An all-on-chip method for testing neutrophil chemotaxis induced by fMLP and COPD patient’s sputum

INNOVATION

Although many microfluidic devices have been developed for quantitative immune cell migration studies and relevant clinical applications, cell isolation in most studies still relies on specialized facilities. The process is labor-intensive and time-consuming and requires specialized skills. Our novel approach provides an easy-to-use microfluidic device for efficient on-chip magnetic negative neutrophil isolation from a small amount of whole blood. Furthermore, we integrated the cell-docking structure and the chemical gradient generator to improve the control of cell patterning and chemotaxis analysis. This developed device and method can be further applied to other cell types such as other leukocyte subsets, cancer cells and stem cells for migration analysis.

INTRODUCTION

Directional cell migration plays important roles in many biological processes and diseases such as host defense, tissue generation and metastatic cancers^1–3. Among the various environmental guiding mechanisms^4,5, chemical gradient can direct the migration of different cell types by chemotaxis. Neutrophils are an important type of white blood cells that play a crucial role in host defense. Its immune functions critically rely on chemotaxis⁶. Neutrophils are well-established cell models for studying cell migration and chemotaxis^5,7, and neutrophil chemotaxis has been targeted for disease-oriented research^8–10. Traditionally, neutrophils are isolated by a method based on density gradient centrifugation¹¹, which is time-consuming, requires large amounts of blood and can cause cell activation and damage. Magnetic cell isolation method can further enrich neutrophils from the polymorphonuclear (PMN) fraction of the blood¹². However, the magnetic method typically still requires centrifugation and has higher reagent cost. Nevertheless, traditional neutrophil isolation methods are still widely used for different cell migration assays.

Compared with conventional cell migration assays, microfluidic devices serve as useful experimental tools for quantitative cell migration and chemotaxis analysis in well-controlled chemical gradients¹³. Various microfluidic gradient devices have been developed and applied to neutrophil chemotaxis analysis¹³. In particular, several studies have demonstrated neutrophil migration testing directly from blood by integrating on-chip neutrophil isolation with adhesion-based neutrophil capturing or geometric confinement^10,14,15. Recently, a new magnetic neutrophil isolation kit for negative neutrophil selection directly from whole blood has been developed (EasySep Direct)¹⁶. In this study, our primary objective was to integrate the EasySep Direct kit with a microfluidic device for all-on-chip neutrophil chemotaxis testing directly from blood. The useful features of this method include: 1) isolation of neutrophils at high purity directly from blood; 2) enabling of rapid neutrophil chemotaxis experiments directly from small amounts of blood using small volumes of cell isolation reagents; and 3) enabling of on-chip chemotaxis testing using negatively selected neutrophils, which avoids possible activation and damage during traditional cell isolation processes. Two types of microfluidic devices were developed in this study. A previously reported microfluidic gradient generator (i.e. type I device)⁸ was adapted to validate the cell isolation method. The type II device integrated a cell-docking structure so that the cells are aligned next to the gradient channel and therefore the subsequent cell migration in response to the applied chemoattactant gradient can be easily and more accurately quantitated. To validate this method, we tested neutrophil chemotaxis to N-formylmethionyl-leucyl-phenylalanin (fMLP), a well-known neutrophil chemoattractant, and to sputum from patients with COPD, which has been shown to attract neutrophils¹⁷.

MATERIALS AND METHODS

Microfluidic device preparation

The preparation of the microfluidic chemotaxis device has been previously described⁸. Briefly, the master of the microfluidic device was fabricated using the standard photolithography method. Two types of master were fabricated. The type I device was fabricated using a single-layer photolithography method. The pattern was designed using AutoCAD software and printed onto a transparency film as the photomask. SU-8 photoresist (~60 µm thick) was spin-coated on a silicon wafer. The pattern design was transferred to the photoresist by UV exposure through the photomask. The type II device with the cell-docking structure was fabricated using a two-layer photolithography method¹⁸. A first layer of SU-8 (~4 µm) pattern was fabricated on the silicon wafer. Then, another layer of SU-8 (~60 µm thick) pattern was aligned on the top of the first layer. The polydimethylsiloxane (PDMS) replica of the pattern was fabricated by the standard soft lithography method. Holes (4 mm in diameter) for the two chemical inlet wells and one outlet well were punched out of the PDMS replica. Thereafter, the PDMS replica was plasma-bonded to a glass slide to complete the microfluidic device assembly. The thin ~4 µm channel in the type II device was designed to pattern cells by preventing cells from entering the gradient channel before the gradient is applied. The microfluidic channel was coated with fibronectin for 1 hour followed by BSA blocking for another hour in room temperature before cell migration experiments.

