Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
micromachines Article Red Blood Cell Responses during a Long-Standing Load in a Microfluidic Constriction Mitsuhiro Horade, Chia-Hung Dylan Tsai *, Hiroaki Ito and Makoto Kaneko Department of Mechanical Engineering, Osaka University, Suita 565-0871, Japan; [email protected] (M.H.); [email protected] (H.I.); [email protected] (M.K.) * Correspondence: [email protected]; Tel.: +81-6-6879-7333 Academic Editors: Kwang W. Oh and Nam-Trung Nguyen Received: 25 January 2017; Accepted: 22 March 2017; Published: 26 March 2017 Abstract: Red blood cell responses during a long-standing load were experimentally investigated. With a high-speed camera and a high-speed actuator, we were able to manipulate cells staying inside a microfluidic constriction, and each cell was compressed due to the geometric constraints. During the load inside the constriction, the color of the cells was found to gradually darken, while the cell lengths became shorter and shorter. According to the analysis results of a 5 min load, the average increase of the cell darkness was 60.9 in 8-bit color resolution, and the average shrinkage of the cell length was 15% of the initial length. The same tendency was consistently observed from cell to cell. A correlation between the changes of the color and the length were established based on the experimental results. The changes are believed partially due to the viscoelastic properties of the cells that the cells’ configurations change with time for adapting to the confined space inside the constriction. Keywords: microfluidics; long-standing load; cell manipulation; red blood cell 1. Introduction There are established relationships between the deformability of red blood cells (RBCs) and human diseases, such as Malaria, Spherocytosis and Cryohydrocytosis [1–5]. While many works have focused on RBC deformability through short-duration compressions, such as transit time through a constriction or micro-slit [6–11], few works have focused on RBC characteristics after long-standing loads [12,13]. RBC responses under a long-standing load reflect how a RBC changes while being plugged in microcirculation. For example, Baez et al. found different flow characteristics of oxygenated and partially deoxygenated red cells plugging in microcirculation [14]. Therefore, this work is motivated by the idea of investigating cell responses during a long-standing load using microfluidic constriction. Figure 1 shows an overview of the proposed idea and sample RBC photos at the beginning and the end of a long-standing load. A constriction channel is placed in a microfluidic device as illustrated in Figure 1a. There are two phases of the experiment. First, a target RBC is positioned in front of the narrow channel for recovering from prior deformation as the “Catching” phase. Afterward, the RBC is moved into the constriction channel for compression as the “Loading” phase. The cell manipulations are performed by a vision-feedback system where the real-time cell position is computed for proper actuator output for moving the cell to a target position. The RBC is maintained inside the constriction for a specified duration as the long-standing load, and a camera is set for observing the responses of the cell during the load. Micromachines 2017, 8, 100; doi:10.3390/mi8040100 www.mdpi.com/journal/micromachines Micromachines 2017, 8, 100 2 of 10 Figure 1b illustrates a RBC before and after the 5 min load with two sample images from the experiments. At the beginning of the load, the RBC is slender and almost transparent. After the 5 min Micromachines 2017, 8, 100 2 of 10 load, the color of the RBC is darkened, and the horizontal length is significantly shortened. A time-lapse sample video can be found in the Supplementary Materials Video S1. Quantitative evaluations of illustrates the amounta of change in and cell color length measured and images presented in the this Figure 1b RBC before after and the 5cell min load are with two sample from paper. The correlation and interpretation from mechanical aspect transparent. are discussed based the experiments. At the beginning of the load, the RBCthe is slender and almost After the on 5 min experimental results. load, the color of the RBC is darkened, and the horizontal length is significantly shortened. Figure 1.1. An An overview overview of of the the proposed proposed method method and and example example results. results. (a) (a) An An illustration illustration of of cell cell Figure manipulation for the long-standing load; (b) illustrations and experimental images of a red blood cell manipulation for the long-standing load; (b) illustrations and experimental images of a red blood cell (RBC) before and after a 5 min load. (RBC) before and after a 5 min load. In this paper, we particularly focus on the evolution of the RBC color and length during a longA time-lapse sample video can be found in the Supplementary Materials Video S1. Quantitative standing load. These newly found cell properties could provide new insights or a different evaluation evaluations of the amount of change in cell color and cell length are measured and presented in this index for cell evaluation. paper. The correlation and interpretation from the mechanical aspect are discussed based on the experimental results. 2. Related Works In this paper, we particularly focus on the evolution of the RBC color and length during a RBC deformability hasnewly been found be correlated certain diseases example, long-standing load. These foundtocell propertieswith could provide new [5,15–18]. insights orFor a different RBCs withindex reduced are found in obese patients and in those with diabetic foot diseases evaluation for deformability cell evaluation. [15,16]. Malaria is a well-known blood-related disease that reduces RBC deformability [17,18]. 