Subscriber access provided by KING MONGKUT'S UNIVERSITY OF TECHNOLOGY THONBURI (UniNet)

Article

All-in-One Centrifugal Microfluidic Device for Sizeselective Circulating Tumor Cell Isolation with High Purity Ada Lee, Juhee Park, Minji Lim, Vijaya Sunkara, Shine Young Kim, Gwang Ha Kim, Mi-Hyun Kim, and Yoon-Kyoung Cho Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 15 Oct 2014 Downloaded from http://pubs.acs.org on October 16, 2014

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

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

Page 1 of 12

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

Analytical Chemistry

All-in-One Centrifugal Microfluidic Device for Size-selective Circulating Tumor Cell Isolation with High Purity Ada Lee,† Juhee Park,† Minji Lim,† Vijaya Sunkara,† Shine Young Kim,‡ Gwang Ha Kim,§ Mi-Hyun Kim,§ and Yoon-Kyoung Cho*†ǁ †

Department of Biomedical Engineering, Ulsan National Institute of Science and Technology (UNIST), UNIST-gil 50, Ulsan, 689-798, Republic of Korea ‡ Department of Clinical Laboratory Medicine, Pusan National University Hospital, 179, Gudeok-ro, Seo-Gu, Busan, 602739, Republic of Korea. §

Department of Internal Medicine, Pusan National University School of Medicine and Biomedical Research Institute, Pusan National University Hospital, 179, Gudeok-ro, Seo-Gu, Busan, 602-739, Republic of Korea. ǁ

Center for Soft and Living Matter, Institute for Basic Science (IBS), UNIST-gil 50, Ulsan 689-798, Republic of Korea

ABSTRACT: Circulating tumor cells (CTCs) have gained increasing attention owing to their roles in cancer recurrence and progression. Due to the rarity of CTCs in the bloodstream, an enrichment process is essential for effective target cell characterization. However, in a typical pressure-driven microfluidic system, the enrichment process generally requires complicated equipment and long processing times. Furthermore, the commonly used immunoaffinity-based positive selection method is limited, as its recovery rate relies on EpCAM expression of target CTCs, which shows heterogeneity among cell types. Here, we propose a centrifugal force-based size-selective CTC isolation platform that can isolate and enumerate CTCs from whole blood within 30 s with high purity. The device was validated using the MCF-7 breast cancer cell line spiked in phosphate-buffered saline (PBS) and whole blood, and an average capture efficiency of 61% was achieved, which is typical for size-based filtration. The capture efficiency for whole blood samples varied from 44% to 84% under various flow conditions and dilution factors. Under the optimized operating conditions, a few hundred white blood cells (WBCs) per milliliter of whole blood were captured, representing a 20-fold decrease compared to those obtained using a commercialized size-based CTC isolation device. In clinical validation, normalized CTC counts varied from 10 to 60 per 7.5 mL of blood from gastric and lung cancer patients, yielding a detection rate of 50% and 38%, respectively. Overall, our CTC-isolation device enables rapid and label-free isolation of CTCs with high purity, which should greatly improve downstream molecular analyses of captured CTCs.

Metastasis, the spread of cancer, is the major cause of cancerrelated mortality.1 During the metastatic process, primary tumor cells, as well as pre-existing metastatic tumor cells, experience a series of steps to spread the disease from its original residing site to distant organs of the human body. Initially, tumor cells disseminated from solid tumors undergo a process known as epithelial-to-mesenchymal transition in order to achieve their migratory and invasive properties.2,3 These tumor cells, termed circulating tumor cells (CTCs), then enter the peripheral blood by a process called intravasation and move along the bloodstream until they reach a favorable secondary site. CTCs in the peripheral blood of cancer patients can serve as excellent diagnostic tools and prognostic markers. Therefore, quantification and characterization of these malignant cells can provide important clinical information for patients with metastatic cancer, thereby offering potential to design effective and individualized cancer therapies. Thus, it is important for clinicians to have reliable diagnostic tools to detect CTCs in order to improve clinical outcomes. One major hurdle for CTC studies is their extreme rarity in the whole blood (1 – 10 cells per 106 – 107 hematologic cells),

and the blood volume that can be processed is limited.4 Because of the rare occurrence of CTCs in nature, development of an effective enrichment process has become a critical step for detecting and characterizing CTCs down to a frequency of 1 – 10 cells per milliliter of blood. In this regard, the microfluidic platform has emerged as a possible solution because the microfabricated structures of the device allow for precise fluid control, and can also provide a biocompatible environment for cells.5 Although a variety of CTC detection platforms using microfluidic technology have been reported to date, these have some major limitations. Thus far, proposed microfluidic chipbased CTC isolation techniques have either relied on the biological properties (i.e. surface marker expression) or physical properties (i.e. size, density, dielectric properties, etc.) of cancer cells.6 Immunoaffinity-based methods7-10 and immunomagnetism-based methods11,12 which exploit surface markers of tumor cells, are two popular strategies used for CTC isolation. The capture efficiency of these methods heavily relies on the expression level of the prominent CTC marker, epithelial cell adhesion molecule (EpCAM). In reality,