Collection of sputum and blood samples

The sputum samples from COPD patients were collected at the Seven Oak General Hospital in Winnipeg, Manitoba (Canada). Approval from a research ethics board was obtained from the University of Manitoba. Supernatant from the sputum was prepared following the protocol as described previously⁸. The supernatant was used as a chemoattractant for neutrophil chemotaxis experiments. Blood samples from healthy donors were collected at the Victoria General Hospital in Winnipeg under a separate approved human research ethics protocol.

On-chip cell isolation, cell migration experiment and analysis

The type I microfluidic device and the experimental procedures for the all-on-chip neutrophil chemotaxis testing are illustrated in Fig. 1. The microfluidic device follows the same design we reported previously⁸, but the area near the outlet was specially shaped to allow easy attachment of the magnets. The serpentine input channel in each side is 60 mm long and 200 µm wide. The gradient channel is ~6 mm long and 350 µm wide. This device allows rapid generation of chemical gradient in the microfluidic channel for chemotaxis experiments in a pump-free manner.

For on-chip neutrophil isolation, 10 μL of whole blood was added into a 1.5 mL Eppendorf tube. Then, 2 µL of antibody cocktail and the magnetic particles (Ab–MP) from the neutrophil isolation kit for negative neutrophil selection (EasySep^™ Direct Human Neutrophil Isolation Kit, STEMCELL) were added to the tube and mixed with the blood. The blood–Ab–MP mixture was incubated in room temperature for 5 minutes. Two small magnetic disks (5 mm in diameter and 1 mm thick; CAT# 44202-1, Indigo Instruments) were vertically attached to the two sides of the outlet well. The medium in the inlet and outlet wells was aspirated before loading the blood–Ab–MP mixture. Two microliters of this mixture were loaded to the center of the outlet well using a pipette tip. The magnetically conjugated cells moved to the side walls of the outlet well, driven by the magnetic force, while neutrophils were free to enter and settle on the fibronectin-coated gradient channel. Next, the chemoattractant solution (prepared in RPMI-1640 with 0.4% BSA) and medium were added to the inlet wells. Two chemoattractant solutions were used: 1) 100 nM fMLP; and 2) 10× diluted supernatant of the sputum from the COPD patients. Because of the pressure difference between the inlets and outlet of the gradient channel, the solutions flowed into the gradient channel and a concentration gradient of the chemoattractant was generated based on laminar flow mixing⁸. FITC–dextran 10 kDa (Sigma-Aldrich, MO; final concentration 5 μM in RPMI-1640) was added to the chemoattractant solution to verify gradient formation by measuring the fluorescence intensity profile in the channel. At this point, the magnets were removed from the device, and flow prevented cells in the outlet well from entering the channel.

The type II device is illustrated in Fig. 2. The dimensions of the gradient channels are similar to the type I device except the width of the gradient channel of the type II device is 280 µm. The gradient generation method and cell isolation procedures were similar to those described for the type I device. To test the stability of the gradient, we measured the fluorescence intensity profile of FITC–dextran in the gradient channel. The results showed stable gradient profile over an hour, which is sufficiently long for the cell migration experiment in this study (Fig. 2e). Instead of isolating cells from the outlet as in the type I device, the magnetic cell isolation takes place through a separate cell-loading port and channel on one side of the gradient channel. The cell-loading channel connects to the cell-docking region. All the inlets and outlet were emptied before loading the blood. However, the medium remained inside the cell-loading channel and the bottom of the cell-loading port. A few microliters of the blood–Ab–MP mixture were added to the cell-loading port and mixed with the medium. The isolated neutrophils flowed into the cell-docking region driven by the pressure difference between the cell-loading port and the gradient channel inlets/outlet. The cells were trapped and aligned next to the gradient channel in the cell-docking regions because of the larger cell size compared to the ~4 μm high docking channel.