2. Relatedmorphology Works Unusual of RBCs has been found from those of the compound heterozygous for hemoglobin S and hemoglobin C [19]. There are different approaches for investigating RBC RBC deformability has been found to be correlated with certain diseases [5,15–18]. For example, deformability [20,21]. For example, Tözeren et al. used micropipette for developing a constitutive RBCs with reduced deformability are found in obese patients and in those with diabetic foot equation for the RBC membrane [22]. Brandao et al. employed optical tweezers for measuring RBC diseases [15,16]. Malaria is a well-known blood-related disease that reduces RBC deformability [17,18]. elasticity [23]. Dulinska et al. applied an atomic force microscope (AFM) for measuring the stiffness Unusual morphology of RBCs has been found from those of the compound heterozygous for of erythrocytes [24]. Among different evaluation methods, microfluidic constrictions have become hemoglobin S and hemoglobin C [19]. There are different approaches for investigating RBC popular in the last decade because of their precise fabrication and accurate control [25–27]. For deformability [20,21]. For example, Tözeren et al. used micropipette for developing a constitutive example, Zheng et al. put RBCs through a constriction for high-throughput biophysical equation for the RBC membrane [22]. Brandao et al. employed optical tweezers for measuring RBC measurements [28]. Sakuma et al. evaluated the cell fatigue state by inducing continuous and elasticity [23]. Dulinska et al. applied an atomic force microscope (AFM) for measuring the stiffness repetitive deformation [29]. There are also works investigating RBC responses after long-standing of erythrocytes [24]. Among different evaluation methods, microfluidic constrictions have become loads. For example, Fischer discovered the shape memory of RBCs based on their recovery from the popular in the last decade because of their precise fabrication and accurate control [25–27]. For example, deformation by shear stress up to 4 h [12]. Markle et al. investigated the viscoelastic properties of the Zheng et al. put RBCs through a constriction for high-throughput biophysical measurements [28]. RBC membrane by characterizing their shape after constant-pressure aspiration up to 90 min [13]. Sakuma et al. evaluated the cell fatigue state by inducing continuous and repetitive deformation [29]. Through the microfluidic platform, it becomes possible to stably monitor cells’ behaviors and There are also works investigating RBC responses after long-standing loads. For example, Fischer responses during a fixed-displacement load. discovered the shape memory of RBCs based on their recovery from the deformation by shear stress up In our previous works, we succeeded in applying long-standing loads to RBCs by high-speed to 4 h [12]. Markle et al. investigated the viscoelastic properties of the RBC membrane by characterizing cell manipulation and discussed how the RBCs recover after the load [30,31]. Preliminary results of their shape after constant-pressure aspiration up to 90 min [13]. Through the microfluidic platform, it becomes possible to stably monitor cells’ behaviors and responses during a fixed-displacement load. Micromachines 2017, 8, 100 3 of 10 Micromachines 2017, 8, 100 3 of 10 In our previous works,changes we succeeded applying long-standing loads observed to RBCs by high-speed cell the RBC color and length duringinthe load were independently and presented at manipulation and discussed how the RBCs recover after the load [30,31]. Preliminary results of the recent conferences [32,33]. In this work, we firstly present a comprehensive analysis of cell changes RBC color length changes the loadawere independently andofpresented at recent during theand long-standing load.during Furthermore, correlation between observed the changes the RBC color and conferences [32,33]. In this work, we firstly present a comprehensive analysis of cell changes during length is found according to the experimental results. the long-standing load. Furthermore, a correlation between the changes of the RBC color and length is found according to the experimental results. 3. Experimental System and Procedure 3. Experimental System and Procedure 3.1. Cell Manipulation System 3.1. Cell System CellManipulation manipulation is the key technique for this work. Without an accurate cell manipulation for keeping RBCs inside the constriction, RBC for response under long-standing load cannot be properly Cell manipulation is the key technique this work. Without an accurate cell manipulation for observed. Micropipettes or robotic microgrippers may also apply a constant load to RBCs, but they keeping RBCs inside the constriction, RBC response under long-standing load cannot be properly usually take much longer to locate a cell [12,13,21,34]. Thus,a the microfluidic systembut offers observed. Micropipettes ortime robotic microgrippers may also apply constant load to RBCs, theya more convenient platform for the test. usually take much longer time to locate a cell [12,13,21,34]. Thus, the microfluidic system offers a more Figure 2 shows the experimental setup of cell manipulation system, which is composed of a polyconvenient platform for the test. dimethylsiloxane (PDMS) chip, a piezoelectric actuator (MESS-TEK: M-2655s(c), MESS-TEK Ltd., Figure 2 shows the experimental setup of cell manipulation system, which is composed of a Saitama, Japan), a syringe (SGE Analytical Science: 2.5MDF-LL-GT, LGE Japan Ltd., Osaka, Japan), a poly-dimethylsiloxane (PDMS) chip, a piezoelectric actuator (MESS-TEK: M-2655s(c), MESS-TEK high-speed camera (Photoron: IDP-Express R2000, Photron Ltd., Tokyo, Japan), a microscope Ltd., Saitama, Japan), a syringe (SGE Analytical Science: 2.5MDF-LL-GT, LGE Japan Ltd., Osaka, (OLYMPUS: IX71, Olympus Co., Tokyo, Japan), and a personal computer (PC). The high-speed Japan), a high-speed camera (Photoron: IDP-Express R2000, Photron Ltd., Tokyo, Japan), a microscope camera and high-speed actuator are two main components for the cell manipulation inside a (OLYMPUS: IX71, Olympus Co., Tokyo, Japan), and a personal computer (PC). The high-speed camera microfluidic device. The camera provides real-time cell position with the frame rate of 1000 frames and high-speed actuator are two main components for the cell manipulation inside a microfluidic per second (fps) while the high-speed piezoelectric actuator can rapidly adjust cell position by device. The camera provides real-time cell position with the frame rate of 1000 frames per second (fps) generating appropriate fluid flow inside the microfluidic chip. An animation demonstrating the while the high-speed piezoelectric actuator can rapidly adjust cell position by generating appropriate manipulation concept can be found in Supplementary Materials Video S2. The captured image frames fluid flow inside the microfluidic chip. An animation demonstrating the manipulation concept can are processed by image methods for determining the cell position, and the position is sent to a be found in Supplementary Materials Video S2. The captured image frames are processed by image proportional-integral-derivative (PID) controller for computing a suitable output signal to move the methods for determining the cell position, and the position is sent to a proportional-integral-derivative actuator. The algorithms are implemented in a C program. The sampling rate in the program is 1000 (PID) controller for computing a suitable output signal to move the actuator. The algorithms are Hz, which means a slight change of cell position would be immediately compensated by the system implemented in a C program. The sampling rate in the program is 1000 Hz, which means a slight within one millisecond [35]. change of cell position would be immediately compensated by the system