1

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 2 of 12

Figure 1. Detailed illustrations showing the design and working principle of the CTC-isolation disc. (a) Expanded view of the CTCisolation disc showing top, middle, body, and bottom polycarbonate (PC) layers (from top to bottom), and double-sided pressure sensitive adhesive tapes sandwiched between the PC layers. The track-etched PC membranes were integrated on the reverse side of the body layer at the carved membrane insertion site. (b) The top view of the CTC-isolation disc showing detailed microfluidic features. The device is composed of three individual filtration units; each unit contains a sample loading chamber, filtration chamber, waste chamber, ventilation chambers, and connected channels. (c) Cross-sectional view of the CTC-isolation disc. The track-etched PC membranes are tightly sealed by a chemical bonding method. (d) A photograph of the fabricated lab-on-a-disc. (e) The schematic illustration showing the working principle of the CTC-isolation disc. Operation images of the CTC-isolation disc were taken (f, g) during filtration, (h) after initial washing, and (i) after staining.

however, EpCAM expression is highly heterogeneous in different types of tumor cells, and is even absent in some tumor types that are non-epithelial in origin.13 Moreover, since the premise of this technique is based on the assumption of sufficient collision frequency between the cell and the antibody-coated substrate, it is critical to have a flow rate that is low enough to maximize the chance of the antigen-antibody interaction. Because of this limitation, the processing time required for immunoaffinity-based detection methods may exceed several hours for analysis of 7.5 mL of patient blood.7 Such limitations of low sensitivity and long processing time of immunoaffinity-based CTC isolation methods can be overcome to a certain degree by taking advantage of differences in the physical properties of target cells from hematologic cells. For example, Moon et al. demonstrated an

effective breast cancer cell isolation process from whole blood by using a dielectrophoretic separation method.14 Densitybased separation is another blood cell separation method that could be applied to CTC isolation. Using a density-gradient medium, it is possible to generate layered separation of cell types based on cellular density.15 However, density difference between CTCs and leukocytes are not big and therefore additional steps are often necessary to increase the purity of the final product. Deterministic lateral displacement (DLD)-based separation method and hydrodynamic force-driven filtration method are two representative size-selective on-chip CTC separation methods. The primary assumption underlying these methods is that, in the fluidic system with low Reynolds number, the cells gently follow in streamlines.21 The DLD-based method uses

2

ACS Paragon Plus Environment

Page 3 of 12

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

Analytical Chemistry

microfabricated post arrays to separate a stream of fluid containing cells with different sizes. In the hydrodynamic force-driven method, a spiral shaped microfluidic channel that can be used to separate CTCs is proposed by Lim et al.22 The advantage of these platforms is that, compared to the immunoaffinity-based platforms, samples can be processed with higher flow rate. In addition, they do not require blood pre-processing steps (i.e. RBC lysis or hematologic cell fixation), yet still yields the final product with high capture efficiency and purity. The most straightforward approach for size-based CTC isolation is, perhaps, mechanical filtration. This approach takes advantage of the well-known characteristic that CTCs are larger than normal hematologic cells. Many groups have proposed CTC filtration platforms using fabricated throughhole membranes designed to restrict the passage of cells smaller than the “critical size,” which is reported to be in the rage of 5 – 8 µm.16,17 Indeed, various types of materials have been used to separate CTCs based on their size. Lin et al. developed a CTC isolation platform using a parylene membrane,8 and Adams et al. developed an SU-8 through-hole membrane with circular micropore arrays.18 High purity is important in order to perform genotype analyses of CTCs after isolation to reduce analytical variation caused by interference of neighboring hematologic cells. Sajay et al. and Hosokawa et al. introduced filters with a rectangular micro-slit array, which they claimed yielded higher purity.19,20 However, these platforms may require the use of several fluidic connectors which may lead to frequent sample losses. In addition, they can be inefficient for the detection and characterization of captured cells because device disassembly is required in order to withdraw the membrane for the subsequent enumeration process.18,23,24 In this report, we present a centrifugal force-based sizeselective CTC detection system for the rapid and label-free isolation of CTCs with high purity (Figure 1). The lab-on-adisc platform utilizes centrifugal force to rapidly transfer unprocessed whole blood samples from one chamber to another.25 Compared to the previously mentioned membranebased microfluidic systems for CTC detection, the lab-on-adisc system can offer more efficient fluid control because it does not require external interconnectors. Instead, only a simple rotary motor is needed to actuate the fluid flow. Such advantages of the lab-on-a-disc system allow for reduced manual handling steps between the filtration, staining, and detection processes. To selectively isolate CTCs based on size differences between CTCs and the surrounding hematologic cells, we integrated a commercially available track-etched polycarbonate (PC) membrane filter on a lab-on-a-disc system. We tested the performance of our novel CTC-isolation disc using a breast cancer cell line spiked in both phosphate buffered saline (PBS) and whole blood. We also conducted a clinical test for the isolation and detection of CTCs from patients with lung cancer and gastric cancer.

