Inertial microfluidic cube for automatic and fast extraction of white blood cells from whole blood

Shu Zhu a, Dan Wu b, Yu Han a, Cailian Wang c, Nan Xiang *a and Zhonghua Ni *a
aSchool of Mechanical Engineering and Jiangsu Key Laboratory for Design and Manufacture of Micro-Nano Biomedical Instruments, Southeast University, Nanjing, 211189, China. E-mail: nan.xiang@seu.edu.cn; nzh2003@seu.edu.cn
bDepartment of Oncology, Jiangyin People's Hospital, Jiangyin, 214400, China
cTumor Center of Zhongda Hospital, Southeast University, Nanjing 210009, China

Received 21st September 2019 , Accepted 2nd December 2019

First published on 4th December 2019


We report here a novel inertial microfluidic (IM) cube integrated with lysis, storage and extraction modules for extracting white blood cells (WBCs) from whole blood automatically, harmlessly and quickly. Lysis, storage, and extraction modules are designed to realize the purposes of complete mixing of whole blood and lysing buffer, thorough lysis of red blood cells (RBCs), and automatic extraction of WBCs from the lysed background RBCs, respectively. After demonstrating its conceptual design, we characterize the performances of the lysis and extraction modules. The results show that a high mixing efficiency of 94.2% can be achieved using our lysis modules for complete mixing of whole blood and lysing buffer. In the extraction module, an extraction efficiency of 88.1% can be achieved for the extraction of WBCs. Finally, we successfully apply our IM cube for the high throughput extraction of WBCs from human whole blood with an extraction efficiency of 83.9% and a cell viability of 96.6%, which are comparable to those using centrifugation and even better in some aspects. Our IM cube is based on passive secondary-flow mixing and inertial sorting, offers the advantages of small footprint, high stability and simple fabrication, and is a promising alternative method for extracting WBCs from human blood.


Introduction

White blood cells (WBCs) usually account for 0.1% of total human blood cells and contain abundant information for blood-based diagnostics.1–3 For example, WBC count has proven to be a clinical indicator of coronary heart disease (CHD).4 In addition, epidemiologic studies have proposed the relationships between cancers (e.g., bladder, breast and stomach cancers) and global WBC methylation.5 The successful detection and characterization of WBCs usually require the depletion of red blood cells (RBCs) and other blood cell components.6–8 Therefore, the extraction of WBCs from whole blood has been regarded as a critical sample preparation step for the WBC detection. In the conventional WBC extraction method, the background RBCs in whole blood are selectively lysed using a hypotonic solution, and then the lysed RBC debris was removed through repeated centrifugation.9,10 However, it has been reported that a high acceleration shock during high-speed centrifugation may probably impair the membrane of the extracted WBCs and induce the death of cells.11–13 In addition to the above RBC lysing method, other methods, such as density gradient centrifugation14 and cell filtration,15 have been invented for the extraction of WBCs in some circumstances. These macroscale methods are time-consuming, labor-intensive and high cost, and may introduce variability in the extraction of WBCs owing to their imprecise operations and nonuniform conditions.16,17

With the advent of microfluidics, high-efficiency microfluidic devices have been reported for the extraction of WBCs at the microscale. As compared with conventional methods, the microfluidic WBC extraction methods offer potential advantages of small footprints, relatively simple structures and low cost.18–20 According to their working principles, the reported microfluidic WBC extraction devices can be divided into active and passive ones. In active microfluidic devices, active external forces (e.g., electric and magnetic) are employed to extract WBCs according to their dielectric property or magnetic susceptibility. For example, Cetin et al.21 reported an alternating current-dielectrophoresis microfluidic chip to separate latex particles and WBCs based on their difference in electrical property. Han et al.22 developed three-stage cascade paramagnetic capture (PMC) mode magnetophoretic micro-separators to separate WBCs and RBCs at a volumetric flow rate of 5 μl h−1. Although these active techniques offer high extraction efficiency and high sample purity, the processing throughput is very low. In addition, the complex sample pre-preparation operations and complicated external field generators are still indispensable, which significantly hampers the wide application of these devices.23

