Precise Size-Based Cell Separation via the ... - ACS Publications

Subscriber access provided by KEAN UNIV

Article

Precise size-based cell separation via the coupling of inertial microfluidics and deterministic lateral displacement Nan Xiang, Jie Wang, Qiao Li, Yu Han, Di Huang, and Zhonghua Ni Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.9b02863 • Publication Date (Web): 14 Jul 2019 Downloaded from pubs.acs.org on July 17, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1 2 3

Precise size-based cell separation via the coupling of inertial microfluidics and deterministic lateral displacement

4 5

Nan Xiang1*, Jie Wang1, Qiao Li1, Yu Han1, Di Huang2* and Zhonghua Ni1*

6 7 8 9

1School

of Mechanical Engineering, and Jiangsu Key Laboratory for Design and Manufacture of Micro-Nano Biomedical Instruments, Southeast University, Nanjing, 211189, China.

10 11

of Mechatronic Engineering, China University of Mining and Technology, Xuzhou, 221116, China.

12

*E-mails: [email protected]; [email protected]; [email protected].

2School

13 14

Abstract

15

We report here a novel two-stage i-DLD sorter through coupling inertial microfluidics with deterministic lateral displacement (DLD), allowing for precise, continuous and size-based cell separation. The 1st stage spiral inertial microfluidic sorter is responsible for removing the overwhelming majority of background blood cells at a highthroughput manner. The precise and flow-rate insensitive DLD sorter with triangular posts serves as the 2nd stage sorter which further removes the residual blood cells for obtaining high-purity tumor cells. After demonstrating the conceptual design, we characterize the performances of our two-stage i-DLD sorter for the separation of differently-sized particles and cells. The characterization results show that a 100% complete separation of 15 μm and 7 μm particles was achieved while a separation efficiency of over 99.9% and a target sample purity of 93.59% was realized for the separation of differently-sized cells. Finally, we successfully apply our sorter for the separation of rare tumor cells from the diluted whole blood or WBCs at good performances. Our two-stage i-DLD sorter offers numerous advantages of label- and external field-free operation, high-efficiency and high-reliability separation, and highthroughput processing without clogging, and is promise to be applied as a potential tool for precise cell separation in low-resource settings.

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

1

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Introduction

2

Separation of desired cells from complex biofluids is an important but challenging sample pretreatment step for various downstream disease diagnosis or biomedical research.1-3 For example, the isolation of rare circulating tumor cells (CTCs) from peripheral blood has been regarded as a non-invasive “liquid biopsy” technique for therapeutic efficacy monitoring, personalized therapy and potential diagnosis of cancer at the early stage.4,5 Up until now, the flow cytometry,6 fluorescence activated cell sorting (FACS)7 and magnetic-activated cell sorter (MACS)8 are still the most commonly employed methods for realizing the cell separation. However, these methods require expensive instruments and complex pre-labeling of fluorophores or magnetic beads, which prevents the wide application of these methods in low-resource settings. In addition, these methods cannot reliably handle small numbers of cells and thus are unable to sort rare cells. The advent of microfluidics has provided new insights for the on-chip cell separation. As a promising technique for engineering cells at microscale, microfluidics offers advantages of small volume, low cost, ease of integration and high precision.9,10 According to the working principle, the reported microfluidic sorters can be divided into two categories: active and passive ones. The active microfluidic sorters employ the external (e.g., electric,11 magnetic,12 acoustic13 and optical14) fields to achieve the cell separation according to the differences in size, refractive index, magnetic susceptibility or dielectric property. Instead, the passive microfluidic sorters apply the inherent fluidic effects (e.g., inertial,15,16 viscoelastic17 and hydrophoresis18) and the specificallydesigned microstructures (e.g., deterministic lateral displacement (DLD)19-21 and microfilter22) to sort cells on the basis of their differences in size, shape or deformability. Although great successes have been achieved, there remain some limitations for the microfluidic sorters solely using single techniques. The hybrid technique separation is promise for addressing the limitations of single technique and has attracted increasing interests in recent years.23 The first hybrid separation strategy is the coupling of both active and passive techniques. For examples, the researchers coupled the inertial microfluidics with dielectrophoresis24,25 or magnetic deflection26,27 for improving the separation performance. Toner et al.28 developed a CTC-iChip in which the DLD was coupled with magnetophoresis for the tumor antigen–independent separation of CTCs. However, the bulk external field generator or complex labeling are still required, which significantly limits the practical application of these devices in low-resource settings. The microfluidic sorters integrated with multiplex passive techniques are more suitable for use in low-resource settings due to the offered specific advantages of labeland external force-free operation, simple self-contained device and relatively high processing throughput. One straightforward integration strategy is the cascading of single passive techniques (e.g., multiorifice flow fractionation,29,30 straight or spiral

