Microfluidic Fluorescence Activated Cell Sorter (μFACS). Lab Chip 2019, 19 (14), 2435–2443. 10.1039/C9LC00163H. [DOI] [PubMed] [Google Scholar]
  39. Ai Y.; Li P. Label-Free Multivariate Biophysical Phenotyping-Activated Acoustic Sorting at the Single-Cell Level. Anal. Chem. 2021, 93 (8), 4108–4117. 10.1021/acs.analchem.0c05352. [DOI] [PubMed] [Google Scholar]
  40. Li P.; Ma Z.; Zhou Y.; Collins D. J.; Wang Z.; Ai Y. Detachable Acoustophoretic System for Fluorescence-Activated Sorting at the Single-Droplet Level. Anal. Chem. 2019, 91 (15), 9970–9977. 10.1021/acs.analchem.9b01708. [DOI] [PubMed] [Google Scholar]
  41. Ung W. L.; Mutafopulos K.; Spink P.; Rambach R. W.; Franke T.; Weitz D. A. Enhanced Surface Acoustic Wave Cell Sorting by 3D Microfluidic-Chip Design. Lab Chip 2017, 17 (23), 4059–4069. 10.1039/C7LC00715A. [DOI] [PubMed] [Google Scholar]
  42. Ren L.; Chen Y.; Li P.; Mao Z.; Huang P. H.; Rufo J.; Guo F.; Wang L.; McCoy J. P.; Levine S. J.; Huang T. J. A High-Throughput Acoustic Cell Sorter. Lab Chip 2015, 15 (19), 3870–3879. 10.1039/C5LC00706B. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Afzal M.; Park J.; Destgeer G.; Ahmed H.; Iqrar S. A.; Kim S.; Kang S.; Alazzam A.; Yoon T. S.; Sung H. J. Acoustomicrofluidic Separation of Tardigrades from Raw Cultures for Sample Preparation. Zoological Journal of the Linnean Society 2020, 188 (3), 809–819. 10.1093/zoolinnean/zlz079. [DOI] [Google Scholar]
  44. Zhang J.; Hartman J. H.; Chen C.; Yang S.; Li Q.; Tian Z.; Huang P. H.; Wang L.; Meyer J. N.; Huang T. J. Fluorescence-Based Sorting of Caenorhabditis Elegans via Acoustofluidics. Lab Chip 2020, 20 (10), 1729–1739. 10.1039/D0LC00051E. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Skowronek V.; Rambach R. W.; Schmid L.; Haase K.; Franke T. Particle Deflection in a Poly(Dimethylsiloxane) Microchannel Using a Propagating Surface Acoustic Wave: Size and Frequency Dependence. Anal. Chem. 2013, 85 (20), 9955–9959. 10.1021/ac402607p. [DOI] [PubMed] [Google Scholar]
  46. Collins D. J.; Ma Z.; Ai Y. Highly Localized Acoustic Streaming and Size-Selective Submicrometer Particle Concentration Using High Frequency Microscale Focused Acoustic Fields. Anal. Chem. 2016, 88 (10), 5513–5522. 10.1021/acs.analchem.6b01069. [DOI] [PubMed] [Google Scholar]
  47. Collins D. J.; Ma Z.; Han J.; Ai Y. Continuous Micro-Vortex-Based Nanoparticle Manipulation via Focused Surface Acoustic Waves. Lab Chip 2017, 17 (1), 91–103. 10.1039/C6LC01142J. [DOI] [PubMed] [Google Scholar]
  48. Collins D. J.; Khoo B. L.; Ma Z.; Winkler A.; Weser R.; Schmidt H.; Han J.; Ai Y. Selective Particle and Cell Capture in a Continuous Flow Using Micro-Vortex Acoustic Streaming. Lab Chip 2017, 17 (10), 1769–1777. 10.1039/C7LC00215G. [DOI] [PubMed] [Google Scholar]
  49. Settnes M.; Bruus H. Forces Acting on a Small Particle in an Acoustical Field in a Viscous Fluid. Phys. Rev. E 2012, 85 (1), 016327. 10.1103/PhysRevE.85.016327. [DOI] [PubMed] [Google Scholar]
  50. Lighthill J. Acoustic Streaming. J. Sound Vib. 1978, 61 (3), 391–418. 10.1016/0022-460X(78)90388-7. [DOI] [Google Scholar]
  51. Destgeer G.; Ha B. H.; Jung J. H.; Sung H. J. Submicron Separation of Microspheres via Travelling Surface Acoustic Waves. Lab Chip 2014, 14 (24), 4665–4672. 10.1039/C4LC00868E. [DOI] [PubMed] [Google Scholar]