Meanwhile, passive devices through using special sieving structures to filter the WBCs from human blood have been reported. For example, Kumar et al.24 developed a three-plate sieving structure based on the cross-flow filtration technique to separate WBCs, RBCs, and platelets from whole blood. These devices have the advantages of simple structures and operations, but their sample processing throughput and recovery efficiency of WBCs are relatively low. In recent years, inertial microfluidics has achieved the separation of particles/cells by utilizing size-dependent hydrodynamic effects. As a sized-based passive cell separation method, inertial microfluidics has been considered as a promising method for realizing the WBC extraction owing to the advantages of high processing throughput, label- and external field-free operations and simple channel structures.25–27 For example, Nivedita et al.28 employed an Archimedean spiral channel with one inlet and three outlets to separate WBCs from the diluted blood (500-fold dilution) at a throughput of 1800 μl min−1 (3.6 μl min−1 for whole blood). Wu et al.17 reported a spiral channel with a trapezoidal cross-section to separate the subpopulation of WBCs (i.e., polymorphonuclear leukocytes and mononuclear leukocytes) from the diluted whole blood (1–2% hematocrit) with an efficiency of 80% and at a throughput of 10 μl whole blood per min. Wu et al.29 developed a straight structure with a contraction–expansion array for the separation of WBCs from the diluted blood (400-fold dilution) with a throughput of 150 μl min−1. After the massive parallelization of this device, the throughput could reach 10[thin space (1/6-em)]800 μl min−1 (only 27 μl min−1 for whole blood). Zhang et al.30 proposed a serpentine channel for extracting WBCs from the diluted blood (20-fold dilution) at a processing throughput of 600 μl min−1 (30 μl min−1 for whole blood). However, the purity of collected WBCs is only 48% after the two consecutive processes. By utilizing the viscoelastic focusing and separation, Tan et al.10 introduced a two-stage straight channel for separating of WBCs from the diluted blood (∼4500-fold dilution) under viscoelastic flow at a throughput of 150 μl min−1. However, the extraction efficiency and the purity of WBCs are relatively low by sized-based passive cell separation methods owing to the heterogeneity of WBCs.31 In addition, high degree of blood pre-dilution through manual operation is still required, which decreases the real throughput for the processing of whole blood and the automation of these methods.

Herein, we develop a novel inertial microfluidic (IM) cube through integrating lysis, storage and extraction modules for the automatic and fast extraction of WBCs from whole blood. The lysis module is employed to mix whole blood and lysing buffer automatically and quickly for the preliminary lysis of RBCs. The storage module is designed to enhance the lysis effect of RBCs. Then, the extraction module integrated with eight extraction units can simultaneously remove the RBC debris and extract WBCs at a high throughput. After the conceptual design, we characterize the performances of the lysis and extraction modules, respectively. Finally, the IM cube is applied for directly extracting WBCs from whole blood, and the performances of our IM cube are compared with those of the centrifugation method. Our IM cube is based on passive secondary-flow mixing and inertial sorting, made of cheap materials and fabricated by well-established technologies. We envision that our IM cube can be used as an automatic and fast method for extracting WBCs from whole blood.

Materials and methods

Conceptual design

Fig. 1a and S1 show the photograph of our IM cube, which consists of three inlets (1, 2 and 3) and two outlets (4 and 5) for the automatic and fast extraction of WBCs from whole blood. When utilizing our cube, the whole blood and lysing buffer were infused into our cube through inlet 1 and inlet 2, respectively. After running for about five minutes, a PBS solution was pumped into the cube via inlet 3. As a result, WBC and lysed RBC samples could be collected via the two outlets (4 and 5), respectively. To achieve the above function, our IM cube is integrated with the lysis module, storage module, extraction module and two world-to-chip covers, as illustrated in Fig. 1b. The detailed microfluidic structures of the lysis and extraction modules are illustrated in Fig. 1c and d.
image file: c9lc00942f-f1.tif
Fig. 1 (a) Photograph of the fabricated IM cube. The whole blood was pumped via inlet 1, the lysing buffer was pumped via inlet 2, and PBS was pumped via inlet 3. The WBC and lysed RBC samples could be collected from the two outlets (4 and 5) separately. (b) Exploded view of our IM cube, which is integrated with three different functional modules and two world-to-chip covers. (c and d) Schematic diagrams illustrating the microfluidic structures of the lysis module (c) and the extraction module (d).

In our IM cube, the whole blood and lysing buffer were firstly flowed into the lysis module (see Fig. 1c) after being pumped through inlet 1 and inlet 2, and the lysis module was designed to achieve the fast and automatic mixing of whole blood and lysing buffer for the preliminary lysis of RBCs. To enhance the mixing effect, we employed a 3D structure channel (see Fig. 2a), which was fabricated through adhering two channel layers (the upper lysis layer and the lower lysis layer) with a double-sided tape layer (see Fig. S2a). In the upper lysis layer (see layer 1 in Fig. S2b), a series of planar split-and-recombine (P-SAR) structures was designed. Through using the P-SAR structures, the flow was firstly split into four streams and subsequently recombined into one, yielding exponential decreases in diffusion time scales between two mixing liquids.32 In addition, the sustaining transverse secondary flow and the multi-vortex flow were proven to be meaningful for increasing the interfacial area between the two co-flowing liquids and thus, enhancing the mixing effect.33–35 To achieve sustained transverse secondary flow, the lower lysis layer (see layer 3 in Fig. S2b) contains five concave–convex structure segments, which were laid along a circle. On the basis of this channel design, the transverse secondary flow and the multi-vortex flow can be generated to accelerate the interspecies transport of the whole blood and the lysing buffer.36,37 After stacking these lysis layers, the injected whole blood and lysing buffer will alternately flow in the upper and lower lysis layers to mix these two liquids well. The flow path in this module is illustrated in Fig. S3. To further enhance the mixing effect of whole blood and lysing buffer, the connecting portions between the upper and lower lysis layers were also composed of concave–convex structures, which were formed by stacking a cylinder between two sectors (see Fig. S4).


image file: c9lc00942f-f2.tif
Fig. 2 (a) Schematic structure of the lysis module. (b) Principle of extracting WBCs using the inertial separation in the extraction module. (c and d) Photographs of the lysis (c) and extraction (d) modules.