EXPERIMENTAL SECTION Cell Culture and Sample Preparation. The MCF-7 breast cancer cell line, AGS gastric cancer cell line, and NCI-H460 non-small cell lung cancer cell line were purchased from American Type Culture Collection (Manassas, VA, USA). All cell lines were cultivated in RPMI culture medium

supplemented with 5% fetal bovine serum (FBS) and 1% antibiotics/antimycotics, under humidified conditions at 37°C with 5% CO2. Prior to each experiment, cells grown to confluence were trypsinized by Cell Stripper (Cellgro; Manassas, VA, USA) for 1 min and resuspended in RPMI culture medium. For harvesting cells, the centrifugation condition for live cancer cell lines was fixed to 100 g in order to minimize cell damages. For blood sample preparation, peripheral blood samples from healthy donors and cancer patients were obtained from Pusan National University Hospital (Busan, Korea). This study was reviewed and approved by the Institutional Review Board (IRB) of the Pusan National University Hospital. Written informed consent was obtained from all patients. Samples were collected in vacutainer tubes (BD Vacutainer; Franklin Lakes, NJ, USA) with ethylene diamine tetraacetic acid (EDTA) to prevent coagulation. Collected blood samples from healthy donors were used for experiments within 12 h. For clinical studies, 2 – 4 mL blood samples were collected from 10 lung cancer patients and 13 gastric cancer patients and used for experiments within 6 h. All blood samples were processed without any pre-treatment processes such as RBC removal or blood cell fixation. For spike-in experiments, exact cell counts of 1 – 1000 cancer cells were obtained using the bright-field channel of an inverted microscope (IX71; Olympus Corp.; Tokyo, Japan) and retrieved by trypsinization for each experiment. Retrieved cells were directly added to 1 mL of PBS or whole blood samples, which were pre-loaded into loading chambers of the CTC-isolation disc. Device Fabrication. A lab-on-a-disc for CTC isolation was designed with SolidWorks CAD design software (Waltham, MA, USA) as shown in Figure 1b. The device is composed of three individual filtration units, enabling simultaneous processing of three different blood samples. Each unit contains a sample loading chamber, filtration chamber, waste chamber, ventilation chambers, and channels connected to the chambers. General fabrication processes of the lab-on-a-disc are described in detail in previous reports.26,27 Briefly, a computer numerical control micromachining apparatus (M&I CNC lab; Kyunggi, Korea) was employed to create designed features on PC plates, and a cutting plotter (Graphtec CE300-60; Graphtec Corporation; Yokohama, Japan) was used to generate features on double-sided pressure sensitive adhesive tapes that were sandwiched between the PC layers to achieve a tight assembly of the device. Specifically, for our CTC-isolation disc, four separate layers of PC plates were established to make up a single device: a top layer, middle layer, body layer, and bottom layer. Holes were created on the top layer for sample injection and ventilation. The middle layer was used for the creation of the inlet channel, and main fluidic chambers were created on the body layer (Figure 1a). The reverse side of the filtration chamber was carved for the creation of an outlet channel and the insertion of a commercially available tracketched PC membrane (Whatman; Florham Park, NJ, USA). Specifically, hydrophilic track-etched PC membranes with a pore size of 8 µm, a diameter of 13 mm, and a thickness of ~10 µm were used for all experiments. The pores size of 8 µm was chosen according to previous reports.16,17 To achieve a leak-free and wrinkle-free sealing of the track-etched PC membrane, a chemical bonding method was used (Figure 1c). Specifically, a drop of acetone was directly applied onto a

3

ACS Paragon Plus Environment

Analytical Chemistry

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

bonding area. Prior to CTC isolation processes, the fabricated CTC-isolation disc was incubated with PBS supplemented with 1% bovine serum albumin (BSA) for 2 h in order to prevent non-specific adsorption of cells. A detailed optimization process for the surface passivation of the PC substrates is described in Figure S1. Figure 1d shows a photograph of the fabricated lab-on-a-disc for a size-selective CTC isolation. Experimental Setup and Device Operation. Figure 1e represents a schematic illustration showing the working principle of the CTC-isolation disc. When the disc rotates, the centrifugal force drives a blood sample through the isolation chamber where target CTCs are trapped on a membrane by size selectivity. Blood cells that are smaller than the size of pores are passed through the membrane and move to the waste chamber. As such, our CTC-isolation disc was operated via centrifugal force in a programmable manner using an operation system described in previous reports.28,29 In brief, a programmable spinning motor was used for application of the centrifugal force, and the use of an imaging module with a charge-coupled device digital camera (Toshiba Corp.; Tokyo, Japan) and a strobe lamp (B&B Corp.; Seoul, Korea) allowed for a real-time visualization of the rotating disc. Determination of Pressure Drop Across the Tracketched Membrane. When dealing with fluids that contain fragile cells, it is critical that the pressure drop, ∆P, across the porous membrane is kept below a threshold value to prevent cell damage. The pressure drop across the membrane in the CTC-isolation disc was calculated by using the following equation30: 128‫ ܮ‬24 ܳ + ସ൨ ߂ܲ = ߤ ൤ ߨ݀ ସ ݀ ܰ where µ is the coefficient of dynamic viscosity, L is the thickness of the membrane, d is the pore diameter, Q is the flow rate, and N is the number of pores. In order to determine the flow rate Q of the above equation, we utilized the previously mentioned operating system with a high-resolution imaging module. Specifically, digital images were acquired while the disc was spun at a specific spin speed. Acquired images were analyzed to determine the volume reduction in the fluidic chamber between each consecutive time point. CTC Analysis by Fluorescence Microscopy. An immunostaining method was used for the analysis of captured cells. For the identification and enumeration of captured cancer cell lines and CTCs, predefined CTC determination criteria- DAPI positive, cytokeratin positive, and CD45 negative -were used. Prior to the staining process, cells were first blocked with 500 µg/mL of Fc blocker (Human IgG; R&D Systems; Minneapolis, MN, USA) for 15 min at room temperature, followed by a washing step with 0.5 % BSA in PBS. Fc blocker was used to block non-specific binding of staining dyes. The cells were then fixed with 4 % paraformaldehyde for 15 min followed by a permeabilization step with 0.1 % Triton X-100 for 5 min. The fixed cells were then washed with 0.5% BSA in PBS. Finally, cells were incubated with a staining solution cocktail containing 4′6diamidino-2-phenylindole (DAPI; Sigma-Aldrich; St. Louis, MO, USA), cytokeratin (eBioscience; San Diego, CA, USA), and CD45 (BD Bioscience; San Jose, CA, USA). Optimized concentrations for each staining solution were determined as