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

2


Page 2 of 18

Page 3 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

inertial microfluidics31-33 and hydrodynamic-vortex separation34,35) in series. Another strategy is the integration of two or more different separation techniques. As the integrated devices can inherit the advantages from different techniques, this strategy is more benefit for improving the separation performance. Sun et al.36 integrated the spiral inertial microfluidics with a membrane filter and successfully applied the integrated device for the isolation of CTCs from blood. The tumor cells captured by the membrane holes can be directly used for the immunostaining enumeration, but the recovery of these captured cells is not easy. Xuan et al.37 proposed an inertia-enhanced pinched flow fractionation (iPFF) technique for significantly improving the particle separation in the traditional PFF. However, the throughput of this technique is still relatively low. Wang et al.38 developed a device by combining inertial focusing with hindrance filtration for achieving the cell separation. However, the purity of the collected samples is still not attractive enough due to the low-efficiency of the hindrance filtration. Even though advances have been made, the new hybrid separation techniques for precise, continuous and label-free cell sorting are still in urgent demands. Herein, we propose a novel two-stage i-DLD sorter through coupling the inertial microfluidics with the DLD for achieving the precise, continuous and label-free cell separation. The high-throughput spiral inertial microfluidics is employed as the 1st stage sorter to remove the overwhelming majority of background cells for reducing the clogging risk of next stage. The precise DLD sorter with triangular posts serves as the 2nd stage sorter which further removes the residual background cells to obtain highpurity target cells. Our two-stage i-DLD sorter is totally passive in principle and offers numerous advantages of label- and external field-free operation, high-efficiency and high-reliability separation, and high-throughput processing without clogging. After demonstrating the design concept, we characterize the performance of our sorter for the size-based separation of particles and cells. Finally, we explore the capability of our sorter for the separation of rare tumor cells from the diluted blood or WBCs. We envision the wide application of our two-stage i-DLD sorter as a potential tool for precise cell separation in low-resource settings.

30 31

Materials and methods

32

Conceptual design and device optimization Figure 1(a) illustrates the working principle of our two-stage i-DLD device which couples the spiral inertial microfluidic sorter with the DLD sorter for the precise, continuous and label-free separation of rare tumor cells from background blood cells. The blood sample was first pumped into the 1st stage spiral inertial microfluidic sorter through the inlet. When flowing in the spiral channel at specific flow rates, the cells will simultaneously suffer from the net inertial lift force (FL) induced by the shear

33 34 35 36 37 38

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

gradient and the wall effect, and the Dean drag force (FD) induced by the Dean vortex in curving channels.39-41 Under the balance of these two lateral forces, the randomlydispensed cells will be gradually focused into cell strings when arriving at the outlet. To successfully achieve the cell focusing, two criterions need to be satisfied. (1) A specific driving flow rate needs to be provided to induce the fluidic inertial effects in finite-Reynolds-number flows (particle Reynolds number Re𝑝 = 𝜌𝑈𝑎2𝑝 𝜇𝐻 ≥ 1, where 𝜌 is the fluid density, U is the average fluid velocity, ap is the cell diameter, μ is the dynamic viscosity, and H is the characteristic channel dimension and can be approximated as channel height (h) for low-aspect-ratio channels).42 (2) The particle confinement ratio (CR = 𝑎𝑝 𝐻) needs to satisfy CR ≫ 0.07.43,44 To ensure the complete focusing of small blood cells, the channel height (h) was determined to be 50 μm according to the above criterion. Other geometry parameters of the spiral channel can be found in Table S1. This spiral channel design was proved to be able to separate tumor cells and blood cells according to their size difference over wide flow rates. As observed in our previous study45, the large tumor cells form a cell string close to the channel centerline while the abundant small blood cells focus into a cell band near the inner wall under the optimal flow rate of 400 μL/min. In this work, we directly employed this channel design as the 1st stage sorter for removing the majority of background blood cells. Detailed discussions on the effects of flow rate, particle size and channel geometry on cell separation in spiral channels can be found in our previous studies44-47 and are not repeated here. Before integrating with the 2nd stage DLD sorter, it is necessary to optimize the length of the spiral channel for the purpose of reducing the flow resistance of the whole device. Figure S1 illustrates the separation performances of 7 μm and 15 μm particles at different channel loops under the flow rate of 400 μL/min. Form the experimental result, it was found that the two-sized particles can be well separated at the loop 4. Therefore, the length of the spiral channel was determined to be ~12 cm (4 loops). At the end of the spiral channel, an expansion channel was employed to amplify the distance between the two separate cell strings and a bifurcated outlet system (upper part: lower part=1:2) was applied to remove the separated blood cells or to export the target cells into the 2nd stage sorter.