To ensure the complete lysis of RBCs, the storage module (Fig. S5a–c) was designed. The mixed liquid of whole blood and lysing buffer takes about 5 min to flow from the bottom to the top of the chamber in the storage module, as illustrated in Fig. S5d. The storage module can be treated as a large and long channel, and the lysis of RBCs could be completed when it flows out from the top of the chamber. As the mixed liquid in the chamber keeps flowing, the cells will not settle in the storage module. The residual mixed liquid in the storage module can be supplanted with PBS.

After the complete lysis of RBCs with the help of the above two modules, the WBCs need to be extracted from the mixed liquid containing the lysing buffer and the RBC debris in the extraction module (see Fig. 1d). To realize this function, spiral inertial microfluidics was employed in the extraction module. In spiral inertial microfluidic channels, the flowing WBCs will simultaneously experience the inertial lift force (FL) caused by the inertial flow and the Dean drag force (FD) caused by the Dean flow in curving channels.28 The inertial lift force (FL) is a combination of shear-induced and wall-induced inertial lift forces. The shear-induced inertial lift force generated by the parabolic velocity profile of the Poiseuille flow pushes the cells to migrate toward the channel walls, while the wall-induced inertial lift force repels the cells away from the channel walls.38 The radial pressure gradient of fluid flow in curving channels generates cross-sectional Dean flow, which exerts a lateral Dean drag force (FD) on cells.39 The superposition of FL/FD determines the focusing modes and the lateral positions of cells. The value of this force ratio mainly depends on the particle diameter and the channel height.40 To realize the focusing of specific sized cells, the confinement ratio (CR) needs to satisfy the following criterion (CR = ap/H ≥ 0.07, where ap is the particle diameter and H is the channel height). It is reported that the cells could focus at a position of about 20% width from the inner wall when CR = ∼0.1.41,42 According to the above theory, the height H of the spiral channels in the extraction module was determined to be 100 μm to ensure the successful focusing of WBCs (with a typical average size of 10 μm). Meanwhile, the RBC debris (CR = ap/H ≪ 0.07) experiences the dominant Dean drag force (FD) and will circulate along with the liquid infused from the outer inlet.43,44 When arriving at the outlet, the large WBCs will enter into the sheath flow infused from the inner inlet and equilibrate near the inner wall, whereas the lysing buffer and RBC debris infused from the outer inlet will recirculate to the outer wall again (see Fig. 2b). To increase the processing throughput of the extraction module, eight spiral channels were integrated in the extraction module, which was fabricated through stacking seven layers in the vertical direction (see Fig. S6a). Specifically, two extraction layers were stacked and each extraction layer contains four spiral channels (see layers 3 and 5 in Fig. S6b and c). The lower and upper flow guide layers (see layers 1 and 7 in Fig. S6b) were also stacked to achieve the average distribution of the driving flow rates for the eight units or converge the collected liquids. These layers mentioned above were adhered using the double-sided tape layers (see layers 2, 4 and 6 in Fig. S6b). The flow path in this module is illustrated in Fig. S7.

Device fabrication and assembly

The lysis layers (layers 1 and 3 in Fig. S2) in the lysis module, the flow guide layers (layers 1 and 7 in Fig. S6) and extraction layers (layers 3 and 5 in Fig. S6) in the extraction module have a three-sheet structure (Fig. S8). All these layers were fabricated using the UV laser cutting and oxygen plasma bonding techniques. Briefly, silicon films were used to fabricate the lysis channels (with a thickness of 180 μm), flow guide channels (with a thickness of 180 μm) and extraction channels (with a thickness of 100 μm). Polymer films with a thickness of 130 μm were used as the upper and lower covers and were bonded with the patterned silicon films via the oxygen plasma treatment (PDC-002, Harrick Plasma) to seal the through channel patterns. Double-sided tape layers with a thickness of 250 μm (layer 2 in Fig. S2; layers 2, 4 and 6 in Fig. S6) were patterned with through-holes for adhering the different channel layers. All the patterns on these layers were cut using a UV laser machine (TH-UV2000A, Tianhong Laser, China). The chamber and chamber cover of the storage module (Fig. S5a) were made of a photosensitive resin material using 3D printing, and were bonded by polyurethane potting (711, Hasuncast). The upper and lower covers (Fig. S9) of the IM cube were machined from acrylic sheets (5 mm × 51 mm × 53 mm), metal tubes were bonded to its side holes with an AB glue. The two covers served as the word-to-chip interfaces for our IM cube.

Fig. S10 shows the assembly processes of the lysis and extraction modules after fabricating each layer. Four locating holes were fabricated at the four corners of each layer, and a custom fixture was employed to realize the quick assembly of the lysis and extraction modules. Fig. 2c and d show the photographs of the lysis and extraction modules after assembly. Fig. S11 shows the assembly process of our IM cube after fabricating and assembling each module. Our IM cube was assembled within three main steps by respectively stacking different modules and parts. Studs and nuts were used to assemble all the modules in our IM cube. Rubber O-rings were used to seal the different modules. After assembling, our device appears like a cube with a small footprint (53 mm × 25 mm × 51 mm). Our IM cube can bear an ultrahigh pressure without leakage and is fabricated with cheap materials and using well-established technologies.

