Portable resistive pulse-activated lens-free cell imaging system

Abstract

We demonstrate a portable lens-free cell imaging system activated by resistive pulses that are induced by biological cells in a microfluidic channel. This imaging on-chip flow cytometer integrates CMOS with microfluidics as a standalone on-chip analytical device. The preliminary version of this system is able to provide dual-parametric study (light contrast image and modulated electrical pulse) for each biological cell. Red blood cells and a tumor cell line (HepG2) are used to demonstrate the performance of this chip for flow analysis of biological samples.

---

A microfluidic chip-based Coulter counter has been introduced as a low-cost and portable analytical platform for characterization of biological cells such as human blood cells,^1–3 and circulating tumor cells (CTCs).^4–6 However, the modulated pulses under the Coulter principle can provide very limited information on the biological sample compared to the conventional optical microscopic methods, which can provide straightforward visualization of the biological samples and therefore pervade many aspects of modern biomedicine and bioscience. Moreover, the implementation cost of conventional optical microscopy is very high, which has limited its applications in point-of-care diagnosis.^7–18 With the current rapid development of biomedical technologies, there is a growing need to develop compact, cost-effective and portable imaging systems for rapid and reliable point-of-care diagnosis, particularly in personalized healthcare. For this purpose, on-chip optofluidic microscopy has recently been introduced as an innovative and powerful platform based on the semiconductor technologies, which is able to image human blood cells or tumor cells with high resolution.¹⁹ Moreover, this on-chip optofluidic microscope requires continuous imaging on the microfluidic chip, even in the absence of cell events, which produces a huge amount of image data that may overload and slow down the signal or image processing. It is also quite challenging for the microcontroller and memory storage of this device to handle such big data.

In this short report, we demonstrate a portable resistive pulse-activated lens-free cell imaging system to address the abovementioned challenge of the big data issue involved in imaging on-chip flow cytometry. As shown in Fig. 1, the chip consists of a microfluidic channel (2 mm in length, 200 μm in width, 60 μm in height) and a complementary metal-oxide-semiconductor (CMOS) imaging sensor on the chip substrate. The microfluidic channel comprises a constricted aperture (30 μm in length, 30 μm in width, 30 μm in height), whereas the CMOS working pixels are located near the downstream exit of the sensing aperture. When a microparticle is transported through the aperture that is subjected to an applied voltage bias, the electric current through the microchannel decreases because the non-conducting particle replaces the conducting electrolyte solution resulting in a significant resistance increase. A threshold current (2% modulation to the baseline current) was set to detect this pulse and activate the imaging component (Fig. 1); when the current returns to a value above this threshold, it indicates that the particle is leaving the aperture, and a microcontroller will generate a signal to activate the CMOS sensor that is working in snapshot mode to capture the image of this particle. In this way, both the particle modulated pulse profile and image can be concurrently recorded by the system, which can be used for correlation analysis between peak amplitude and pixel size of biological cells.

Fig. 1 The working principle of the system: the imaging sensor is off when no resistive pulse is detected, whereas it is on when a resistive pulse is detected.

The mold of the polydimethylsiloxane (PDMS) microfluidic channel was fabricated on a glass slide, which was cleaned in acetone, methanol, and deionized water, and dried on a hotplate for 30 min at 250 °C. SU-8 25 (Microchem, USA) was spin-coated on a clean glass slide (at 2000 rpm), soft-baked for 5 min at 65 °C and 15 min at 95 °C, and exposed to UV irradiation. The glass slide was then baked on a hot plate for 1 min at 65 °C and 5 min at 95 °C. After post-baking, the mold was developed in SU-8 developer solution for 5 min, and baked on a hot plate for 20 min at 250 °C. PDMS pre-polymer and curing agents were mixed in a ratio of 10 : 1 and degassed in a vacuum chamber. The mixture was cast on the SU-8 mold and cured in an oven for 2 h at 95 °C. PDMS channels were then sliced and peeled off from the SU-8 mold and reservoir holes were punched. The PDMS channel and a clean substrate with a CMOS image sensor were plasma-treated for 10 s before they were bonded to form the final chip.