4


Page 4 of 18

Page 5 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2 3 4 5 6 7

Figure 1. (a) Working principle of our two-stage i-DLD device which couples the spiral inertial microfluidic sorter with the DLD sorter for the precise, continuous and labelfree separation of tumor cells from background blood cells. (b) The CAD drawing illustrating the detailed structures of our two-stage i-DLD device. (c) Photograph of the final device fabricated in PDMS using soft lithography. The channel was fully filled with red ink for clear visualization. (d) SEM image of the fabricated triangular posts.

8 9 10 11 12 13 14 15

Although the 1st stage inertial microfluidic sorter is capable of separating the tumor cells from the blood cells, a small part of blood cells will still enter into the target sample outlet, which results in the significant deterioration of the sample purity. To address this issue, we connected the DLD sorter with the inertial microfluidic sorter in series to improve the separation performance. The DLD sorter is capable of continuously separating cells based on their sizes with a separation resolution up to 10 nm,19 and thus is a good choice for separating cells with close sizes. The post in the DLD sorter was 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

designed to be in the triangular shape with its side length (S) of 60 μm. As compared to DLD sorter with other post shapes (e.g., circular, quadrilateral and I-shaped), the DLD sorter with triangular posts has specific advantages of reducing clogging and lowering pressure drop.19,48 The period (N) was selected to be 20 (the row shift fraction (ε=1/N) equals 1/20), which means that there are N flow streamlines between the two adjacent posts. The gap distance (G) was measured to be 25 μm, and the boundary interface was modified to reduce the aberrant flow in the areas adjacent to the boundaries49 (see Figure S2). In the DLD sorter, the target tumor cells will move in the displacement mode due to the cell-post interaction and shift a row down after passing 20 lateral posts (20 lateral posts are served as a post array unit). We designed 22 repeated post array units to ensure that the large tumor cells can shift towards the lower boundary and can be collected via the outlet III. Meanwhile, the residual blood cells with diameters smaller than the critical diameter (Dc) will migrate in the zigzag mode and can be exported via the outlet II. To verify our DLD design, we explored the particle/cell migration dynamics in a separate DLD sorter. Figure S3 shows the distributions of 15 μm and 7 μm particles near the outlet of the separate DLD sorter at the flow rates of 40~400 μL/min. In addition to the particles, the migration behaviors of tumor and blood cells were explored and illustrated in Figure S4. The above experimental results well support two conclusions: (1) The flow rate has little effect on particle/cell migration, which makes our triangular post DLD sorter be an ideal choice as the 2nd stage sorter (the precise matching of the operational flow rates of the two-stage sorters is not required); (2) Our DLD sorter is capable of separating the large tumor cells from the background blood cells. The critical diameter (Dc) of our DLD sorter under the inertial-flow condition (Rep >1) was estimated to be between the sizes of tumor cells and blood cells. When coupling these two sorters, the focused tumor cells along with the residual blood cells will enter into the upper half of the DLD region. The cell-free fluid near the bottom half of the spiral channel can subtly act as the sheath flow for confining the cells in the upper half of the DLD region. This feature makes that our 2nd stage DLD sorter can be operated in a sheath-free manner. To explore the capability of our sorter for processing high-concentration samples, the blood samples with different concentrations of 106~108 counts/mL were tested (see Figure S5). It was found that the cell distribution widths in both channels will increase with increasing the cell concentration. To achieve the best performance, it is better to control the initial concentration below 107 counts/mL.

36 37 38 39

Device fabrication Our two-stage i-DLD device was fabricated using the well-established soft lithography technique. The detailed fabrication process can be found in the 6


Page 6 of 18

Page 7 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2 3 4

supplementary section S1. Figure 1(b) shows the CAD drawing of our two-stage i-DLD sorter which has only one inlet for permitting the sheathless operation. Figure 1(c, d) respectively illustrates the photograph of our two-stage i-DLD sorter fabricated in the material of PDMS and the SEM image of the triangular post array.