Sample preparation

To characterize the fluid mixing performances of the lysis module, rhodamine-B (1 μg ml−1, Maikun Chemical) and fluorescein (1 μg ml−1, Maikun Chemical) powders were separately dissolved in a magnetic-activated cell sorting (MACS) buffer for preparing the rhodamine-B and fluorescein fluorescent solutions. The MACS buffer consists of 1× PBS (Sigma-Aldrich) and 2 mM ethylenediaminetetraacetic acid (EDTA) supplemented with 0.5% bovine serum albumin (BSA) (Miltenyi Biotec). The BSA was used to prevent the nonspecific adhesion of the rhodamine-B and fluorescein fluorescent molecules to the channel walls.

Two types of polystyrene particles with diameters of 4 μm and 10 μm (Thermo Scientific, Inc) were respectively used to mimic the lysed RBC debris and WBCs for optimizing the structural and operational parameters in the extraction module. The particle solutions of 4 μm and 10 μm with an initial solids content of 1% were mixed together and then diluted with deionized (DI) water to form the uniform particle mixture suspension with a specific concentration. For operating the spiral inertial microfluidics in the extraction module, PBS (Sigma-Aldrich) acted as the sheath fluid.

Human whole blood was acquired from healthy volunteers through a Vacutainer collection tube (BD Biosciences) containing anticoagulant K2EDTA. This study was approved by the institutional committee of the Institutional Ethical Committee (IEC) for Clinical Research of Zhongda Hospital (Southeast University) and informed consent was obtained from the volunteers. All experiments were performed in compliance with the Chinese laws and following the institutional guidelines. An ammonium-chloride-potassium (ACK) lysing buffer (Thermo Fisher Scientific) was employed to lyse the RBCs. As a control experiment, the conventional method for extracting the WBCs using the multi-step centrifugation was also performed. Briefly, the whole blood and the ACK lysing buffer were first manually mixed and then incubated for 5 min to selectively lyse the RBCs in the whole blood. The lysed liquid was centrifuged for 5 min at a speed of 300 × g, then the supernatant was carefully aspirated, and finally, the settled WBCs were re-suspended in the sterile PBS. To greatly remove the lysing buffer and RBC debris, the above steps were repeated.

To stain the WBCs collected by our IM cube, the cells attached on the adhesive glass slides were fixed with 4% paraformaldehyde in PBS (Sigma-Aldrich), permeabilized using 0.05% Triton X-100 (PBST) solution, and subsequently blocked with 2% BSA and 3% normal goat serum in PBS. Then, the samples were immunofluorescently stained with mouse anti-CD45 antibody (ab8216, Abcam) and fluorescein isothiocyanate (FITC)-labeled anti-mouse Ig antibody (ab6785, Abcam). Finally, the samples were mounted with antifade mounting medium with DAPI (Vector Laboratories).

The WBC samples collected by the IM cube and centrifugation method were also processed using a flow cytometer (BD Biosciences). Before the WBCs were analyzed by flow cytometry, the samples were added with propidium iodide (PI) (BD Biosciences) and incubated for 3–5 min.

Experimental setup

For the characterization of the performances, the lysis and extraction modules were respectively clamped with two transparent PMMA plates with inlets and outlets. The stable and precise infusion of the samples or PBS solutions was realized using two syringe pumps (Legato 100, KD Scientific). The connection between the PMMA plates and syringes or collecting tubes was realized via Teflon tubes.

The lysis module was mounted onto the platform of an upright fluorescence microscope (80i, Nikon) equipped with a color camera (DS-Ri1, Nikon) to capture the mixing status of the fluorescent solutions under the illumination of fluorescent light with the excitation wavelengths of 465–495 nm and 510–560 nm, respectively. The captured image frames were merged using the ImageJ software (NIH) to create the composite pictures illustrating the mixing effects in the lysis module. The intensity profiles of the composite pictures were also measured and normalized for quantitatively evaluating the mixing performances.

In the experiments of extraction performance characterization, the images of particle/cell distributions near the outlet of the spiral channel were continuously captured using an inverted microscope (IX 71, Olympus) equipped with a CCD camera (Exi Blue, Qimaging). The discrete image frames captured over a certain time period were stacked using the ImageJ software (NIH) to create the composite images illustrating the focusing performances of differently-sized particles or cells. To quantitatively evaluate the extraction performances, the concentrations of WBCs in the initial sample and samples collected from each outlet were counted using a cell counting plate (Qiujing) several times.

For the operation of our IM cube, three syringe pumps (Legato 100, KD Scientific) were used to provide the required driving flow rates. The collected WBCs were sampled under a microscope (IX 71, Olympus) and the concentrations of WBCs were counted using a Countess II FL automated cell counter (Thermo Fisher Scientific).