within one millisecond [35]. Figure 2. The experimental setup. (a) A photo of the setup; (b) a schematic view of the experimental Figure 2. The experimental setup. (a) A photo of the setup; (b) a schematic view of the experimental system. The RBC is manipulated by a vision-feedback control system with real-time RBC position and system. The RBC is manipulated by a vision-feedback control system with real-time RBC position and microfluidic flow control. microfluidic flow control. Micromachines 2017, 8, 100 4 of 10 3.2. Design and Fabrication of PDMS Chip 3.2. Design and Fabrication of PDMS Chip Figure 3 illustrates the fabrication of the microfluidic chip used in this work. Figure 3a shows Figure 3 illustrates the fabrication of the microfluidic chipchannel used inand thisthe work. Figure 3aare shows the dimensions of the channel design. The widths of the main constriction 10.0 the dimensions of the channel design. The widths of the main channel and the constriction are and 3.24 µm, respectively. The channel height is 3.5 µm for the both parts. The dimensions of10.0 the and 3.24 µm, respectively. The channel height is 3.5 µm for the both parts. The dimensions of the channel are designed for introducing deformation to the tested RBCs. Although RBCs are known channelextraordinary are designeddeformability for introducing to the RBCs. Although are known having anddeformation can go through 1–3tested µm pores without lysis, RBCs the 3.24 µm wide having extraordinary deformability and can go through 1–3 µm pores without lysis, the 3.24 µm wide channel is designed to let the tested RBCs deform around 50% while the length of 15 µm is to make channel is designed to let tested RBCs deform around while thethrough length ofthe 15different µm is to width make sure all deformed RBCs arethe kept inside the constriction. RBC50% deformation sure all deformed RBCs are kept inside the constriction. RBC deformation through the different width of the channel has been previously investigated [36]. An inlet and an outlet are on the two sides of of the channel has for been previously [36]. An inletconnection. and an outlet are on theand twoFigure sides of the the main channel RBCs feedinginvestigated and the syringe pump Figure 3b–f 3g–k main channel for RBCs feeding theand syringe pump connection. Figures 3b–f andmold 3g–kfabrication, illustrate the illustrate the fabrication of the and mold microfluidic chip, respectively. In the a fabrication of the mold and microfluidic chip, respectively. In the mold fabrication, a silicon substrate silicon substrate is spin coated with a layer of photoresist (SU-8 3005, Microchem Corp., Newton, MA, is spinbefore coatedbeing with exposed a layer oftophotoresist Microchem Corp., Newton, USA) before USA) ultraviolet (SU-8 (UV) 3005, light for lithography. The mold isMA, developed using being exposed to ultraviolet (UV) light for lithography. The mold is developed using propylene glycol propylene glycol monomethylether acetate (PM) thinner and isopropyl alcohol (IPA). The mixture of monomethylether (PM) thinner (IPA). TheMI, mixture and curing ® 184, and PDMS and curing acetate agent (Sylgard Dowisopropyl Corning alcohol Corp., Midland, USA) of arePDMS poured over the ® 184, Dow Corning Corp., Midland, MI, USA) are poured over the fabricated mold agent (Sylgard fabricated mold in a container for molding. After the PDMS is cured, the PDMS chip is peeled off in a container molding. After PDMS is the and PDMS chip Finally, is peeledthe offmicrofluidic from the mold and from the moldfor and two holes arethe punched forcured, the inlet outlet. chip is two holes are punched for the inlet and outlet. Finally, the microfluidic chip is bonded to a slide glass bonded to a slide glass using a plasma device (Cute-MP, Femto Science Inc., Gyeonggi-Do, Korea). using a plasma device (Cute-MP, Femto Science Inc., Gyeonggi-Do, Korea). Figure 3. 3. Design Design and and fabrication fabrication of of the the microfluidic microfluidic chip. chip. (a) (a) Schematic Schematic view view of of the the microfluidic microfluidic chip. chip. Figure The device consists of an inlet, an outlet, a microchannel (3.5 µm in height and 10.0 µm in width), and The device consists of an inlet, an outlet, a microchannel (3.5 µm in height and 10.0 µm in width), and narrow channel channel (3.5 (3.5 µm µm in in height, 15.0 µm µm in narrow height, 15.0 in length length and and 3.24 3.24 µm µm in in width); width); (b–f) (b–f) the the step-by-step step-by-step photolithography for mold fabrication; (g–k) the step-by-step procedure for poly-dimethylsiloxane photolithography for mold fabrication; (g–k) the step-by-step procedure for poly-dimethylsiloxane (PDMS) chip fabrication from the mold. (PDMS) chip fabrication from the mold. 3.3. RBC RBC Sample Sample Preparation Preparation and 3.3. and Control Control Sequence Sequence Blood for for the the experiment was provided provided by by aa volunteer subject who who has has read read and Blood experiment was volunteer subject and agreed agreed with with the consent of the test. About 5 µL of blood was taken by a sterilized lancet from a fingertip of the consent of the test. About 5 µL of blood was taken by a sterilized lancet from a fingertip of the the subject. The The blood blood is is diluted diluted 100 100 times times by by standard standard saline saline for for reducing reducing the the number number density density of of the subject. the RBCs in in the the test test sample. sample. The The single single cell cell test test cannot cannot be be performed performed without without the the dilution dilution due due to to the the high high RBCs density of of RBCs RBCs in in the The sample sample solution solution is is injected injected into into the the channel channel from from the the chip chip density the whole whole blood. blood. The inlet, and inlet, and then then the the syringe syringe pump pump is is connected connected to to the the outlet outlet for for the the RBC RBC manipulation. manipulation. Because Because the the changes of RBC color and length are faster at the beginning and gradually approaching to asymptotic changes of RBC color and length are faster at the beginning and gradually approaching to asymptotic values, rates are set 0.1, 1 and for the t = 0–1 s, of t = t1–20 s and values, the theimage imagesampling sampling rates areevery set every 0.1, 10 1 sand 10 intervals s for theofintervals = 0–1 s, 20–290 s, respectively. tt == 1–20 s and t = 20–290 s, respectively. 