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Sample preparation Two types of polystyrene particles with diameters of 7 μm and 15 μm were purchased from Thermo Fisher Scientific or Bangs Laboratories for testing the device performance. The particle solutions were diluted with 0.5 wt% Tween 20 mixed phosphate buffer saline (PBS) buffer (Sigma-Aldrich) for preparing the particle suspensions with specific concentrations. Human breast adenocarcinoma cells (MCF-7 cells) were cultured in the highglucose Dulbecco’s modified Eagle’s medium (DMEM, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific) and 1% penicillin-streptomycin (Thermo Fisher Scientific). After growing to the confluence, the tumor cells were harvested through dissociating using Trypsin-EDTA (0.25%, Thermo Fisher Scientific) solution and then re-dispersed in sterile PBS buffer. To easily discriminate tumor cells from background blood cells, the tumor cells were stained with Calcein AM (Thermo Fisher Scientific) according to the manufacturer's instruction. The collected tumor cells after being processing with our sorter were re-cultured and periodically inspected. Human whole blood was draw from a healthy consenting volunteer using a vacutainer collection tube (BD Biosciences) containing anticoagulant K2EDTA. The study was approved by the institutional committee of Institutional Ethical Committee (IEC) for Clinical Research of Zhongda Hospital (Southeast University). The blood samples were diluted with the PBS buffer (Sigma-Aldrich) to different cell concentrations and then a certain amount of stained tumor cells were spiked into the diluted blood samples. For preparing the WBC samples, the whole blood was lysed with Ammonium-Chloride-Potassium (ACK) lysing buffer (Thermo Fisher Scientific) under gently shaking for 5 min. Then, the RBC debris was removed through centrifugation and the obtained WBCs were re-suspended in sterile PBS buffer.

32 33 34 35 36 37 38 39

Experimental setup The fabricated device was mounted onto the platform of an invert microscope (IX 71, Olympus). The inlet and outlet orifices were respectively connected with a syringe pump (Legato 270, KD Scientific) for driving the sample at a stable flow rate and centrifuge tubes for sample collection via Teflon tubings. Before use, the device was pumped through with the FBS solution to reduce the adhesion between cell and channel structure. The particle/cell motions in the device were captured using a high-speed 7



1 2 3 4 5 6 7 8 9 10

camera (Phantom V611, Vision Research) under both bright-field and fluorescence observation modes. The captured image frames were then processed with the ImageJ software (NIH) to create the composite images illustrating the particle/cell distributions over a certain time period. To characterize the separation performance of our device, the concentrations of the initial samples and the samples collected from the three outlets were sampled and counted using a Countess II FL automated cell counter (Thermo Fisher Scientific) for several times. The volumes of the samples were also recorded. For the experiments concerning the separation of rare cells, the tumor cells in the collected samples were enumerated after centrifugation.

11 12

Results and discussion

13

Complete separation of differently-sized particles After demonstrating the conceptual design, we applied our two-stage i-DLD sorter for the separation of 15 μm and 7 μm particles to verify its effectiveness. The prepared particle suspension was pumped into the inlet of our two-stage i-DLD sorter at the flow rate of 400 μL/min.

14 15 16 17

8


Page 8 of 18

Page 9 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2 3

Figure 2. (a) Composite images illustrating the particle distributions at different locations (i~vi). (b) Microscopic images of the samples collected from three outlets.

4 5 6 7 8 9 10 11 12 13 14 15 16 17

Figure 2(a) illustrates the particle distributions at representative locations (i~vi). In the 1st stage spiral inertial microfluidic sorter (i), the randomly-dispensed particles near the inlet (loop 0) gradually focus into two distinct particle strings when migrating to the loop 4. The string of small 7 μm particles has a lateral position close to the inner wall while the 15 μm particle string locates close to the channel centerline. With the assistance of a sudden expansion channel (ii), the distance between the two particle strings can be significantly amplified for helping separation. After passing through the sudden expansion channel, the string of small particles can be easily removed via the outlet I. However, due to the particle-particle collision and the relatively unstable focusing, a small part of unfocused small particles will enter into the 2nd stage DLD sorter together with the large particle string (iii). In the DLD array (iv, v and vi), the residual small particles migrate in the zigzag mode and can be exported via the outlet II while the large particles migrate in the displacement mode. After passing through the 9



1 2 3 4 5 6

whole DLD array, the large particles migrate towards the lower boundary and can be collected via the outlet III (the supplementary video S1). Figure 2(b) shows the microscopic images of the samples collected from the three outlets. It was found that a 100% complete separation of these two particles is achieved using our two-stage i-DLD sorter and the 2nd stage DLD sorter can effectively remove the residual small particles to significantly increase the purity of target sample (100% purities for both particles).

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Characterization of cell separation performance We next characterized the cell separation performance of our two-stage i-DLD sorter. The diluted blood sample (5 × 106 counts/mL) spiked with a relatively large number of tumor cells (~104 counts/mL) was pumped into the sorter at the flow rate of 400 μL/min. Figure 3(a) illustrates the cell distributions at different locations. After passing through the 1st spiral inertial microfluidic sorter, the large tumor cells (marked with red circulars) can be separated from the band of blood cells at the sudden expansion channel (i). However, these is a considerable part of blood cells being mixed with the target tumor cells (see (ii) in Figure 3(a) and the supplementary video S2), which results in the serious contamination of the target sample. In our two-stage i-DLD sorter, these residual blood cells can be removed through the 2nd stage DLD sorter (iii: middle region and iv: outlet region). As can be observed from the outlet region of the DLD sorter ((iv) in Figure 3(a) and the supplementary video S3), the target tumor cells will gradually shift towards the lower wall boundary and can be well separated from the residual blood cells.