Results and discussion

Characterization of the lysis module

We first characterized the mixing efficiency (ME) of the lysis module. The rhodamine-B and fluorescein solutions were used to determine the ME of the lysis module at a total flow rate of 2400 μl min−1. Fig. 3a shows the mixing performances at the inlet, outlet and nine positions marked in Fig. 2a. At the inlet, a distinct boundary was clearly observed between the two liquids (red and green). The images from position 1 to 7 show that the red and green striations of the two liquids are observed, then gradually disappeared along the flow direction. The images from positions 8 and 9 and outlet show a uniform yellow-green color, and thus, we intuitively considered that the two solutions achieve the complete mixing in the lysis module. To further prove the above conclusion, the intensity profiles of rhodamine-B (red) and fluorescein (green) solutions in the lysis module were measured and normalized, as illustrated in Fig. S12. The empirical mixing efficiency (ME) was acquired as follows: ME = (ARψR + AFψF)/(AR + ψR), where AR and AF respectively represent the area under the normalized intensities of rhodamine-B and fluorescein. ψR and ψF are, respectively, the percentages of the mixed areas of rhodamine-B and fluorescein, and can be calculated as ψ = mixed area/total area × 100%.45 Then, the MEs at these positions were calculated and the results are illustrated in Fig. 3b. It was found that the ME sharply increases from the inlet to position 2. The increase rate of ME slightly decreased from position 3 to the outlet. Ultimately, a ME of 94.2% was achieved at the outlet, which is comparable to or even higher than that of other 3D mixers (most of them offer a ME of >90%).46,47 Then, the whole blood and lysing buffer were infused into the lysis module through the inlet at the flow rates of 160 μl min−1 and 2240 μl min−1, respectively. As shown in Fig. 3c, the mixing status of the whole blood and lysis buffer at different positions are similar to the above observations. The images captured at the positions 8 and 9 and outlet illustrate a uniform dark-grey color in the lysis channel, which indicates a good mixing of the whole blood and lysis buffer.
image file: c9lc00942f-f3.tif
Fig. 3 (a) Images illustrating the mixing effects of rhodamine-B (red) and fluorescein (green) solutions at different positions. (b) Calculated mixing efficiency (ME) at different positions in the lysis module. (c) Images illustrating the mixing effects of the whole blood and lysis buffer.

Characterization of the extraction module

We next characterized the extraction performances of the extraction module to optimize the channel structure and operational flow rate. Seven types of spiral channels with different channel widths of 440–680 μm (with an interval of 40 μm) were tested. The suspension of 4 μm and 10 μm mixed particles was pumped into the outer inlet of the spiral channel at a flow rate of 300 μl min−1 while the sheath was pumped into inner inlet at the flow rates of 700–4000 μl min−1. Fig. 4a shows the distributions of these two particles at the outlet of the spiral channel with a width of 640 μm. From these distribution maps, it was observed that the 10 μm particles gradually focus into a particle string near the inner wall when the flow rate increases from 1900 μl min−1 to 2500 μl min−1, and the focusing string of the 10 μm particles remains relatively stable at flow rates higher than 2500 μl min−1. In addition, the 4 μm particles display as a broad band which gradually shifts towards the outer channel wall when the flow rate increases from 1900 μl min−1 to 3100 μl min−1. By further increasing the flow rate, the band of the 4 μm particles returns to the inner wall. To provide a clear picture of the particle extraction performances, the distributions of the 4 and 10 μm particles at various flow rates were measured, as illustrated in Fig. 4b. It was found that the extraction of the 10 μm particles can be realized when the flow rate increases from 2700 μl min−1 to 3700 μl min−1. In this flow rate range, the position of the 10 μm particle string is above the bifurcation point of the two outlets while the 4 μm particle band is well below this point. The distance between the 4 μm and 10 μm particles gradually increases when the flow rate increases from 2700 μl min−1 to 3100 μl min−1, and decreases by further increasing the flow rate. Therefore, the optimal flow rate for achieving the best extraction performance was determined to be 3100 μl min−1. The distributions of the 4 μm and 10 μm particles in channels with other widths can be found in Fig. S13. It was found that the other channels can also realize the separation of these two particles at specific flow rates. However, the 640 μm wide channel has a high processing throughput, a wide operational flow rate range and a better stability. To further validate the WBC extraction performances in this channel, the lysed blood was infused into the outer inlet at a flow rate of 300 μl min−1 while PBS was infused into the inner inlet at a flow rate of 2800 μl min−1. As shown in Fig. 4c, the sample collected from the inner outlet was colorless, in contrast with a red solution of the lysed RBCs collected from the outer outlet. The cell distribution near the channel outlet also indicates that the WBCs focus near the inner wall and can be collected from the inner outlet ultimately. To quantitatively evaluate the extraction performances, the extraction efficiency (EE = ncollect/ntotal) of the 10 μm particles and WBCs were calculated. Here, ncollect is the number of WBCs or 10 μm particles collected from the inner outlet and ntotal is the number of WBCs or 10 μm particles in the initial sample. Fig. 4d shows the calculated EEs of the 10 μm particles and WBCs at the optimal flow rate. It was found that an EE of 94.9% was achieved for the 10 μm particles. The EE for the WBCs is slightly worse (EE = 88.1%) probably due to the heterogeneity of WBCs.
image file: c9lc00942f-f4.tif
Fig. 4 (a) Images illustrating the distributions of 10 μm and 4 μm particles at different flow rates. (b) The lateral positions of 10 μm and 4 μm particles in the extraction channel of 640 μm width at various flow rates. (c) The focusing image of WBCs at the outlet of the extraction channel and the photograph of the samples collected from the inner outlet and outer outlet at the optimum flow rate. (d) Extraction efficiency of 10 μm particles and WBCs at the optimum flow rate.