A tumor cell line, HepG2 cells (American Type Culture Collection, MD, USA), were cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), penicillin (100 μg mL⁻¹), 1 mM sodium pyruvate, and 0.1 mM MEM non-essential amino acids. The cells were grown at 37 °C under 5% CO₂ in a T75 flask. The red blood cells (RBCs) were separated from the whole blood using OptiPrep™ density gradient medium. Specifically, the whole blood was mixed with OptiPrep at a ratio of 1 : 8. Tricine-buffered saline (TBS: 0.85% NaCl, 20 mM Tricine–NaOH, pH 7.4) was layered at the top of the mixture for protecting it from mixing with the blood. The whole blood sample was centrifuged at 2000g for 15 min. RBCs were collected from the bottom of the centrifuge tube. The acquired samples were washed three times with 1× PBS for 10 min and resuspended in fresh PBS solutions (2 × 10⁶ mL⁻¹).

A power management circuit in the system can provide the voltage bias for the continuous electrokinetic flow in the chip with two electrodes submerged in the inlet and outlet chamber. The current modulation in the microfluidic circuit is monitored by a high-speed precision current-sensor amplifier and a low-pass filter. The output signal is collected and digitalized by an ADC chip, which is controlled by a central controller field programmable gate array (FPGA, Altera, Cyclone II, USA). Once a pulse signal is detected, the FPGA will produce a signal to activate the CMOS image sensor to capture the cell image. The CMOS image sensor (Aptina MT9M032, San Jose, CA, USA) can image an area as small as 11.5 mm (W) × 11.5 mm (L) × 2.3 mm (H) with a pixel size of 2.2 × 2.2 μm and a full-resolution array size of 1472 (H) × 1096 (V). The frame rate of the image sensor is 60 frames per second (fps) at a resolution of 1280 (H) × 720 (V). The frame rate can be further increased if the region of interest (ROI) to readout only covers part of the entire sensor array. In this study, only a small area on the CMOS sensor, located near the exit of the sensing aperture, is needed for cell imaging when it is exiting the aperture. To ensure optimal image contrast, the cover glass of the CMOS sensor is carefully removed using a razor blade such that the distance between the microfluidic channel and the sensor array is minimized. The schematic diagram and a home-made prototype of the proposed system are shown in Fig. 2.

Fig. 2 (a) The schematic structure of the proposed imaging on-chip flow cytometry: the FPGA is the central microcontroller for signal detection and processing; the analog module on the left is to continuously monitor the resistive pulse induced by biological cells; the CMOS imaging sensor module on the right is activated by the FPGA after a resistive pulse is detected by the analog module; (b) a prototype of the imaging on-chip flow cytometer consisting of FPGA controller, current detection module, and CMOS imaging module. Two on-board electrodes are used to provide continuous electrokinetic flow in the microfluidic chip.

Monodisperse polystyrene microbeads of 6.3 μm and 15.6 μm in diameter were used to calibrate this imaging flow system. The microbead suspensions were stripped from their original buffer through multiple dilution steps, and suspended in 1× PBS solution. Bovine serum albumin (BSA) was added to the solution to prevent the beads from forming agglomerates. RBC and HepG2 cells were harvested and resuspended in 1× PBS solution, supplemented with BSA to prevent cell aggregation.

Fig. 3 summarizes the experimental results for polystyrene particles of two different sizes, RBCs, and HepG2 cells with this resistive pulse-activated lens-free imaging flow system. The counting rate of particles or cells is about 100 events per min. The 2nd and 3rd columns in Fig. 3 demonstrate the comparison between the optical contrast images of the particles and cells obtained by a regular optical microscope and the CMOS sensor on this imaging on-chip flow cytometer. The particle or cell diameter can be calculated based on the number of pixels in the image and the actual size of each pixel of the CMOS sensor (4th column in Fig. 3). These results indicate that, for these polystyrene particles, the measured sizes are consistent with the nominal average size as specified by the manufacturer. The peak amplitude of the resistive pulses corresponding to each particle/cell event is also recorded in the system (5th column in Fig. 3). The 6.3 μm particles induce average pulse amplitude of 3.13 μA, whereas the 15.6 μm particles induce an average pulse amplitude of 19.57 μA. For testing biological cells, RBCs exhibit average pulse amplitude of 2.56 μA and diameter of 6.6 μm, whereas HepG2 cells show an average pulse amplitude of 20.76 μA and diameter of 22 μm. The bandwidth of the pulses reveals that each cell passes through the aperture within about 100 ms.

Fig. 3 The images of 6.3 μm and 15.6 μm polystyrene particles, a RBC and a HepG2 cell captured by the imaging on-chip flow cytometer, compared to the contrast image obtained by an optical microscope. The diameter is calculated based on the CMOS pixel size (2.2 × 2.2 μm) and the image pixel size. The pulse amplitude is recorded by the analog module corresponding to each cell/particle event.