10


Page 10 of 18

Page 11 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2 3 4 5 6 7

Figure 3. (a) Composite images illustrating the cell distributions at different locations (i~iv). In these images, the tumor cells were marked with red circulars. (b) Photograph and microscopic images of the samples collected from the three outlets. The fluorescence images at the upper-right corners of each microscopic image were used to identify the stained tumor cells. (c) Separation efficiencies of blood cells and tumor cells at each outlet.

8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 3(b) illustrates the photograph and the microscopic images of the samples collected from the three outlets. It was clearly observed that the sample collected from outlet I appears red color due to the existence of large numbers of blood cells. In the microscopic images sampled from outlet III (see Figures 3(b) and S6), very high-purity tumor cells can be easily obtained using our sorter, which enables various subsequent biological studies (e.g., next generation gene sequencing and bioinformatics analysis) that require high-purity samples to be possible. To further quantitatively evaluate the separation performance of our two-stage i-DLD sorter, the separation efficiency (number of specific cells in the one outlet/number of specific cells in all outlets) was calculated. Figure 3(c) illustrates the separation efficiencies of both tumor cells and blood cells. The total separation efficiency of blood cells in outlets I and II can be regarded as the blood-cell removing ratio. The 1st stage spiral inertial microfluidic sorter can remove 96.31% of blood cells while an additional 3.63% of blood cells can 11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

be further removed through the 2nd stage DLD sorter. Through using our two-stage separation strategy, a total blood-cell removing ratio of 99.94% was achieved. In addition, the separation efficiency of tumor cells in outlet III was calculated to be as high as 100%. The purity of the tumor cells collected from outlet III was measured through enumerating cells in sampled windows (see Figure S7). The purity of the separated tumor cells can reach 93.59%, which is much higher than the purity (only 0.94%) that can be achieved solely using the spiral inertial microfluidic sorter.45 In addition to high-efficiency, another important feature of our two-stage i-DLD sorter is high-reliability. The separation performance of the 1st stage spiral inertial microfluidic sorter was found to be highly sensitive to the operational flow rate and the cell concentration.15,45,46 Our two-stage i-DLD sorter can achieve a good separation performance even when the 1st stage sorter does not work well under the improper flow rate or cell concentration. Figure S8(a) shows the cell distributions at different locations. It was found that nearly all the tumor cells can be separated into the DLD array, but the focusing performance of small blood cells is unsatisfactory and a lot of blood cells will also enter into the DLD sorter. However, in our two-stage i-DLD device, the 2nd stage DLD sorter can make up for the low-efficiency of the 1st stage sorter as the separation performance of the 2nd stage DLD sorter is to a certain degree insensitive to flow rate or cell concentration. Figure S8(b) shows the microscopic images of the initial sample and the samples collected from the three outlets. Figure S8(c) illustrates the separation efficiency of blood cells at each stage. It was found that the 1st stage sorter only removes 25.51% of blood cells while the 2nd stage sorter removes 72.19% of blood cells. The clogging was not observed in the DLD array (see the supplementary video S4) even when processing the sample with a concentration of large tumor cells up to 104 counts/mL. In future, our two-stage i-DLD sorter is promise to be widely employed for the label-free and size-based separation of various cells at a good separation efficiency.

28 29 30 31 32 33 34 35 36 37 38

Separation of rare tumor cells We next applied our two-stage i-DLD sorter for the separation of rare tumor cells from blood. The blood sample with a cell concentration of 5 × 107 counts/mL was spiked with 500 tumor cells per mL and pumped into the sorter at the flow rate of 400 μL/min. Figure 4(a) shows the photograph of the collected samples. Figure 4(b-d) and Figure S9 are the sampled microscopic images under both bright-field and fluorescence observation modes. It was observed that an overwhelming majority of blood cells can be removed via the outlet I after the 1st separation stage, and thus the collected sample appears blood red. Meanwhile, the residual blood cells can be removed via the outlet II after passing through the 2nd stage DLD sorter. The total removing ratio of blood cells 12


Page 12 of 18

Page 13 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2 3 4 5 6 7 8

9 10 11 12 13 14

is as high as 99.95%, and there is no tumor cell being observed in the sampled windows of samples collected from these two outlets. In the microscopic images of target samples, there are few blood cells being observed around the target tumor cell (see Figure 4(d)). As the cell concentration in the target sample is extremely low, the sample was concentrated and an example of the microscopic image of the concentrated sample was illustrated in the inset of Figure 4(d). In addition to WBCs, the target tumor cells were still contaminated by a few RBCs which were mixed into the target sample at the startup time for achieving focusing or at the end of the sample injection.