Application

Finally, we applied our IM cube for the automatic and fast extraction of WBCs from whole blood. The whole blood and lysing buffer were infused into our IM cube at the flow rates of 160 μl min−1 and 2240 μl min−1 (a flow ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]14), respectively. Five minutes later, PBS was infused into our IM cube at a flow rate of 22.4 ml min−1. Fig. 5a shows the WBC extraction process and the photograph of the collected WBC sample. It was observed that the collected WBC sample appears colorless, which indicates that an overwhelming majority of RBCs has been removed after being processed with our IM cube. In the process of extracting WBCs in the extraction module, it is inevitable that a very small amount of the lysed background liquid may diffuse into the target WBC sample due to the diffusion effect. As the lysed RBCs will release hemoglobin into the lysing buffer, we used UV spectroscopy to measure the hemoglobin concentration in the collected WBC sample. Based on the Cripps method,48,49 the hemoglobin concentration in a sample is proportional to 2A(576) − A(560) − A(592) (Cripps' sum), where A(560), A(576) and A(592) are, respectively, the absorbances measured at 560, 576 and 592 nm. A reference sample was made by mixing the lysing buffer and whole blood at a ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]14 (corresponds to the ratio of the flow rate) and incubated at room temperature for 5 min to lyse all the RBCs and release the hemoglobin into the lysing buffer. Successive dilution of the reference solution with PBS was performed and Cripps' sum of these obtained solutions was also measured to establish a calibration curve between hemoglobin concentration and Cripps' sum (see Fig. S14). Then, the relative hemoglobin concentration of the collected WBC sample could be measured. It was found that the WBC sample extracted by our IM cube has a relative hemoglobin concentration of 1.8% of the reference solution, which indicates that PBS accounts for 98.2% of the total liquid (see Fig. 5b). To further validate the extraction performances of our IM cube, the samples collected from the two outlets were analyzed. Fig. S15 shows the photograph and microscopic images of the samples collected from the two outlets of our IM cube. It was clearly observed that the background RBCs were all lysed and only a very few WBCs was observed in the waste. Fig. 5c shows the extraction efficiency (EE) of our IM cube. An average EE of 83.9% can be achieved using our IM cube. Cell viability test was also performed to determine whether the processing of our IM cube will induce the death of WBCs. The trypan blue exclusion test shows that a cell viability rate (VR) of 96.6% for the collected WBCs was maintained after being processed by our IM cube. In addition, the immunofluorescence staining result of WBCs using CD45 and DAPI and the result of flow cytometry indicate that our IM cube has a negligible effect on the viability and integrity of WBCs (see Fig. S16 and S17). The WBC sample collected by our IM cube has a concentration of about 2.3 × 105 ml−1, which is sufficient for many downstream assays. For example, flow cytometry usually requires a cell concentration of over 1 × 105 ml−1 and a sample volume of 0.5–2 ml. To obtain samples with a higher concentration, the WBC sample collected by our IM cube was further concentrated using our previously-developed inertial microfluidic concentrator.50 As can be seen from Fig. S18, the original sample with a volume of 12.5 ml could be easily reduced to only 1.3 ml. To further increase the throughput of our device, two IM cubes were paralleled and each cube was also stacked with 16 spiral extraction units (see Fig. S19; Video S1 in the ESI), and a throughput of 720 μl min−1 for whole blood could be achieved. Therefore, our IM cube is capable of extracting high-purity WBCs from whole blood automatically, quickly, directly and harmlessly.
image file: c9lc00942f-f5.tif
Fig. 5 Characterization of the IM cube. (a) The WBC sample extracted by our IM cube. (b) The PBS proportion of the WBC sample collected by our IM cube and centrifugation (one-step). (c) The extraction efficiency for WBCs acquired by our IM cube and the repeated centrifugation (two-step).

As a control experiment, the WBC extraction performances using the centrifugation method were also characterized (Fig. 5b and c; S17). It was observed that PBS accounts for 97.2% of the WBC sample liquid collected after the one-step centrifugation, which is slightly lower than that (98.2%) using our IM cube. An EE of 85.6% is achieved after two-step centrifugation, which is slightly higher than that (83.9%) using our IM cube. However, the uncertainty (the error of EE) of the repeated centrifugation is much larger than that of our IM cube due to the manual operation. In addition, the results of flow cytometry (see Fig. S17) indicate that the viability of the samples after the two-step centrifugation is lower than that of the samples processed with our IM cube. The deterioration of cell viability indicates that a high centrifugal force caused by the centrifugation may induce the death of cells. The above comparison results demonstrate that the performances of our IM cube are comparable to those of the centrifugation and it does better in some aspects. Therefore, our IM cube can potentially act as a small footprint substitute for the automatic and fast extraction of WBCs from whole blood.