  52. Hasegawa T.; Yosioka K. Acoustic-radiation Force on a Solid Elastic Sphere. J. Acoust. Soc. Am. 1969, 46 (5B), 1139–1143. 10.1121/1.1911832. [DOI] [Google Scholar]
  53. Ciplys D.; Rimeika R. Measurements of Electromechanical Coupling Coefficient for Surface Acoustic Waves in Proton-Exchanged Lithium Niobate. Ultrasound 1999, 33 (3), 14–20. [Google Scholar]
  54. Pagliuca F. W.; Millman J. R.; Gürtler M.; Segel M.; Van Dervort A.; Ryu J. H.; Peterson Q. P.; Greiner D.; Melton D. A. Generation of Functional Human Pancreatic β Cells in Vitro. Cell 2014, 159 (2), 428–439. 10.1016/j.cell.2014.09.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Saliba Y.; Bakhos J. J.; Itani T.; Farès N. An Optimized Protocol for Purification of Functional Islets of Langerhans. Lab. Invest. 2017, 97 (1), 70–83. 10.1038/labinvest.2016.123. [DOI] [PubMed] [Google Scholar]
  56. Buitrago A.; Gylfe E.; Henriksson C.; Pertoft H. Rapid Isolation of Pancreatic Islets from Collagenase Digested Pancreas by Sedimentation through PercollTM at Unit Gravity. Biochem. Biophys. Res. Commun. 1977, 79 (3), 823–828. 10.1016/0006-291X(77)91185-8. [DOI] [PubMed] [Google Scholar]
  57. Canny J. A Computational Approach to Edge Detection. IEEE Trans. Pattern Anal. Mach. Intell. 1986, PAMI-8 (6), 679–698. 10.1109/TPAMI.1986.4767851. [DOI] [PubMed] [Google Scholar]
  58. Buchwald P.; Bernal A.; Echeverri F.; Tamayo-Garcia A.; Linetsky E.; Ricordi C. Fully Automated Islet Cell Counter (ICC) for the Assessment of Islet Mass, Purity, and Size Distribution by Digital Image Analysis. Cell Transplant. 2016, 25 (10), 1747–1761. 10.3727/096368916X691655. [DOI] [PubMed] [Google Scholar]
  59. Kilimnik G.; Jo J.; Periwal V.; Zielinski M. C.; Hara M. Quantification of Islet Size and Architecture. Islets 2012, 4 (2), 167–172. 10.4161/isl.19256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Woods E. J.; Zieger M. A. J.; Lakey J. R. T.; Liu J.; Critser J. K. Osmotic Characteristics of Isolated Human and Canine Pancreatic Islets. Cryobiology 1997, 35 (2), 106–113. 10.1006/cryo.1997.2029. [DOI] [PubMed] [Google Scholar]
  61. Dentry M. B.; Yeo L. Y.; Friend J. R. Frequency Effects on the Scale and Behavior of Acoustic Streaming. Phys. Rev. E 2014, 89 (1), 013203. 10.1103/PhysRevE.89.013203. [DOI] [PubMed] [Google Scholar]
  62. Cheeke J. D. N.Fundamentals and Applications of Ultrasonic Waves, 1st ed.; CRC Press LLC, 2002. [Google Scholar]
  63. Farmehini V.; Kiendzior S.; Landers J. P.; Swami N. S. Real-Time Detection and Control of Microchannel Resonance Frequency in Acoustic Trapping Systems by Monitoring Amplifier Supply Currents. ACS Sens. 2021, 6 (10), 3765–3772. 10.1021/acssensors.1c01580. [DOI] [PubMed] [Google Scholar]
  64. Dual J.; Hahn P.; Leibacher I.; Möller D.; Schwarz T. Acoustofluidics 6: Experimental Characterization of Ultrasonic Particle Manipulation Devices. Lab Chip 2012, 12 (5), 852–862. 10.1039/c2lc21067c. [DOI] [PubMed] [Google Scholar]
  65. Shindo Y.; Kalivarathan J.; Saravanan P. B.; Levy M. F.; Kanak M. A.. Assessment of Culture/Preservation Conditions of Human Islets for Transplantation. Cell Transplantation 2022, 31, 10.1177/09636897221086966. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

la3c02934_si_001.pdf (1.1MB, pdf)
la3c02934_si_002.avi (224.8KB, avi)
la3c02934_si_003.avi (340KB, avi)

Articles from Langmuir are provided here courtesy of American Chemical Society

RESOURCES