Page 4 of 12

follows: 100 ng/mL of DAPI, 8 µg/mL of Anti-PanCytokeratin-eFluor® 615, 240 ng/mL of Anti-Cytokeratin-PE, and 4 µg/mL of Human CD45-FITC. Reagents used for fixation, permeabilization, and staining were added through the sample loading chamber. By instantaneously rotating the disc for 0.5 sec, it was possible to completely fill up the space above the track-etched PC membrane with each reagent. To test the viability of captured cells, a live/dead assay was performed according to the manufacturer’s protocol (LIVE/DEAD® Viability/Cytotoxicity Kit; Invitrogen; Carlsbad, CA, USA). After the immunostaining process, the whole device was mounted on a fluorescence inverted microscope (IX71, Olympus Corp.; Tokyo, Japan). Isolated cells were analyzed and enumerated using image analysis software (MetaMorph; Molecular Devices; Sunnyvale, CA, USA).

RESULTS AND DISCUSSION Isolation of Cancer Cells Using a Rotating Device. For a typical microfluidic system setup, complex peripheral apparatus such as syringe pumps, tubing, and connectors are mandatory components. In the case of our lab-on-a-disc system, however, only a single motor was required to process samples up to 3 mL without requiring any extra connectors or tubing. This provides great potential for establishing a rapid, accurate, and cost-effective CTC detection technology, which is critical for rare cell-based diagnostics. In this study, the entire process of CTC isolation, staining, and detection was conducted on a disc, using a programmable operation system. The operation procedure is summarized in Table 1, and images showing the sample filtration process and staining process is shown in Figures 1f – 1i. The 20 sec of spinning was long enough to filter 3 mL of whole blood without significant RBC sedimentation in the radial direction. Table 1. Operation program of the lab-on-a-disc system for the rapid isolation and detections of CTCs no. 1 2

operation sample filtration washing

G-force

fluidic state flow

(rpm) 254 g* (2400 rpm) 64 g

flow

volume

time

3 mL

20 sec

1 mL

3 sec

250 µL

15 min

1 mL

3 sec

23 sec

(1200 rpm) 3

Fc blocking

incubation

4

washing

flow

64 g (1200 rpm)

5 6 7

fixation permeabilization washing

incubation

-

250 µL

15 min

incubation

-

250 µL

5 min

1 mL

3 sec

250 µL

15 min

1 mL

3 sec

64 g

flow

50.2 min

(1200 rpm) 8

staining

incubation

9

washing

flow

64 g (1200 rpm)

Total Processing Time

50.5 min

*Spin speed from 1200 rpm to 3600 rpm (corresponding G-force from 64 g to 570 g, respectively) was used for the samples (PBS, whole blood from healthy donors, and whole blood from patients) tested in this study.

4

ACS Paragon Plus Environment

Page 5 of 12

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

Analytical Chemistry

The range of G-force used in all the experiments in this study was restricted to 16 g (600 rpm), 64 g (1200 rpm), 254 g (2400 rpm), and 570 g (3600 rpm), which represent conditions in which cells should not experience significant damage.31 The G-force of the chosen rpm conditions was calculated based on the assumption that the captured cells were at the farthest outer edge of the filtration chamber (Figure 1b). We also determined the pressure drop across the membrane in order to ensure that the pressure difference across the membrane generated by the rotation of the disc is comparable to that obtained in previous reports.20,30 The average pressure drop generated by our CTC-isolation disc was 4.8 mbar, 10.36 mbar, and 18.4 bar at spin speeds of 1200 rpm, 2400 rpm, and 3600 rpm, respectively. These values fall within the pressure drop range obtained from a track-etched membrane-based platform reported by Coumans et al., where the capture efficiency reached a maximum at 10 mbar.30 For the spike-in experiments for PBS samples, 1 mL of the cell suspension (cell numbers ranging from 1 to 962) was injected into the loading chamber. The sample was then transferred to the filtration chamber by spinning the disc at a spin speed of 2400 rpm, followed by a subsequent washing step with a spin speed of 1200 rpm. Only 20 s were required to