Conclusion

In this work, a novel IM cube integrated with lysis, storage and extraction modules is designed and fabricated, allowing for the simultaneous depletion of RBC debris and extraction of WBCs. These three modules are respectively used to completely mix whole blood and lysing buffer, thoroughly lyse RBCs, and automatically extract WBCs from the lysed background RBCs. After the conceptual design of our device, we respectively characterize the performances of the lysis and extraction modules. The results show that a mixing efficiency of 94.2% is achieved while extraction efficiencies of 94.9% and 88.1% are achieved for the extraction of 10 μm particles and WBCs. Finally, we apply the IM cube for the extraction of WBCs from human whole blood at a high throughput. It is found that an extraction efficiency of 83.9% for WBCs from whole blood and a cell viability of 96.6% can be achieved by using our IM cube. The performances of our IM cube are comparable to those using the centrifugation method and better in some aspects. Our IM cube is based on passive secondary-flow mixing and inertial sorting, offers the advantages of small footprint, high stability, simple fabrication and cheap materials, and is promising to be applied as an alternative method for extracting WBCs from whole blood.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research work is supported by the National Natural Science Foundation of China (81727801, 51875103 and 51775111), the Natural Science Foundation of Jiangsu Province (BK20190064), the Six Talent Peaks Project of Jiangsu Province (SWYY-005) and the Zhishan Youth Scholar Program of SEU.