RBCs and HepG2 cells were spiked and mixed in the same suspension to demonstrate the capability of this system to distinguish tumor cells from RBCs. Fig. 4 shows a series of cell events during a short time, recording the resistive pulses and the corresponding cell images activated by each pulse. The pulses of smaller amplitude represent the RBCs, which is concurrently verified by the image pixel size. The pulses of higher amplitude represent the HepG2 cells with a significantly larger pixel size. Although both RBCs and HepG2 cells show some size variation, these two different subpopulations can be clearly distinguished based on the image size and resistive pulse amplitude in Fig. 4.

Fig. 4 The recorded resistive pulses induced by several consecutive cell events, indicating the detection of a total of 9 biological cells (4 RBCs and 5 HepG2 cells); each resistive pulse activates a cell image taken by the on-chip CMOS imaging module.

In summary, a resistive pulse-activated lens-free imaging on-chip flow cytometer has been developed to record both the resistive pulse induced by a single biological cell and its optical contrast image. This proof-of-concept study showed that HepG2 tumor cells can be easily distinguished from red blood cells using this system. As a major advantage, the imaging module is at rest when there is no cell detected, and is activated only by resistive pulses induced by each cell event, thereby significantly reducing the image data size and improving the signal and image processing speed. This on-demand automatic imaging strategy could be applied to develop portable analytical devices for haematological studies and other biomedical applications. The current prototype, however, is limited by low image resolution and system throughput compared to a commercial imaging flow cytometer. More importantly, the capability of fluorescence imaging is still lacking in this system. Future work could further improve the image resolution by advanced image processing algorithms and the system throughput using multiple channel arrays, and integrate more advanced optical sensors to realize fluorescence imaging with this on-chip flow cytometer.

Notes and references

1. W. Asghar, Y. Wan, A. Ilyas, R. Bachoo, Y. T. Kim and S. M. Iqbal, Lab Chip, 2012, 12, 2345.
2. X. Wu, Q. Chen, Y. Sun and X. Fan, Appl. Phys. Lett., 2013, 102, 203706.
3. Q. Chen, H. Liu, W. Lee, Y. Sun, D. Zhu, H. Pei, C. Fan and X. Fan, Lab Chip, 2013, 13, 3351.
4. J. Guo, T. S. Pui, A. R. R. Rahman and Y. Kang, IEEE Trans. Biomed. Eng., 2013, 60, 3269.
5. W. Xu, R. Mezencev, B. Kim, L. Wang, J. McDonald and T. Sulcheck, PLoS One, 2012, 7, e46609.
6. S. Byun, S. Son, D. Amodei, N. Cermak, J. Shaw, J. H. Kang, V. C. Hecht, M. M. Winslow, T. Jacks, P. Mallick and S. R. Manalis, Proc. Natl. Acad. Sci. U. S. A., 2013, 110(19), 7580.
7. X. Heng, D. Erickson, L. Baugh, Z. Yaqoob, P. Sternberg, D. Psaltis and C. Yang, Lab Chip, 2006, 6, 1274.
8. W. Xu, M. Jericho, I. Meinertzhagen and H. Kreuzer, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 11301.
9. J. Garcia-Sucerquia, W. Xu, M. Jericho and H. Kreuzer, Opt. Lett., 2006, 31, 1211.
10. X. Cui, L. Lee, X. Heng, W. Zhong, P. Sternberg, D. Psaltis and C. Yang, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 10670.
11. N. Lindquist, A. Lesuffleur, H. Im and S. Oh, Lab Chip, 2009, 9, 382.
12. S. Seo, T. Su, D. Tseng, A. Erlinger and A. Ozcan, Lab Chip, 2009, 9, 777.
13. E. Schonbrun, W. Ye and K. Crozier, Opt. Lett., 2009, 34, 2228.
14. G. Zheng, X. Cui and C. Yang, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 9043.
15. E. Schonbrun, A. Abate and P. Steinvurzel, Lab Chip, 2010, 10, 852.
16. G. Zheng and C. Yang, Opt. Lett., 2010, 35, 2636.
17. D. Breslauer, R. Maamari, N. Switz, W. Lam, D. Fletcher and M. Pai, PLoS One, 2009, 4, e6320.
18. J. Wu, X. Cui, G. Zheng, Y. M. Wang, L. M. Lee and C. Yang, Opt. Lett., 2010, 35, 2188.
19. G. Zheng, S. A. Lee, S. Yang and C. Yang, Lab Chip, 2010, 10, 3125.

---