Figure 2. Cancer cell isolation using the MCF-7 breast cancer cell line in PBS. (a) Linearity of MCF-7 cell capture efficiency at various cell concentrations. (b) Fluorescent images of isolated MCF-7 cells. MCF-7 cells were stained with DAPI and PE labeled cytokeratins.

process 3 mL of the sample, which is a significantly reduced time compared to previously reported methods. Prior to the staining of captured cells, the initial process included direct injection of Fc blocker into the filtration chamber. After 14 min of incubation, cells were fixed for 15 min followed by a permeabilization step for 5 min. Finally, a cocktail of staining solution containing DAPI (for nuclear staining), PE-labeled cytokeratin (a marker for CTCs) was injected to the filtration chamber and then incubated for 15 min. PBS washing was performed in between each step, and the spin speed used for washing was 1200 rpm. The total processing time from sample injection to staining was less than 1h. In terms of the performance of the device, the overall capture efficiency of the MCF-7 cells in PBS was 61% (Figure 2a). In addition, the results showed high linearity with a correlation coefficient (R2) higher than 0.99 (Figure 2a). The captured MCF-7 cells were stained with DAPI and PE-labeled cytokeratins after the isolation process, and morphology of the cells was investigated via fluorescence microscopy. Figure 2b shows circular but slightly flattened MCF-7 cells with strong DAPI and cytokeratin signals. All the isolated cells were carefully inspected using pre-defined CTC determination criteria, and none of the cells showed a ruptured shape. Since we used “deformable” live cancer cells during the filtration process, it is possible that a loss in capture efficiency may have resulted from the escapes of cells through fused pores.18,32 Isolation of Cancer Cells in Whole Blood Samples. To mimic the process of CTC isolation from patient blood samples, the effectiveness of the operation conditions of the CTC-isolation disc system was confirmed by using MCF-7 cells spiked in whole blood from healthy donors. First, in order to investigate the effect of the spin speed on the capture efficiency, the samples were processed at four different conditions: 600 rpm, 1200 rpm, 2400 rpm, and 3600 rpm (Figure 3a). For the 600 rpm condition, the disc was rotated in clockwise and counter-clockwise manners repeatedly in order to induce a more efficient blood transfer. For each trial, approximately 100 MCF-7 cells were spiked into blood samples from a healthy donor. The average capture efficiencies for the 600 rpm, 1200 rpm, 2400 rpm, and 3600 rpm conditions were 84 ± 3% (n = 3), 51 ± 3% (n = 6), 50 ± 3% (n = 3), and 50 ± 5% (n = 3), respectively. The highest capture efficiency was observed at 600 rpm, showing a significant difference from that obtained under other spin conditions (P < 0.01, t-test). Within the range of 1200 to 3600 rpm, however, no significant difference was observed (P > 0.01, ANOVA). Average WBC counts for 600 rpm, 1200 rpm, 2400 rpm, and 3600 rpm were 3092 ± 286 cells (n = 3), 266 ± 51 cells (n = 6), 461 ± 61 cells (n = 6), and 181 ± 53 cells (n = 3), respectively. The largest number of WBCs was obtained with a spin speed of 600 rpm, showing a significant difference compared to that in the other spin conditions (P < 0.01, t-test). However, within the range of 1200 to 3600 rpm, WBC counts decreased to very low numbers, with no significant difference among them (P > 0.01, ANOVA). This indicated that, although the spin speed of 600 rpm showed the highest capture efficiency, the purity was low and blood transfer was inefficient.

5

ACS Paragon Plus Environment

Analytical Chemistry

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

A live/dead assay was performed to ensure that all of the captured cells including CTCs and WBCs were viable under the processed spin conditions. The experiment was conducted at two extreme conditions (600 rpm and 3600 rpm), and the results were compared with those obtained using the commercialized ScreenCell® system. As a result, viability of approximately 98% was achieved for both the 600 rpm and 3600 rpm conditions (Figure S2a). In addition, a cell culture experiment was performed to further confirm the viability of captured MCF-7 cells. For this specific experiment, we used a separately designed “membrane-detachable” disc. Using the same operational conditions, MCF-7 cells which were isolated from whole blood were released from the disc, and cultured in RPMI containing culture plate for 8 days (Figure S2b). It was confirmed that using the CTC-isolation disc, it is possible to capture viable MCF-7 cells and culture them. Moreover, contaminated WBCs did not have any notable effect on the growth of MCF-7 cells. Therefore, we concluded that the spin conditions in the range of 1200 rpm and 3600 rpm were more reasonable for CTC isolation experiments. The default spin speed was set as 1200 rpm, but higher spin speeds of 2400

Page 6 of 12

rpm and 3600 rpm were used for viscous blood samples. Membrane clogging is a frequently encountered issue in most size-selective CTC isolation platforms. In order to avoid membrane clogging, RBCs can be removed by lysing the cells23,33 or by using density gradient centrifugation.24 However, pre-treatment of blood samples by RBC lysis or density gradient centrifugation may lead to significant loss of target cells. Dilution is an alternative blood pre-treatment method that is often used to avoid membrane clogging.30,33-35 Using the CTC-isolation disc, the effect of the dilution factor on the capture efficiency was studied (Figure 3b). 1 mL of whole blood was diluted with various amounts of PBS (0 ~ 3 mL). For each trial, approximately 100 MCF-7 cells were spiked into each blood sample, and each sample was processed at 2400 rpm. Average capture efficiencies for each condition were 53 ± 1%, 56 ± 2%, and 54 ± 4%, 55% ± 2%, respectively (n=3). Average WBC counts for each condition were 530 ± 40 cells, 512 ± 71 cells, 460 ± 144 cells, and 495 ± 83 cells respectively (n=3). These results imply that the dilution factor is not a critical component affecting either the capture efficiency or purity (P > 0.01, ANOVA).