Cell migration was recorded by time-lapse imaging at six frames per minute using a Nikon Ti-microscope with an environmental control chamber to maintain the temperature at 37°C throughout the experiment. After the experiments, 10–100 individual cells from the time-lapse images for each experiment were tracked to calculate quantitative cell migration parameters including chemotactic index (CI) and migration speed using the established method¹⁹. CI is the ratio of cell displacement toward the gradient to the total migration distance. The migration speed is the ratio of total migration distance to the experiment period. The cell morphology analysis was performed by measuring the aspect ratio following the previously described method⁸. Basically, the aspect ratio was calculated as the ratio of the long and short axes of the best-fitting ellipse of the cell contour. Because the cells were patterned with the same starting position relative to the gradient in the type II device, we analyzed the cell migration distance along the direction of the gradient into the gradient channel as a measure of chemotaxis (Fig. 5b).

Cell staining and imaging

Cells (isolated by the on-chip method) in the microfluidic channel were fixed with methanol and then stained by 1:20 diluted Giemsa stain solution (Rowley Biochemical, Danvers, MA)²⁰. Then the stained cells were washed with deionized water and imaged by an inverted light microscope. In a separate experiment, cells (isolated by the on-chip method) in the microfluidic channel were stained with FITC-conjugated anti-CD66b antibody (CAT# 10419, STEMCELL) for cell surface CD66b expression followed by washing with PBS twice and imaging with a fluorescence microscope.

Statistical analysis of the data

At least three independent experiments were performed for each condition. All parameters are presented as the average ± standard error of the mean (s.e.m.). Statistical analyses were performed using the two-sample student’s t-test. p < 0.05 (*) was considered significantly different.

RESULTS

Characterization of on-chip neutrophil isolation

First, we verified that without using the microfluidic device, the EasySep Direct kit enabled isolation of PMN cells at high purity (i.e. >95%) from larger volumes of blood, as shown by the standard flow cytometric analysis (Fig. 3a). Next, we validated the on-chip method by on-chip Giemsa staining (Fig. 3b) and CD66b staining (Fig. 3c). The Giemsa staining showed typical ring-shaped and lobe-shaped nuclei feature of the neutrophils. The CD66b staining also showed a positive result. These results suggest that the on-chip isolation method enabled effective isolation of PMN cells at high purity (near 100%) from much smaller volumes of blood. Because the on-chip method starts from very small volumes of blood, the amount of isolation reagent is also significantly reduced (i.e. 25-fold less than the standard bulk isolation with the EasySept Direct kit), while still providing enough cells for the subsequent chemotaxis experiments in the same device. The whole on-chip cell isolation can be completed in less than 10 minutes.

Direct on-chip testing of neutrophil migration to fMLP and sputum from COPD patients using the type I device

We next validated the type I device by testing neutrophil chemotaxis to a well-known chemoattractant, fMLP. After the on-chip cell isolation, a 100 nM fMLP gradient or medium flow alone as the control was applied to the cells in the gradient channel, and cell migration was recorded for 15 minutes. Because the cell migration experiment was immediately following the on-chip cell isolation in the same device, the whole experiment from blood to chemotaxis testing was done in less than 25 minutes. CI and migration speed in the fMLP gradient were significantly higher compared to the medium control, thus clearly demonstrating chemotaxis to fMLP using this all-on-chip method (Fig. 4a).

We previously demonstrated neutrophil chemotaxis induced by sputum from COPD patients using a similar microfluidic device and the traditional cell isolation method⁸. This study suggested the potential of microfluidics-based neutrophil chemotaxis test for clinical COPD diagnosis. However, the lengthy traditional cell isolation method and the associated requirements in facility, as well as cost and skills, present a major bottleneck for clinical applications. We believe the all-on-chip method provides a rapid and cost effective solution. To further validate the on-chip method, we evaluated neutrophil chemotaxis to COPD sputum. Our results clearly demonstrated neutrophil chemotaxis and motility induced by COPD sputum, as shown by the CI and migration speed (Fig. 4a). Compared with the medium control, both CI and cell speed in the fMLP and COPD sputum gradient are much higher. These results demonstrate that the chemoattractant gradient stimulated neutrophil chemotaxis. Furthermore, CI but not cell speed in the fMLP gradient was higher than it in the COPD sputum gradient. These results are in qualitative agreement with our previous study using a similar type of device⁸. However, the underlying biological mechanisms require further investigation. The quantitative cell migration parameters in different sets of experiments can also depend on the chemoattractant doses, blood donors and COPD sputum samples. Consistent with our previous results using the traditional cell isolation method⁸, we found elongated cell morphology (higher aspect ratio) in the COPD sputum but fan-shaped cell morphology in the fMLP gradient (lower aspect ratio) using the on-chip method (Fig. 4b). Consistent with our previous work⁸, the cell morphology in the COPD sputum was similar to it in the IL-8 gradient, suggesting that neutrophil chemotaxis to the COPD sputum gradient is mainly induced by tissue-derived chemoattractants such as IL-8 in the sputum. The observed difference in cell morphologies induced by different chemoattractant gradients requires further investigation to elucidate the underlying mechanisms.