4. Experimental Results Figure 4 shows an example RBC in the test. The RBC was firstly caught on the left of the constriction from a constant pressure-driven flow, as shown in Figure 4a. During the catching phase, Micromachines 2017, 8, 100 5 of 10 4. Experimental Results Micromachines 2017, 8, 100 Figure 2017, 4 shows Micromachines 8, 100 an 5 of 10 example RBC in the test. The RBC was firstly caught on the left of 5 ofthe 10 constriction from a constant pressure-driven flow, as shown in Figure 4a. During the catching phase, the RBC was held steady for 2 s to recover from prior deformations, such as deformation caused by the RBC was held steady for 2 s to recover from prior deformations, such as deformation caused by the shear stress from the flow. Afterward, a new target position was given as the right end of the the shear stress from the flow. flow. Afterward, a new target position was given as the right end of the constriction, and as a result, the RBC was moved into the constriction and was compressed. Figure constriction, andas asaaresult, result,the the RBC was moved into the constriction and was compressed. Figure constriction, and RBC was moved constriction and wasand compressed. Figure 4b–e 4b–e are the images of the RBC staying inside into the the channel for 0, 30, 100, 300 s, respectively. In 4b–e areimages the images ofRBC the RBC staying inside the channel for 0,100, 30, and 100, 300 and s,300 s, respectively. In are the of the staying inside the channel for 0, 30, respectively. In this this example, we can directly observe the changes of the RBC color and length during the loading this example, we directly can directly observe the changes the color RBC and colorlength and length during the loading example, welength can observe the changes theofRBC during the loading phase. The of the RBC became shorter,of while the color got darker. A time-lapse samplephase. video phase. The of length of the RBC became shorter, while thegot color got darker. A time-lapse sample video The length the RBC became shorter, while the color darker. A time-lapse sample video can be can be found in Supplementary Materials Video S1. can be in found in Supplementary Materials Video S1. found Supplementary Materials Video S1. Figure 4. An example of a RBC in the microfluidic channel before and during loading. (a) The RBC at Figure 4. of in the channel before loading. (a)(a) The RBC at Figure 4. An An example example ofaaRBC RBC themicrofluidic microfluidic channel beforeand andduring during loading. The RBC the catching phase; (b–e) The in RBC response to the long-standing load at 0, 30, 100, and 300 s, the catching phase; (b–e) The RBC response to the long-standing load at 0, 30, 100, and 300 atrespectively. the catching phase; response to the long-standing at 0, 30,time. 100, and 300 s, s, The RBC’s(b–e) colorThe andRBC length were changed with respect toload the loading respectively. The RBC’s RBC’s color color and and length length were were changed changed with with respect respect to to the the loading loadingtime. time. respectively. The Figure 5a–c and Figure 5d-f show the color distribution of two different RBCs during the same Figure 5a–c and Figure 5d-f show the color distribution of two different RBCs during the same long-standing load. y and axes distribution in the figureof represent the x, y positions, and the Figures 5a–c andThe 5d–fx,show thez color two different RBCs during the same long-standing load. The x, y and z axes in the figure represent the x, y positions, and the corresponding color value, respectively. The color is in standard 8-bit grayscale, where the color of long-standing load. The x, y and z axes in the figure represent the x, y positions, and the corresponding corresponding color value, respectively. The color is in standard 8-bit grayscale, where the color of each value, pixel ranges from 0 The (white) tois255 durations of loading shown 5a,d,ranges Figure color respectively. color in (black). standardThe 8-bit grayscale, where the colorin ofFigure each pixel each pixel ranges from 0 (white) to 255 (black). The durations of loading shown in Figure 5a,d, Figure 5b,e and Figure 5c,f are 1, 15 and 290, respectively. One consistent tendency for the two different from 0 (white) to 255 (black). The durations of loading shown in Figure 5a,d, Figure 5b,e and Figure 5c,f 5b,e and Figure 5c,f are 1, 15 and 290, respectively. One consistent tendency for the two different RBCs, well290, as all the tested RBCs, was that the overall for pixel color the cellsRBCs, got darker with are 1, 15asand respectively. One consistent tendency the twoof different as well as respect all the RBCs, as well as all the tested RBCs, was that the overall pixel color of the cells got darker with respect to the RBCs, loading time. thepixel colorcolor patterns were Thewith pattern around theloading center of the tested was thatHowever, the overall of the cellsdifferent. got darker respect to the time. to the loading time. However, the color patterns were different. The pattern around the center of the RBC in Figure 5c is flat whereas the one in Figure 5f is like a valley where the color of the central However, the color patterns were different. The pattern around the center of the RBC in Figure 5c is RBC in Figure 5c is flat whereas the one in Figure 5f is like a valley where the color of the central pixels is lighter the surrounding Thewhere shrinkages of the RBCcentral lengths during the load can flat whereas the than one in Figure 5f is likepixels. a valley the color of the pixels is lighter than pixels is lighter than the surrounding pixels. The shrinkages of the RBC lengths during the load can alsosurrounding be observedpixels. in Figure The lengthofchanges RBC Sample andload 2 can bealso seenbeinobserved Figure 5a–c the The5.shrinkages the RBCoflengths during1the can in also be observed in Figure 5. The length changes of RBC Sample 1 and 2 can be seen in Figure 5a–c and Figure 5d–f, respectively. The shrinkages of the RBCs on the top and bottom of Figure 5 were Figure 5. The length changes of RBC Sample 1 and 2 can be seen in Figures 5a–c and 5d–f, respectively. and Figure 5d–f, respectively. The shrinkages of the RBCs on the top and bottom of Figure 5 were about 5% and 14% with respect to top theirand initial lengths at t = 05s.were According to the two samples shown The shrinkages of the RBCs on the bottom of Figure about 5% and 14% with respect about 5% and 14% with respect to their initial lengths at t = 0 s. According to the two samples shown in Figure 5, we find that while different RBCs share a similar tendency of getting darker and shorter, to their initial lengths at t = 0 s. According to the two samples shown in Figure 5, we find that while in Figure 5, we find that while different RBCs share a similar tendency of getting darker and shorter, they alsoRBCs haveshare different responses in terms of the colorand patterns andthey thealso amount shrinkage during different a similar tendency of getting darker shorter, haveof different responses they also have different responses in terms of the color patterns and the amount of shrinkage during the long-standing in terms of the colorload. patterns and the amount of shrinkage during the long-standing load. the long-standing load. Figure Figure5.5.The Thecolor colordistributions distributionsof oftwo twodifferent differentRBC RBCsamples samplesatatthe theloading loadingtimes timesofoft t==11s,s,tt==15 15s,s, Figure 5. The color distributions of two different RBC samples at the loading times of t = 1 s, t = 15 s, and RBCsample sample22at atthe theloading loadingtimes. times. andt t==290 290s.s.