Figure 4. (a) Photograph of the collected samples after performing the experiment on the separation of rare tumor cells from blood. (b~d) Microscopic images of the samples collected from the three outlets. The inset in subfigure (d) is the microscopic image of the concentrated sample from outlet III. (e) Recovery ratio and purity of the target tumor cells.

15 16 17 18 19 20 21 22 23 24 25 26 27 28

To quantitatively access the performance of our sorter for the separation of rare cells, the recovery ratio and purity of the target sample were calculated. The recovery ratio was defined as the ratio between the number of target tumor cells collected from outlet III and the number of tumor cells in the initial sample. Figure 4(e) shows the calculated recovery ratio and purity of the target tumor cells. The average recovery ratio and purity were respectively calculated to be 91.34% and 17.68%. As the blood cells are rather overwhelming when compared to tumor cells, the tumor-to-blood cell ratio has increased for over 105 fold after processing with our sorter. One of the challenge in CTCs enrichment is the separation of CTCs from large WBCs. We next explored the performance of our two-stage i-DLD sorter for separating rare tumor cells from WBCs. To increase the processing efficiency of our sorter, the RBCs in the whole blood were lysed, and then the WBCs were re-suspended in the PBS buffer to the original volume of whole blood, obtaining a WBC suspension with a high 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

concentration up to 106 counts/mL. Then, the prepared WBC suspension was spiked with 500 tumor cells per ml, and pumped into the sorter. Figure S10(a-e) illustrates the photograph of the collected samples and the microscopic images of samples collected from the three outlets. It was found that our sorter can realize the separation of rare tumor cells from WBCs at a good performance. Figure S10(f) shows the recovery ratio and the purity of tumor cells. An average recovery ratio of 71.92% and an average purity of 15.48% were achieved. The decrease of recovery ratio may be due to the heavy interactions between the large WBCs under a high concentration. With the assistance of RBC lysis, our two-stage i-DLD sorter is able to process the 1 ml whole blood within 2.5 min. Further increasing the throughput of our sorter can be realized through stacking multiplex channel layers. Cell viability test was performed to examine whether the shear stress in high-speed inertial flows and the cell-post interaction will induce the death of target tumor cells. The result of trypan blue exclusion test indicates that a cell viability of over 95% was kept for the collected samples after running through our sorter. In addition, the separated tumor cells were re-cultured for 48, 72, 96 and 120 hours. Figure S11 shows the microscopic images of the re-cultured cells at different time scales, which indicates that our sorter has a negligible effect on cell viability, and the separated cells can be recultured for various downstream applications. Therefore, our two-stage i-DLD sorter can be potentially acted as the “liquid biopsy” tool for the label-free separation of highviability and high-purity CTCs from human whole blood.

22 23

Conclusions

24

In this work, we develop a novel two-stage i-DLD sorter for the precise, continuous and label-free cell separation through coupling the two passive techniques of inertial microfluidics and DLD. After verifying the conceptual design, we characterize the separation performances of our two-stage i-DLD sorter for the separation of particles and cells. The characterization results show that a 100% complete separation of 15 μm and 7 μm particles is achieved while a separation efficiency over 99.9% and a target sample purity of 93.59% are realized for the separation of differently-sized cells. In addition to the high-efficiency, our two-stage i-DLD sorter can achieve a highreliability separation even when the 1st stage sorter does not work well. Then, we apply our sorter for the separation of rare tumor cells from blood. It is found that a blood cell removing ratio of 99.95% and an average tumor cell recovery ratio of 91.34% can be achieved. Finally, we show that our sorter is capable of separating rare tumor cells from high-concentration WBCs at a good performance. Our two-stage i-DLD sorter offers numerous advantages of label- and external field-free operation, high-efficiency and high-reliability separation, and high-throughput processing without clogging and is promise to be applied as a new tool for precise cell separation in low-resource settings.

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

14


Page 14 of 18

Page 15 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2

Conflicts of interest

3

There are no conflicts to declare.

4 5

Acknowledgements

6

This research work is supported by the National Natural Science Foundation of China (81727801, 51875103, 51775111 and 51505082), the Natural Science Foundation of Jiangsu Province (BK20150606), the Fundamental Research Funds for the Central Universities (2242017K41031), the Six Talent Peaks Project of Jiangsu Province (SWYY-005) and the Zhishan Youth Scholar Program of SEU.