References

  1. J. W. G. Yarnell, I. A. Baker, P. M. Sweetnam, D. Bainton, J. R. Obrien, P. J. Whitehead and P. C. Elwood, Circulation, 1991, 83, 836–844 CrossRef CAS PubMed .
  2. W. B. Kannel, K. Anderson and P. W. F. Wilson, JAMA, J. Am. Med. Assoc., 1992, 267, 1253–1256 CrossRef CAS PubMed .
  3. X. H. Cheng, D. Irimia, M. Dixon, K. Sekine, U. Demirci, L. Zamir, R. G. Tompkins, W. Rodriguez and M. Toner, Lab Chip, 2007, 7, 170–178 RSC .
  4. M. P. Weijenberg, E. J. M. Feskens and D. Kromhout, Arterioscler., Thromb., Vasc. Biol., 1996, 16, 499–503 CrossRef CAS PubMed .
  5. M. B. Terry, L. Delgado-Cruzata, N. Vin-Raviv, H. C. Wu and R. M. Santella, Epigenetics, 2011, 6, 828–837 CrossRef CAS PubMed .
  6. X. J. Han, C. van Berkel, J. Gwyer, L. Capretto and H. Morgan, Anal. Chem., 2012, 84, 1070–1075 CrossRef CAS PubMed .
  7. M. Hosokawa, M. Asami, S. Nakamura, T. Yoshino, N. Tsujimura, M. Takahashi, S. Nakasono, T. Tanaka and T. Matsunaga, Biotechnol. Bioeng., 2012, 109, 2017–2024 CrossRef CAS PubMed .
  8. R. S. Vasan, Circulation, 2006, 113, 2335–2362 CrossRef PubMed .
  9. C. J. Vanoss, P. M. Bronson, E. A. Dinolfo and K. C. Chadha, Immunol. Commun., 1981, 10, 549–555 CrossRef CAS PubMed .
  10. J. K. S. Tan, S. Y. Park, H. L. Leo and S. Kim, IEEE Trans. Biomed. Circuits Syst., 2017, 11, 1431–1437 Search PubMed .
  11. S. Fukuda and G. W. Schmid-Schonbein, J. Leukocyte Biol., 2002, 72, 133–139 CAS .
  12. K. S. Lang, P. A. Lang, C. Bauer, C. Duranton, T. Wieder, S. M. Huber and F. Lang, Cell. Physiol. Biochem., 2005, 15, 195–202 CrossRef CAS PubMed .
  13. Y. Xie, D. N. Zheng, Q. F. Li, Y. Chen, H. Lei and L. L. Q. Pu, J. Plast. Reconstr. Aesthet. Surg., 2010, 63, 482–487 CrossRef PubMed .
  14. L. Vettore, M. C. Dematteis and P. Zampini, Am. J. Hematol., 1980, 8, 291–297 CrossRef CAS PubMed .
  15. H. N. Chang, W. G. Lee and B. S. Kim, Biotechnol. Bioeng., 1993, 41, 677–681 CrossRef CAS PubMed .
  16. J. F. Tait, C. Smith and B. L. Wood, Blood Cells, Mol., Dis., 1999, 25, 271–278 CrossRef CAS PubMed .
  17. L. D. Wu, G. F. Guan, H. W. Hou, A. A. S. Bhagat and J. Han, Anal. Chem., 2012, 84, 9324–9331 CrossRef CAS PubMed .
  18. E. K. Sackmann, A. L. Fulton and D. J. Beebe, Nature, 2014, 507, 181–189 CrossRef CAS PubMed .
  19. J. Sibbitts, K. A. Sellens, S. Jia, S. A. Klasner and C. T. Culbertson, Anal. Chem., 2018, 90, 65–85 CrossRef CAS PubMed .
  20. D. Kim, M. Sonker and A. Ros, Anal. Chem., 2019, 91, 277–295 CrossRef CAS PubMed .
  21. B. Cetin and D. Q. Li, Electrophoresis, 2010, 31, 3035–3043 CrossRef CAS PubMed .
  22. K. H. Han and A. B. Frazier, Lab Chip, 2006, 6, 265–273 RSC .
  23. C. Iliescu, G. L. Xu, E. Barbarini, M. Avram and A. Avram, Microsyst. Technol., 2009, 15, 1157–1162 CrossRef CAS .
  24. S. P. Kumar, T. Ramesh and R. Aravind, Biomed. Res., 2017, 28, 4181–4186 CAS .
  25. M. Toner and D. Irimia, Annu. Rev. Biomed. Eng., 2005, 7, 77–103 CrossRef CAS PubMed .
  26. H. W. Hou, A. A. S. Bhagat, W. C. Lee, S. Huang, J. Han and C. T. Lim, Micromachines, 2011, 2, 319–343 CrossRef .
  27. D. R. Gossett, H. T. K. Tse, J. S. Dudani, K. Goda, T. A. Woods, S. W. Graves and D. Di Carlo, Small, 2012, 8, 2757–2764 CrossRef CAS PubMed .
  28. N. Nivedita and I. Papautsky, Biomicrofluidics, 2013, 7, 054101 CrossRef PubMed .
  29. Z. L. Wu, Y. Chen, M. R. Wang and A. J. Chung, Lab Chip, 2016, 16, 532–542 RSC .
  30. J. Zhang, D. Yuan, R. Sluyter, S. Yan, Q. B. Zhao, H. M. Xia, S. H. Tan, N. T. Nguyen and W. H. Li, IEEE Trans. Biomed. Circuits Syst., 2017, 11, 1422–1430 Search PubMed .
  31. H. Ramachandraiah, H. A. Svahn and A. Russom, RSC Adv., 2017, 7, 29505–29514 RSC .
  32. A. P. Sudarsan and V. M. Ugaz, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 7228–7233 CrossRef CAS PubMed .
  33. P. B. Howell, D. R. Mott, J. P. Golden and F. S. Ligler, Lab Chip, 2004, 4, 663–669 RSC .
  34. S. P. Vanka, G. Luo and C. M. Winkler, AIChE J., 2004, 50, 2359–2368 CrossRef CAS .
  35. A. P. Sudarsan and V. M. Ugaz, Lab Chip, 2006, 6, 74–82 RSC .
  36. T. J. Johnson, D. Ross and L. E. Locascio, Anal. Chem., 2002, 74, 45–51 CrossRef CAS PubMed .
  37. P. B. Howell, D. R. Mott, S. Fertig, C. R. Kaplan, J. P. Golden, E. S. Oran and F. S. Ligler, Lab Chip, 2005, 5, 524–530 RSC .
  38. D. Di Carlo, D. Irimia, R. G. Tompkins and M. Toner, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 18892–18897 CrossRef CAS PubMed .
  39. H. W. Hou, M. E. Warkiani, B. L. Khoo, Z. R. Li, R. A. Soo, D. S. W. Tan, W. T. Lim, J. Han, A. A. S. Bhagat and C. T. Lim, Sci. Rep., 2013, 3, 1259 CrossRef PubMed .
  40. A. A. S. Bhagat, S. S. Kuntaegowdanahalli and I. Papautsky, Lab Chip, 2008, 8, 1906–1914 RSC .
  41. H. M. Tay, D. C. Yeo, C. Wiraja, C. J. Xu and H. W. Hou, J. Visualized Exp., 2016, 113, e54327 Search PubMed .
  42. H. M. Tay, S. Kharel, R. Dalan, Z. J. Chen, K. K. Tan, B. O. Boehm, S. C. J. Loo and H. W. Hou, NPG Asia Mater., 2017, 9, e434 CrossRef .
  43. H. W. Hou, R. P. Bhattacharyya, D. T. Hung and J. Han, Lab Chip, 2015, 15, 2297–2307 RSC .
  44. D. C. Yeo, C. Wiraja, Y. Y. Zhou, H. M. Tay, C. J. Xu and H. W. Hou, ACS Appl. Mater. Interfaces, 2015, 7, 20855–20864 CrossRef CAS PubMed .
  45. M. Rafeie, M. Welleweerd, A. Hassanzadeh-Barforoushi, M. Asadnia, W. Olthuis and M. E. Warkiani, Biomicrofluidics, 2017, 11, 014108 CrossRef PubMed .
  46. T. W. Lim, Y. Son, Y. J. Jeong, D. Y. Yang, H. J. Kong, K. S. Lee and D. P. Kim, Lab Chip, 2011, 11, 100–103 RSC .
  47. J. Yang, L. Qi, Y. Chen and H. M. Ma, Chin. J. Chem., 2013, 31, 209–214 CrossRef CAS .
  48. C. M. Cripps, J. Clin. Pathol., 1968, 21, 110–112 CrossRef CAS PubMed .
  49. E. Sollier, M. Cubizolles, Y. Fouillet and J. L. Achard, Biomed. Microdevices, 2010, 12, 485–497 CrossRef PubMed .
  50. N. Xiang, X. Shi, Y. Han, Z. G. Shi, F. T. Jiang and Z. H. Ni, Anal. Chem., 2018, 90, 9515–9522 CrossRef CAS PubMed .

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9lc00942f

This journal is © The Royal Society of Chemistry 2020