(a–c) (a–c)RBC RBCsample sample11at atthe theloading loadingtimes. times.(d–f) (d–f)RBC and t = 290 s. (a–c) RBC sample 1 at the loading times. (d–f) RBC sample 2 at the loading times. Figure 6 explains the image processing method for the representative cell color and length. Figure 6 explains the image processing method for the representative cell color and length. Figure 6a shows a captured image from the camera. A background image was subtracted from each Figure 6a shows a captured image from the camera. A background image was subtracted from each captured image for background removal. Figure 6b shows the cell image after background removal. captured image for background removal. Figure 6b shows the cell image after background removal. Although the color on a RBC is not uniform and may have different patterns, such as the two Although the color on a RBC is not uniform and may have different patterns, such as the two Micromachines 2017, 8, 100 6 of 10 examples shown in Figure 5, for the convenience of analysis, the gray level was defined as a representative color Micromachines 2017, 8, 100value for a RBC and was the average of the color values over the RBC. The 6RBC of 10 length is the distance between the left-most and right-most points of the detected cell area image as the green arrow indicates in Figure 6. Figure7a,b 6 explains thechange imageofprocessing for the representative cellwith color and length. Figure show the gray levelsmethod and lengths of all the tested RBCs respect to the Figure 6a shows a captured image from the camera. A background image was subtracted from each elapsed time, while Figure 7c shows the correlation between the gray levels and lengths at different Micromachines 2017, 8, 100 6 of 10 captured image fordifferent background removal. Figure 6binitial shows the levels, cell image instances. Because RBCs have different gray the after gray background level at t = removal. 0 s was Although the RBC is not uniform and may have different patterns, as bar the represent two examples normalized tocolor 0 foron thea convenience of comparison. The data points and thesuch error the examples shown 5,infor Figure 5, for the convenience of analysis, the gray levelaswas defined as a shown in Figure the convenience of analysis, the gray level was defined a representative averages and standard deviations of all the tested RBCs. The results in Figure 7a show that the gray representative value for RBCaverage and was the of theover colorthe values over the RBC. TheisRBC color for color a RBC wasato the oftime theaverage color values RBC. The RBC length the level isvalue increasing withand respect the elapsed of loading. The curves of the gray level show an length is the distance between the left-most and right-most points of the detected cell area image as distance between the left-most and right-most points of the detected cell area image as the green arrow asymptotic approach to converging values for the RBCs. the greenin arrow indicates in Figure 6. indicates Figure 6. Figure 7a,b show the change of gray levels and lengths of all the tested RBCs with respect to the elapsed time, while Figure 7c shows the correlation between the gray levels and lengths at different instances. Because different RBCs have different initial gray levels, the gray level at t = 0 s was normalized to 0 for the convenience of comparison. The data points and the error bar represent the averages and standard deviations of all the tested RBCs. The results in Figure 7a show that the gray level is increasing with respect to the elapsed time of loading. The curves of the gray level show an asymptotic approach to converging values for the RBCs. Figure 6. Determination of RBC gray level and length from a captured image. (a) A raw image frame; Figure 6. Determination of RBC gray level and length from a captured image. (a) A raw image frame; (b) the enhanced image by removing the background. (b) the enhanced image by removing the background. Figure 7a,b show the change of gray levels and lengths of all the tested RBCs with respect to the elapsed time, while Figure 7c shows the correlation between the gray levels and lengths at different instances. Because different RBCs have different initial gray levels, the gray level at t = 0 s was normalized to 0 for the convenience of comparison. The data points and the error bar represent the averages and standard deviations of all the tested RBCs. The results in Figure 7a show that the gray 6. Determination of RBC level andtime length a captured image. of (a) the A raw image frame; level Figure is increasing with respect to gray the elapsed offrom loading. The curves gray level show an (b) the enhanced image by removing the background. asymptotic approach to converging values for the RBCs. Figure 7. The experimental results of the gray level and cell length of the RBCs with respect to the loading time. (a) The change of the gray level with respect to the loading time; (b) the change of the cell length with respect to the loading time; (c) the change of the gray level with respect to the cell length at different instances. The normalized cell length in Figure 7b was calculated as the ratio between every measured cell length and its initial length. The normalization was done for the convenience of comparing RBCs with different initial lengths in the narrow channel. The curves of the length were also logarithmically Figure experimental results results of the the gray gray level cell the Figure 7. 