7 8 9 10 11 12

Supporting Information

13

Particle separation performances at different spiral channel loops, structure of the separate DLD sorter, particle and cell distributions at the outlet of the DLD sorter, effect of cell concertation on blood cell distributions in the spiral channel and the DLD sorter, microscopic images of the samples collected from outlet III, microscopic images of the collected tumor cells for illustrating the sample purity, separation performance of our two-stage i-DLD sorter when the 1st stage sorter does not work well, fluorescence images of the samples collected from the three outlets, separation of rare tumor cells from WBCs, microscopic images of the re-cultured tumor cells at different time scales, dimensions of the spiral channel, fabrication process of our two-stage i-DLD sorter, video illustrating the complete particle separation at the outlet of the 2nd stage DLD sorter, videos illustrating the cell separation performances at the outlets of the 1st stage spiral sorter and the 2nd stage DLD sorter, video illustrating that the clogging was not observed in the DLD array.

14 15 16 17 18 19 20 21 22 23 24 25 26 27

Reference

28 29 30 31 32 33 34 35 36 37 38 39

(1) Mu, X.; Zheng, W.; Sun, J.; Zhang, W.; Jiang, X. Small 2013, 9, 9-21. (2) Wyatt Shields Iv, C.; Reyes, C. D.; López, G. P. Lab Chip 2015, 15, 1230-1249. (3) Plouffe, B. D.; Murthy, S. K. Anal. Chem. 2014, 86, 11481-11488. (4) Rawal, S.; Yang, Y.-P.; Cote, R.; Agarwal, A. Annu. Rev. Anal. Chem. 2017, 10, 321-343. (5) Jackson, J. M.; Witek, M. A.; Kamande, J. W.; Soper, S. A. Chem. Soc. Rev. 2017, 46, 4245-4280. (6) Laerum, O. D.; Farsund, T. Cytometry 1981, 2, 1-13. (7) Bonner, W. A.; Hulett, H. R.; Sweet, R. G.; Herzenberg, L. A. Rev. Sci. Instrum. 1972, 43, 404-409. (8) Adams, J. D.; Kim, U.; Soh, H. T. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 18165-18170. 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