Figure 3. Various conditions for cancer cell isolation in whole blood. (a) The effect of the spin speed on the capture efficiency of MCF-7 cells and white blood cell (WBC) counts. (b) The effect of the dilution factor on the capture efficiency of MCF-7 cells and WBC counts. (c) Capture efficiency of MCF-7 cells (filled circles) and WBC counts (open circles) at various cell concentrations from 0 to 140 cells per milliliter of whole blood. (d) Performance comparison between ScreenCell® and the CTC-isolation disc. For (a), (b), and (d), red bars represent the MCF-7 counts and gray bars represent WBC counts; asterisks (*) and double asterisks (**) indicate P < 0.01 and P > 0.01, respectively; error bars represent the standard error of the mean (s.e.m.).

6

ACS Paragon Plus Environment

Page 7 of 12

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

Analytical Chemistry

Figure 4. Results showing the performance of the CTC-isolation disc in (a) lung cancer patients and gastric cancer patient samples. The CTC counts in each samples is normalized to 7.5 mL, which is the input volume of blood used in Cellsearch® system. (b) and (c) are the representative fluorescent images of captured CTCs from lung cancer patients (L5 and L6), and gastric cancer patients (G4 and G5), respectively. Captured cells were stained with DAPI, PE labeled cytokeratins, and FITC labeled CD45.

Finally, the performance of the device using blood samples was validated by measuring the capture efficiencies at different concentrations of MCF-7 cells (Figure 3c). MCF-7 cells were spiked into whole blood at cell concentrations ranging from 0 to 136 cells per milliliter of whole blood, and each sample was processed at 2400 rpm. The number of contaminated WBCs was enumerated along with the captured input cells for each experiment. The result of the capture efficiency using the whole blood sample was 54% with a correlation coefficient (R2) of 0.98, which shows similarly high linearity compared to that obtained when the cells were spiked in PBS. The WBC counts varied from approximately 200 to 600 cells. Performance Comparison with ScreenCell® System. The spike-in experiment was repeated to compare the performance of the device with that of the ScreenCell® system. For experiments using the ScreenCell® Cyto device, whole blood samples were pre-processed according to the manufacturer’s protocol. Briefly, 1.3 mL of ScreenCell® dilution FC2 buffer was added to 1 mL of MCF-7 cell spiked blood samples. Mixed samples were homogenized by inverting the tubes multiple times followed by an 8 min incubation step. Diluted blood samples were subsequently processed. For the spike-in experiment with the CTC-isolation disc, 1 mL of MCF-7 cell spiked whole blood sample was directly used without any pre- processing step. The capture efficiencies of MCF-7 cells using the ScreenCell® and CTCisolation disc were 69 ± 6% (n = 3) and 56 ± 2% (n = 3), respectively (Figure 3d). The slight (but not statistically

significant; P > 0.01, t-test) difference in the capture efficiency may have resulted from the fact that the pore size of the tracketched PC membranes used in ScreenCell® is smaller (7.5 µm)23 than that used in the CTC-isolation disc (8 µm). In addition, the use of ScreenCell® dilution FC2 buffer which is known to fix the cells, likely also contributed to the higher capture efficiency obtained with the ScreenCell® system. The WBC counts for the ScreenCell® and CTC-isolation disc were 10,983 ± 606 cells (n = 3) and 552 ± 62 cells (n = 3), respectively. It is noteworthy that the number of WBCs captured with the ScreenCell® and CTC-isolation disc differed significantly (P < 0.01, t-test), yielding a 20-fold lower number of WBCs obtained with the CTC-isolation disc than with the ScreenCell® system. Clinical Evaluation. For clinical evaluation of the CTCisolation disc method, 10 lung cancer patients and 13 gastric cancer patients, and 5 healthy donors were enrolled in the study. Patient blood samples were processed within 6 h, and different volumes of whole blood samples ranging from 2.2 to 4.4 mL were directly injected in the device, without any sample pre-treatment process. Because most of the patient blood samples were not transferred at 1200 rpm du o their high viscosity, the spin speed for the filtration process was fixed to 2400 rpm except for samples L2, L7, G1, and G5, which had even higher viscosity. These patient samples were processed using 3600 rpm, assuming that this spin speed would not reduce the capture efficiency. The experiment for sample L7 was not successful because the RBCs from the sample clogged the membrane during the filtration process.