7. The level and cell length length of the RBCs RBCs with with respect respect to to the the converging inThe theexperimental 5 min load. Figureof 7c indicates aand series of dataofsets at different elapsed times loading time. (a) The change of the gray level with respect to the loading time; (b) the change of the loading time. (a) The change of the gray level with respect to the loading time; (b) the change of cell the between the gray level and length. Negative correlations between the gray level and cell length are length with with respect to thetoloading time; (c) the (c) change of the gray level with respect the celltolength at cell length respect the loading time; the change of the gray level withtorespect the cell observed. Aninstances. interesting observation from Figure 7c is that during the first 15 s of the loading, the different length at different instances. The normalized cell length in Figure 7b was calculated as the ratio between every measured cell length and its initial length. The normalization was done for the convenience of comparing RBCs with different initial lengths in the narrow channel. The curves of the length were also logarithmically converging in the 5 min load. Figure 7c indicates a series of data sets at different elapsed times between the gray level and length. Negative correlations between the gray level and cell length are Micromachines 2017, 8, 100 7 of 10 The normalized cell length in Figure 7b was calculated as the ratio between every measured cell length and its initial length. The normalization was done for the convenience of comparing RBCs with different initial lengths in the narrow channel. The curves of the length were also logarithmically converging in the 5 min load. Figure 7c indicates a series of data sets at different elapsed times between Micromachines 2017, 8, 100 7 of 10 the gray level and length. Negative correlations between the gray level and cell length are observed. An interesting observation from Figure is gray that during the first 15 s of the loading, the 15 cells length cell length continuously decreased while7cthe level remained almost constant. From to the continuously decreased while the gray level remained almost constant. From 15 s to the end of the end of the loading, the gray level increased, corresponding to the decrease of the cell length. That loading, graywere levelshrinking increased, corresponding to the decrease of the That means the means thethe RBCs first before changing color in the first 15 cell s of length. the loading. RBCs were shrinking first before changing color in the first 15 s of the loading. 5. Discussion 5. Discussion According to the experimental results in Figure 7, there is a correlation between the change of According to the experimental results in Figure 7, there is a correlation between the change of the the gray level and the cell length during the long-standing loads. The decrease of the cell length gray level and the cell length during the long-standing loads. The decrease of the cell length happened happened earlier than the increase of the gray level. This fact says that the increase of the gray level earlier than the increase of the gray level. This fact says that the increase of the gray level may be may be brought about by the decrease of the cell length. For interpreting the phenomenon, we brought about by the decrease of the cell length. For interpreting the phenomenon, we separately separately consider the cell behaviors in the entrance phase and the loading phase as follows: consider the cell behaviors in the entrance phase and the loading phase as follows: Figure 8a–d show a RBC entering a constriction. The left column of Figure 8 shows a series of Figure 8a–d show a RBC entering a constriction. The left column of Figure 8 shows a series of actual photos indicating the cell behavior during the entrance phase. A video clip of a RBC entering actual photos indicating the cell behavior during the entrance phase. A video clip of a RBC entering and exiting the constriction can be found in Supplementary Materials Video S3. A RBC is usually and exiting the constriction can be found in Supplementary Materials Video S3. A RBC is usually assumed being able to deform freely as long as the total surface area remains constant [37]. Based on assumed being able to deform freely as long as the total surface area remains constant [37]. Based on the assumption, the possible shape changes of the RBC during the entrance phase are illustrated in the assumption, the possible shape changes of the RBC during the entrance phase are illustrated in the the middle and right columns of Figure 8, which are the top view and the cross-sectional view, middle and right columns of Figure 8, which are the top view and the cross-sectional view, respectively. respectively. According to the photos, the RBC seems not only to be compressed but also folded in According to the photos, the RBC seems not only to be compressed but also folded in half for entering half for entering into the constriction. The cross-sectional area of the channel is gradually filled with into the constriction. The cross-sectional area of the channel is gradually filled with the deformed the deformed RBC during the entrance. However, there are still vacancies around the deformed RBCs RBC during the entrance. However, there are still vacancies around the deformed RBCs after the full after the full entrance of the RBC. entrance of the RBC. Figure 8. Interpretations of RBC deformation while squeezing into a constriction channel. (a) RBC Figure 8. Interpretations of RBC deformation while squeezing into a constriction channel. (a) RBC before the deformation; (b) the tip of the RBC is compressed and folded for squeezing into the channel; before the deformation; (b) the tip of the RBC is compressed and folded for squeezing into the channel; (c) a greater area of the cross-section is filled by the RBC as it moves into the channel; (d) the RBC is (c) a greater area of the cross-section is filled by the RBC as it moves into the channel; (d) the RBC is fully squeezed into the narrow channel. fully squeezed into the narrow channel. Figure 99illustrates a possible shape change of the RBC during long-duration load. Illustrations Figure illustrates a possible shape change of the RBCthe during the long-duration load. of another cross-sectional view from the bottom are additionally included in Figure 9. The in decreases of Illustrations of another cross-sectional view from the bottom are additionally included Figure 9. the cell lengthof may as the cell gradually adapts to the confined the space constriction, The decreases thehappen cell length may happen as the cell gradually adaptsspace to theinside confined inside and tries to fill the vacancies. This kind of shape adaption is generally known as strain creep for typical the constriction, and tries to fill the vacancies. This kind of shape adaption is generally known as viscoelastic materials, and it means that the and deformation of a compressed RBCofisaacompressed function of RBC time strain creep for typical viscoelastic materials, it means that the deformation when the load constraint is fixed. Based on the assumption of the surface of area, shape is a function ofor time when the load or constraint is fixed. Based on constant the assumption thethis constant surface area, this shape change makes the cell length shorten in the narrow channel. Furthermore, the adaption results in an equivalently thicker RBC from the top view as shown on the right of Figure 9c. The increase of the RBC thickness would reduce the transparency of the cell, and as a result, the cell is darkened and the gray levels increase as the experimental results show in Figure 7. However, further fluorescence or confocal microscopic images