Direct on-chip testing of neutrophil migration to fMLP and sputum from COPD patients using the type II device

In several experiments using the type I device, we found that the cells could not attach well to the fibronectin-coated channel under flow. Furthermore, in the type I device, there was no control of the cells’ initial positions in the channel relative to the gradient, which complicated the accuracy of the chemotaxis analysis. To overcome these limitations, we further developed a type II device by incorporating cell-docking structures. As shown in Fig. 2d, the cells isolated from the whole blood were successfully trapped in the cell-docking area. The stuff outside the docking area and stuck at the edge of gradient channel should be some broken cells and debris, which was observed to have limited affection to the cell migration. After the on-chip cell isolation, the cells were exposed to the medium flow, and the cell migration was recorded for 15 minutes as the control. During this period, very few cells crossed the barrier and moved into the gradient channel (Fig. 5a). When a 100 nM fMLP gradient was applied to the gradient channel. Many more cells responded quickly and migrated into the gradient channel (Fig. 5a). Similarly, the cells showed migratory response to the COPD sputum gradient in the type II device (Fig. 5a). The results show higher CI and migration speed in the fMLP gradient than those in the COPD sputum gradient (Fig. 5c). In addition, cells migrated a longer distance toward the fMLP gradient than toward the COPD sputum gradient over the 15 minutes experiment (Fig. 5d). This observation indicates the more persistent chemotactic migration of neutrophils to the fMLP gradient and is consistent with the higher CI in the fMLP gradient than the COPD sputum gradient in both the type I and type II devices (Fig. 4a,5c). The variations of quantitative cell migration parameter comparison between the type I and type II devices in this study can be in part due to neutrophils from different blood donors. In addition, the initial cell alignment in the low concentration region of the chemoattractant gradient in the type II device was believed to have caused the significantly higher CI but lower cell speed, and different cell speeds between the fMLP gradient and the COPD sputum gradient compared to the type I device without initial cell alignment.

CONCLUSION

In this study, we demonstrated a simple, effective and robust all-on-chip method for testing neutrophil chemotaxis. Previously, we used a similar microfluidic device to study neutrophil chemotaxis to sputum from COPD patients using both traditional microscopy-based method and a portable imaging system^8,21. However, the traditional cell preparation method significantly limits the efficiency of neutrophil chemotaxis experiment and makes it difficult to run the test in clinical settings. The developed all-on-chip method effectively removed this bottleneck. The previous cell chemotaxis analysis using the simple gradient generator is complicated by the uncontrolled initial cell positions in the gradient⁸, and we have recently shown that cell migration is sensitive to the cell’s current and previous gradient exposure²². To this end, the integrated cell-docking function in the type II device provides an effective solution and the cell migration distance analysis permits easy chemotaxis quantification without cell tracking. It will be interesting and useful to further compare the quantitative chemotaxis parameters between traditional cell isolation method and the on-chip method. Collectively, the all-on-chip method serves as a potentially useful tool for research and clinical applications that require rapid and accurate chemotaxis testing of untouched neutrophils.

Similar methods can be applied to other cell types upon availability of the corresponding magnetic cell isolation kits.

ACKNOWLEDGMENTS

This work is supported by Grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) and financial assistance from CMC Microsystems. F.L. thanks the Winnipeg Rh Institute Foundation and the University of Manitoba for a Rh Award; and J.W. thanks Research Manitoba for graduate fellowships. We thank the Nano-Systems Fabrication Laboratory at the University of Manitoba for their technical support. We also thank the Clinical Institute of Applied Research and Education at the Victoria General Hospital in Winnipeg for helping with the management of blood samples from healthy donors. We thank Dr. Xueling Cui for her suggestions in the Giemsa staining experiment.