7

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 8 of 12

shows representative immunofluorescence images of CTCs and WBCs obtained from lung cancer patients and gastric cancer patients. It can be inferred from the images that the size of CTCs shows high variability. The size distribution of the isolated CTCs from cancer patients was compared with that of the corresponding cancer cell line by measuring the diameter of each cell via MetaMorph image analysis software (Figure 5). For both lung cancer and gastric cancer patients, the size of the cells ranged from 6 to 22 µm, and the median size of the CTCs from cancer patients was larger than that of the cancer cell lines.

CONCLUSION

Figure 5. Histograms showing the size distribution of CTCs isolated from (a) lung cancer patients (purple, solid), and the NCI-H460 cell line (hatched) and (b) gastric cancer patients (orange, solid) and AGS cancer cell line (hatched). The size of the cancer cells varied from 6 to 22 µm.

All other samples were processed without any technical difficulties. Figure 4 shows the plotted result for the performance of the CTC-isolation disc using clinical samples. All counts were normalized to 7.5 mL, which is the input volume used in the CellSearch® system (the only FDAapproved CTC detection device) in order to directly compare the number of captured CTCs in a fixed blood sample volume. The data of actual CTC counts and processed blood volume from lung cancer patients and gastric cancer patients are provided in Table S1. The actual number of isolated CTCs from lung cancer patients ranged from 5 to 21 CTCs, and that from gastric cancer patients ranged from 5 to 29 CTCs. In terms of the detection rates, CTCs were detected in 5 of 10 (50 %) lung cancer patients with lung cancer, and CTCs were detected in 5 of 13 (38.4 %) gastric cancer patients. Since the average capture efficiencies obtained from spike-in experiments is approximately 50 %, it is possible that a number of CTCs have either escaped through membrane pores, or were not properly immuno-stained due to a low CK expression level. Improvements can be made through future studies, by integrating a membrane with smaller and uniform pore size and developing CTC detection markers which are applicable to detect a broad range of CTCs. Figure 4b and 4c

In this work, we developed a size-selective lab-on-a-disc platform for the rapid and efficient detection of CTCs from whole blood. The main strength of the proposed CTC-capture disc is that it requires less than 30 s to process 3 mL of whole blood samples. Considering that most of the immunoaffinitybased CTC isolation platforms require at least a few hours to complete the sample processing procedure, our system significantly reduces the processing time. Another advantage of our platform is that, compared to membrane-based CTC isolation platforms introduced in previous works, the CTC isolation processes do not require a sample pre-treatment step, and it is still possible to acquire viable CTCs with high purity. Moreover, our system reduces unnecessary manual handling since it does not require device disassembly to release the membrane for the analysis of captured cells. It could be also possible to perform downstream molecular analysis by integrating the active valving system on the device. For the future direction, however, it is necessary that we enhance the capture efficiency by integrating the membrane with regularly arrayed and smaller pores. In addition, the development of more accurate CTC detection markers is necessary in order to achieve a higher detection rate in clinical settings. Overall, we have successfully developed an effective “all-in-one” CTC isolation device, from sample injection to staining, using whole blood without requiring any pre-processing step. Owing to these advantages, the CTC-isolation disc platform can become a highly competitive candidate in the field of cancer diagnosis.

ASSOCIATED CONTENT Supporting Information Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author *Email: [email protected]. Phone: +82-52-217-2511. Fax: +8252-217-2509

ACKNOWLEDGEMENTS This work was supported by the National Research Foundation (NRF) grant (2013R1A2A2A05004314) and a grant from the Korean Health Technology R&D Project, Ministry of Health & Welfare (A121994) funded by the Korean government. The

8

ACS Paragon Plus Environment

Page 9 of 12

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

Analytical Chemistry

biological specimens for this study were provided by the Pusan National University Hospital, a member of the National Biobank of Korea, which is supported by the Ministry of Health Welfare. All samples derived from the National Biobank of Korea were obtained with informed consent under institutional review board-approved protocols.