are needed to confirm the morphology changes Micromachines 2017, 8, 100 8 of 10 change makes the cell length shorten in the narrow channel. Furthermore, the adaption results in an equivalently thicker RBC from the top view as shown on the right of Figure 9c. The increase of the RBC thickness would reduce the transparency of the cell, and as a result, the cell is darkened and the gray levels increase as the experimental results show in Figure 7. However, further fluorescence or confocal microscopic images are needed to confirm the morphology changes of RBCs during the long-standing Micromachines 2017, 8, 100 8 of 10 load and are the future direction of this work. Figures 88 and and 99 provide provide aa possible possible interpretation interpretation for for the the correlation correlation between between the the changes changes of of the the Figures RBC color color and and length length from from the the perspectives perspectives of of mechanical mechanical deformation. deformation. We We would would like like to to emphasize emphasize RBC that the interpretation does not exclude other possible changes inside the RBC, such as reformation of that the interpretation does not exclude other possible changes inside the RBC, such as reformation the cytoskeleton or the exhaust of adenosine triphosphate inside the RBCs. In fact, it is very possible of the cytoskeleton or the exhaust of adenosine triphosphate inside the RBCs. In fact, it is very that there were fundamental changes changes of the RBCs based onbased the findings from the from previous study on possible that there were fundamental of the RBCs on the findings the previous cell fatigue [24]. The other possible effects of RBCs under long-standing loads are worth investigating study on cell fatigue [24]. The other possible effects of RBCs under long-standing loads are worth further from further different perspectives, such as chemistry biology, forbiology, RBC deformability. investigating from different perspectives, such asand chemistry and for RBC deformability. Figure InterpretationsofofRBC RBC changes during the long-standing in the constriction. (a) Figure 9. Interpretations changes during the long-standing load inload the constriction. (a) Loading Loading time t = 60 s; (b) time loading times;t (c) = 180 s; (c) loading time t = 60 s; (b) loading t = 180 loading time t = time 300 s.t = 300 s. 6. 6. Conclusions Conclusions The The responses responses of of RBCs RBCs during during aa long-standing long-standing load load in in aa microfluidic microfluidic constriction constriction are are experimentally experimentally investigated. investigated.The Thecolor colorof ofRBCs RBCsin in the the constriction constriction becomes becomes darker darker while while the the length length of respect to to thethe loading time. The The spatial and temporal variations of theof color of the theRBCs RBCsshrinks shrinkswith with respect loading time. spatial and temporal variations the change are analyzed. The results show both logarithmic changes of the color and the length of color change are analyzed. The results show both logarithmic changes of the color and the lengththe of RBCs. We also find an unanticipated negative correlation between the color and the RBC length the RBCs. We also find an unanticipated negative correlation between the color and the RBC length according according to to the the experimental experimental results. results. Interpretation Interpretation from from the the perspective perspective of ofmechanical mechanicaldeformation deformation is is discussed. discussed. RBC RBCresponses responsesunder undersuch suchaa long-standing long-standingload loadare aresimilar similarto to aa RBC RBC being being plugged plugged in in microcirculation inside a body, and the results could provide new insights into the time-dependent microcirculation inside a body, and the results could provide new insights into the time-dependent quality quality of of plugged plugged RBCs. RBCs. Supplementary areare available online at www.mdpi.com/2072-666X/8/4/100/s1; Video Supplementary Materials: Materials: The Thefollowing following available online at www.mdpi.com/2072-666X/8/4/100/s1; Video S1: A time-lapse theresponse cell response during a long-standing Video S2: The conceptual animation S1: A time-lapse video video of the of cell during a long-standing load;load; Video S2: The conceptual animation of of cell manipulation; Video S3: A slow-motion video of a RBC passing through a constriction channel showing the cell manipulation; Video S3: A slow-motion video of a RBC passing through a constriction channel showing the dynamic response of cell folding and unfolding. dynamic response of cell folding and unfolding. Acknowledgments: This work was partially supported by JSPS KAKENHI Grant Nos. JP15H05761, JP16K14197, Acknowledgments: This work was partially supported by JSPSInstitute, KAKENHI Grant Nos. JP15H05761, JP16K14197, JP16K18051, JP16H06933, Foundation Advanced Technology and “Nanotechnology Platform Project JP16K18051, JP16H06933, Foundation Advanced Technology Institute, “Nanotechnology Platform Project (Nanotechnology Open Facilities in Osaka University)” of MEXT Japan and [F-16-OS-0006, S-16-OS-0005]. (Nanotechnology Open Facilities in Osaka University)” of MEXT Japan [F-16-OS-0006, S-16-OS-0005]. Author Contributions: All authors conceived and designed the experiments; M.H., C.-H.D.T. and H.I. performed the experiments; M.H., C.-H.D.T. and H.I. contributed to the data analysis and interpretations; M.H., C.-H.D.T. and M.K. wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. Micromachines 2017, 8, 100 9 of 10 Author Contributions: All authors conceived and designed the experiments; M.H., C.-H.D.T. and H.I. performed the experiments; M.H., C.-H.D.T. and H.I. contributed to the data analysis and interpretations; M.H., C.-H.D.T. and M.K. wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Glenister, F.; Coppel, R.; Cowman, A.F.; Mohandas, N.; Cooke, B.M. Contribution of parasite proteins to altered mechanical properties of malaria-infected red blood cells. Blood 2002, 99, 1060–1063. [CrossRef] [PubMed] Lim, C.T. Single cell mechanics study of the human disease malaria. J. Biomech. Sci. Eng. 2006, 1, 82–92. [CrossRef] Tomaiuolo, G. Biomechanical properties of red blood cells in health and disease towards microfluidics. Biomicrofluidics 2014, 8, 51501. [CrossRef] [PubMed] Mokken, F.; Kedaria, M.; Henny, C. The clinical importance of erythrocyte deformability, a hemorrheological parameter. Ann. Hematol. 