(9) Sackmann, E. K.; Fulton, A. L.; Beebe, D. J. Nature 2014, 507, 181. (10) Sibbitts, J.; Sellens, K. A.; Jia, S.; Klasner, S. A.; Culbertson, C. T. Anal. Chem. 2018, 90, 65-85. (11) Kim, D.; Sonker, M.; Ros, A. Anal. Chem. 2019, 91, 277-295. (12) Chen, Q.; Li, D.; Malekanfard, A.; Cao, Q.; Lin, J.; Wang, M.; Han, X.; Xuan, X. Anal. Chem. 2018, 90, 8600-8606. (13) Li, P.; Huang, T. J. Anal. Chem. 2019, 91, 757-767. (14) Wang, M. M.; Tu, E.; Raymond, D. E.; Yang, J. M.; Zhang, H.; Hagen, N.; Dees, B.; Mercer, E. M.; Forster, A. H.; Kariv, I.; Marchand, P. J.; Butler, W. F. Nat. Biotechnol. 2005, 23, 83-87. (15) Amini, H.; Lee, W.; Di Carlo, D. Lab Chip 2014, 14, 2739-2761. (16) Zhang, J.; Yan, S.; Yuan, D.; Alici, G.; Nguyen, N.-T.; Ebrahimi Warkiani, M.; Li, W. Lab Chip 2016, 16, 10-34. (17) Lu, X.; Liu, C.; Hu, G.; Xuan, X. J. Colloid Interface Sci. 2017, 500, 182-201. (18) Choi, S.; Song, S.; Choi, C.; Park, J.-K. Anal. Chem. 2009, 81, 1964-1968. (19) McGrath, J.; Jimenez, M.; Bridle, H. Lab Chip 2014, 14, 4139-4158. (20) Huang, L. R.; Cox, E. C.; Austin, R. H.; Sturm, J. C. Science 2004, 304, 987-990. (21) Tottori, N.; Nisisako, T. Anal. Chem. 2019, 91, 3093-3100. (22) Liu, Y.; Li, T.; Xu, M.; Zhang, W.; Xiong, Y.; Nie, L.; Wang, Q.; Li, H.; Wang, W. Lab Chip 2019, 19, 68-78. (23) Yan, S.; Zhang, J.; Yuan, D.; Li, W. Electrophoresis 2017, 38, 238-249. (24) Zhang, J.; Yuan, D.; Zhao, Q.; Yan, S.; Tang, S.-Y.; Tan, S. H.; Guo, J.; Xia, H.; Nguyen, N.-T.; Li, W. Sens. Actuators, B 2018, 267, 14-25. (25) Moon, H.-S.; Kwon, K.; Kim, S.-I.; Han, H.; Sohn, J.; Lee, S.; Jung, H.-I. Lab Chip 2011, 11, 1118-1125. (26) Chen, Q.; Li, D.; Lin, J.; Wang, M.; Xuan, X. Anal. Chem. 2017, 89, 6915-6920. (27) Zhou, Y.; Song, L.; Yu, L.; Xuan, X. Microfluid. Nanofluid. 2017, 21, 14. (28) Karabacak, N. M.; Spuhler, P. S.; Fachin, F.; Lim, E. J.; Pai, V.; Ozkumur, E.; Martel, J. M.; Kojic, N.; Smith, K.; Chen, P.-i.; Yang, J.; Hwang, H.; Morgan, B.; Trautwein, J.; Barber, T. A.; Stott, S. L.; Maheswaran, S.; Kapur, R.; Haber, D. A.; Toner, M. Nat. Protoc. 2014, 9, 694. (29) Sim, T. S.; Kwon, K.; Park, J. C.; Lee, J.-G.; Jung, H.-I. Lab Chip 2011, 11, 9399. (30) Moon, H.-S.; Kwon, K.; Hyun, K.-A.; Sim, T. S.; Park, J. C.; Lee, J.-G.; Jung, H.-I. Biomicrofluidics 2013, 7, 014105. (31) Tanaka, T.; Ishikawa, T.; Numayama-Tsuruta, K.; Imai, Y.; Ueno, H.; Matsuki, N.; Yamaguchi, T. Lab Chip 2012, 12, 4336-4343. (32) Miller, B.; Jimenez, M.; Bridle, H. Sci. Rep. 2016, 6, 36386. (33) Abdulla, A.; Liu, W.; Gholamipour-Shirazi, A.; Sun, J.; Ding, X. Anal. Chem. 2018, 90, 4397-4405. (34) Wang, X.; Papautsky, I. Lab Chip 2015, 15, 1350-1359. (35) Wang, X.; Yang, X.; Papautsky, I. TECHNOLOGY 2016, 04, 88-97. (36) Wang, J.; Lu, W.; Tang, C.; Liu, Y.; Sun, J.; Mu, X.; Zhang, L.; Dai, B.; Li, X.; Zhuo, H.; Jiang, X. Anal. Chem. 2015, 87, 11893-11900. (37) Lu, X.; Xuan, X. Anal. Chem. 2015, 87, 4560-4565. (38) Shen, S.; Ma, C.; Zhao, L.; Wang, Y.; Wang, J.-C.; Xu, J.; Li, T.; Pang, L.; Wang, J. Lab Chip 2014, 14, 2525-2538. (39) Sun, J.; Li, M.; Liu, C.; Zhang, Y.; Liu, D.; Liu, W.; Hu, G.; Jiang, X. Lab Chip 16


Page 16 of 18

Page 17 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19


2012, 12, 3952-3960. (40) Liu, C.; Xue, C.; Sun, J.; Hu, G. Lab Chip 2016, 16, 884-892. (41) Zhang, J.; Yuan, D.; Zhao, Q.; Teo, A. J. T.; Yan, S.; Ooi, C. H.; Li, W.; Nguyen, N.-T. Anal. Chem. 2019, 91, 4077-4084. (42) Di Carlo, D.; Irimia, D.; Tompkins, R. G.; Toner, M. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 18892-18897. (43) Kuntaegowdanahalli, S. S.; Bhagat, A. A. S.; Kumar, G.; Papautsky, I. Lab Chip 2009, 9, 2973-2980. (44) Xiang, N.; Shi, Z.; Tang, W.; Huang, D.; Zhang, X.; Ni, Z. RSC Adv. 2015, 5, 77264-77273. (45) Huang, D.; Shi, X.; Qian, Y.; Tang, W.; Liu, L.; Xiang, N.; Ni, Z. Anal. Methods 2016, 8, 5940-5948. (46) Xiang, N.; Yi, H.; Chen, K.; Sun, D.; Jiang, D.; Dai, Q.; Ni, Z. Biomicrofluidics 2013, 7, 044116. (47) Xiang, N.; Chen, K.; Dai, Q.; Jiang, D.; Sun, D.; Ni, Z. Microfluid. Nanofluid. 2015, 18, 29-39. (48) Loutherback, K.; Chou, K. S.; Newman, J.; Puchalla, J.; Austin, R. H.; Sturm, J. C. Microfluid. Nanofluid. 2010, 9, 1143-1149. (49) Inglis, D. W. Appl. Phys. Lett. 2009, 94, 013510.

20 21

17



1

For Table of Contents Only

2

18


Page 18 of 18

Precise Size-Based Cell Separation via the ... - ACS Publications

Recommend Documents