1992, 64, 113–122. [CrossRef] [PubMed] Li, X.; Vlahovska, P.; Karniadakis, G. Continuum- and particle-based modeling of shapes and dynamics of red blood cells in health and disease. Soft Matter 2013, 9, 28–37. [CrossRef] [PubMed] Roth, K.B.; Eggleton, C.D.; Neeves, K.B.; Marr, D.W.M. Measuring cell mechanics by optical alignment compression cytometry. Lab Chip 2013, 13, 1571–1577. [CrossRef] [PubMed] Gossett, D.R.; Henry, T.K.; Lee, S.A.; Ying, Y.; Lindgren, A.G.; Yang, O.O.; Rao, J.; Clark, A.T.; Di Carlo, D. Hydrodynamic stretching of single cells for large population mechanical phenotyping. Proc. Natl. Acad. Sci. USA 2012, 109, 7630–7635. [CrossRef] [PubMed] Otto, O.; Rosendahl, P.; Mietke, A.; Golfier, S.; Herold, C.; Klaue, D.; Girardo, S.; Pagliara, S.; Ekpenyong, A.; Jacobi, A.; et al. Real-time deformability cytometry: On-the-fly cell mechanical phenotyping. Nat. Methods 2015, 12, 199–202. [CrossRef] [PubMed] Tsai, C.D.; Sakuma, S.; Arai, F.; Taniguchi, T.; Ohtani, T.; Sakata, Y.; Kaneko, M. Geometrical alignment for improving cell evaluation in a microchannel with application on multiple myeloma red blood cells. RSC Adv. 2014, 4, 45050–45058. [CrossRef] Salehyar, S.; Zhu, Q. Deformation and internal stress in a red blood cell as it is driven through a slit by an incoming flow. Soft Matter 2016, 12, 3156–3164. [CrossRef] [PubMed] Park, S.Y.; Dimitrakopoulos, P. Transient dynamics of an elastic capsule in a microfluidic constriction. Soft Matter 2013, 19, 8844–8855. [CrossRef] [PubMed] Fischer, T.M. Shape memory of human red blood cells. Biophys. J. 2004, 86, 3304–3313. [CrossRef] Markle, D.R.; Evans, E.A.; Hochmuth, R.M. Force relaxation and permanent deformation of erythrocyte membrane. Biophys. J. 1983, 42, 91–98. [CrossRef] Baez, S.D.; Kaul, M.E.; Nagel, R.L. Microvascular determinants of blood flow behavior and HbSS erythrocyte plugging in microcirculation. Blood Cells 1982, 8, 127–137. [PubMed] Wiewiora, M.; Sosada, K.; Wylezol, M.; Slowinska, L.; Zurawinski, W. Red blood cell aggregation and deformability among patients qualified for bariatric surgery. Obes. Surg. 2007, 17, 365–371. [CrossRef] [PubMed] Cahn, A.; Livshits, L.; Srulevich, A.; Raz, I.; Yedgar, S.; Barshtein, G. Diabetic foot disease is associated with reduced erythrocyte deformability. Int. Wound J. 2015, 13, 500–504. [CrossRef] [PubMed] Dondorp, A.M.; Kager, P.A.; Vreeken, J.; White, N.J. Abnormal blood flow and red blood cell deformability in severe malaria. Parasitol. Today 2000, 16, 228–232. [CrossRef] Pivkin, I.V.; Peng, Z.; Karniadakis, G.E.; Buffet, P.A.; Dao, M.; Suresh, S. Biomechanics of red blood cells in human spleen and consequences for physiology and disease. Proc. Natl. Acad. Sci. USA 2016, 113, 7804–7809. [CrossRef] [PubMed] Lawrence, C.; Fabry, M.E.; Nagel, R.L. The unique red cell heterogeneity of SC disease: Crystal formation, dense reticulocytes, and unusual morphology. Blood 1991, 78, 2104–2112. [PubMed] Zheng, Y.; Nguyen, J.; Wei, Y.; Sun, Y. Recent advances in microfluidic techniques for single-cell biophysical characterization. Lab Chip 2013, 13, 2464–2483. [CrossRef] [PubMed] Micromachines 2017, 8, 100 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 10 of 10 Reich, E.; Goldberg, I.H. Actinomycin and nucleic acid function. Prog. Nucleic Acid Res. Mol. Boil. 1964, 3, 183–234. Tözeren, A.; Skalak, R.; Sung, K.L.; Chien, S. Influence of physicochemical factors on rheology of human neutrophils. Biophys. J. 1982, 39, 23–32. [CrossRef] Brandao, M.M.; Fontes, A.; Barjas-Castro, M.L.; Barbosa, L.C.; Costa, F.F.; Cesar, C.L.; Saad, S.T.O. Optical tweezers for measuring red blood cell elasticity: Application to the study of drug response in sickle cell disease. Eur. J. Haematol. 2003, 70, 207–211. [CrossRef] [PubMed] Dulińska, I.; Targosz, M.; Strojny, W.; Lekka, M.; Czuba, P.; Balwierz, W.; Szymoński, M. Stiffness of normal and pathological erythrocytes studied by means of atomic force microscopy. J. Biochem. Biophys. Methods 2006, 66, 1–11. [CrossRef] [PubMed] Hirose, Y.; Kaneko, M.; Kawahara, T.; Yamanishi, Y.; Arai, F. High speed cell stiffness evaluation toward 100% reliability. In Proceedings of the 2011 IEEE SENSORS, Limerick, Ireland, 28–31 October 2011; pp. 962–965. Hirose, Y.; Tadakuma, K.; Higashimori, M.; Arai, T.; Kaneko, M.; Iitsuka, R.; Yamanishi, Y.; Arai, F. A new evaluation toward high speed cell sorter. In Proceedings of the 2010 IEEE International Conference Robotics and Automation (ICRA), Anchorage, MD, USA, 3–7 May 2010; pp. 4113–4118. Tsai, C.D.; Sakuma, S.; Arai, F.; Kaneko, M. A new dimensionless index for evaluating cell stiffness-based deformability in microchannel. IEEE Trans. Biomed. Eng. 2014, 61, 1187–1195. [CrossRef] [PubMed] Zheng, Y.; Shojaei-Baghini, E.; Azad, A.; Wang, C.; Sun, Y. High-throughput biophysical measurement of human red blood cells. Lab Chip 2012, 13, 2560–2567. [CrossRef] [PubMed] Sakuma, S.; Kuroda, K.; Tsai, C.D.; Fukui, W.; Arai, F.; Kaneko, M. Red blood cell fatigue evaluation based on close-encountering point between extensibility and recoverability. Lab Chip 2014, 14, 1135–1141. [CrossRef] [PubMed] Ito, H.; Murakami, R.; Sakuma, S.; Tsai, C.D.; Gutsmann, T.; Brandenburg, K.; Ito, H.; Murakami, R.; Sakuma, S.; Tsai, C.D.; et al. Mechanical diagnosis of human erythrocytes by ultra-high speed manipulation unraveled critical time window for global cytoskeletal remodeling. Sci. Rep. 2017, 7, 43134. [CrossRef] [PubMed] Murakami, R.; Tsai, C.D.; Ito, H.; Tanaka, M.; Sakuma, S.; Arai, F.; Kaneko, M. Catch, Load and Launch toward On-Chip Active Cell Evaluation. In Proceedings of the 2016 IEEE International Conference Robotics and Automation (ICRA), Stockholm, Sweden, 16–21 May 2016; pp. 1713–1718. Horade, M.; Tsai, C.D.; Ito, H.; Tanaka, M.; Kaneko, M. Chameleon effect of RBC under loading in microfluidic channel. In Proceedings of the 19th International Conference on Miniaturized Systems for Chemistry and Life Sciences (MicroTAS16), Dublin, Ireland, 9–13 October 2016; pp. 311–312. Tsai, C.D.; Horade, M.; Ito, H.; Tanaka, M.; Kaneko, M. High-resolution cell manipulation for longstanding load on red blood cells. In Proceedings of the IEEE International Conference on Mechatronics and Automation (ICMA2016), Harbin, China, 7–10 August 2016; pp. 914–919. Horade, M.; Kojima, M.; Kamiyama, K.; Kurata, T.; Mae, Y.; Arai, T. Development of an optimum end-effector with a nano-scale uneven surface for non-adhesion cell manipulation using a micromanipulator. J. Micromech. Microeng. 2015, 25, 5002. [CrossRef] Monzawa, T.; Kaneko, M.; Tsai, C.D.; Sakuma, S.; Arai, F. On-chip actuation transmitter for enhancing the dynamic response of cell manipulation using a macro-scale pump. Biomicrofluidics 2015, 9, 014114. [CrossRef] [PubMed] Tsai, C.D.; Tanaka, J.; Kaneko, M.; Horade, M.; Ito, H.; Taniguchi, T.; Ohtani, T.; Sakata, Y. An on-chip RBC deformability checker significantly improves velocity-deformation correlation. Micromachines 2016, 7, 176–185. [CrossRef] Evans, E.A.; Waugh, R.; Melnik, L. Elastic area compressibility modulus of red cell membrane. Biophys. J. 1976, 16, 585–595. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).