REFERENCES (1) Kaiser, J. Science 2010, 327, 1072-1074. (2) Schroeder, A.; Heller, D. A.; Winslow, M. M.; Dahlman, J. E.; Pratt, G. W.; Langer, R.; Jacks, T.; Anderson, D. G. Nat Rev Cancer 2012, 12, 39-50. (3) Mocellin, S.; Keilholz, U.; Rossi, C. R.; Nitti, D. Trends in Molecular Medicine 2006, 12, 130-139. (4) Hong B, Z. Y. Theranostics 2013, 3, 377-394. (5) van Midwoud, P. M.; Janse, A.; Merema, M. T.; Groothuis, G. M. M.; Verpoorte, E. Analytical Chemistry 2012, 84, 39383944. (6) den Toonder, J. Lab on a Chip 2011, 11, 375-377. (7) Nagrath, S.; Sequist, L. V.; Maheswaran, S.; Bell, D. W.; Irimia, D.; Ulkus, L.; Smith, M. R.; Kwak, E. L.; Digumarthy, S.; Muzikansky, A.; Ryan, P.; Balis, U. J.; Tompkins, R. G.; Haber, D. A.; Toner, M. Nature 2007, 450, 1235-1239. (8) Stott, S. L.; Hsu, C.-H.; Tsukrov, D. I.; Yu, M.; Miyamoto, D. T.; Waltman, B. A.; Rothenberg, S. M.; Shah, A. M.; Smas, M. E.; Korir, G. K.; Floyd, F. P.; Gilman, A. J.; Lord, J. B.; Winokur, D.; Springer, S.; Irimia, D.; Nagrath, S.; Sequist, L. V.; Lee, R. J.; Isselbacher, K. J.; Maheswaran, S.; Haber, D. A.; Toner, M. Proceedings of the National Academy of Sciences 2010. (9) Adams, A. A.; Okagbare, P. I.; Feng, J.; Hupert, M. L.; Patterson, D.; Göttert, J.; McCarley, R. L.; Nikitopoulos, D.; Murphy, M. C.; Soper, S. A. Journal of the American Chemical Society 2008, 130, 8633-8641. (10) Zhang, N.; Deng, Y.; Tai, Q.; Cheng, B.; Zhao, L.; Shen, Q.; He, R.; Hong, L.; Liu, W.; Guo, S.; Liu, K.; Tseng, H.-R.; Xiong, B.; Zhao, X.-Z. Advanced Materials 2012, 24, 27562760. (11) Hoshino, K.; Huang, Y.-Y.; Lane, N.; Huebschman, M.; Uhr, J. W.; Frenkel, E. P.; Zhang, X. Lab on a Chip 2011, 11, 3449-3457. (12) Kang, J. H.; Krause, S.; Tobin, H.; Mammoto, A.; Kanapathipillai, M.; Ingber, D. E. Lab on a Chip 2012, 12, 2175-2181. (13) Mikolajczyk, S. D.; Millar, L. S.; Tsinberg, P.; Coutts, S. M.; Zomorrodi, M.; Pham, T.; Bischoff, F. Z.; Pircher, T. J. Journal of Oncology 2011, 2011. (14) Moon, H.-S.; Kwon, K.; Kim, S.-I.; Han, H.; Sohn, J.; Lee, S.; Jung, H.-I. Lab on a Chip 2011, 11, 1118-1125. (15) Paterlini-Brechot, P.; Benali, N. L. Cancer Letters, 253, 180-204. (16) Kim, M. S.; Sim, T. S.; Kim, Y. J.; Kim, S. S.; Jeong, H.; Park, J.-M.; Moon, H.-S.; Kim, S. I.; Gurel, O.; Lee, S. S.; Lee, J.-G.; Park, J. C. Lab on a Chip 2012, 12, 2874-2880. (17) Coumans, F. A. W.; van Dalum, G.; Beck, M.; Terstappen, L. W. M. M. PLoS ONE 2013, 8, e61770. (18) Adams, D. L.; Zhu, P.; Makarova, O. V.; Martin, S. S.; Charpentier, M.; Chumsri, S.; Li, S.; Amstutz, P.; Tang, C.-M. RSC Advances 2014, 4, 4334-4342. (19) Sajay, B.; Chang, C.-P.; Ahmad, H.; Chung, W.; Puiu, P.; Rahman, A. Biomed Microdevices 2013, 15, 699-709.

(20) Hosokawa, M.; Yoshikawa, T.; Negishi, R.; Yoshino, T.; Koh, Y.; Kenmotsu, H.; Naito, T.; Takahashi, T.; Yamamoto, N.; Kikuhara, Y.; Kanbara, H.; Tanaka, T.; Yamaguchi, K.; Matsunaga, T. Analytical Chemistry 2013, 85, 5692-5698. (21) Liu, Z.; Huang, F.; Du, J.; Shu, W.; Feng, H.; Xu, X.; Chen, Y. Biomicrofluidics 2013, 7, -. (22) Hou, H. W.; Warkiani, M. E.; Khoo, B. L.; Li, Z. R.; Soo, R. A.; Tan, D. S.-W.; Lim, W.-T.; Han, J.; Bhagat, A. A. S.; Lim, C. T. Sci. Rep. 2013, 3. (23) DESITTER, I.; GUERROUAHEN, B. S.; BENALIFURET, N.; WECHSLER, J.; JÄNNE, P. A.; KUANG, Y.; YANAGITA, M.; WANG, L.; BERKOWITZ, J. A.; DISTEL, R. J.; CAYRE, Y. E. Anticancer Research 2011, 31, 427-441. (24) Kim, E. H.; Lee, J. K.; Kim, B. C.; Rhim, S. H.; Kim, J. W.; Kim, K. H.; Jung, S. M.; Park, P. S.; Park, H. C.; Lee, J.; Jeon, B. H. Analytical Biochemistry 2013, 440, 114-116. (25) Madou, M.; Zoval, J.; Jia, G.; Kido, H.; Kim, J.; Kim, N. Annual Review of Biomedical Engineering 2006, 8, 601-628. (26) Lee, B. S.; Lee, Y. U.; Kim, H.-S.; Kim, T.-H.; Park, J.; Lee, J.-G.; Kim, J.; Kim, H.; Lee, W. G.; Cho, Y.-K. Lab on a Chip 2011, 11, 70-78. (27) Kim, T.-H.; Park, J.; Kim, C.-J.; Cho, Y.-K. Analytical Chemistry 2014, 86, 3841-3848. (28) Kim, T.-H.; Hwang, H.; Gorkin, R.; Madou, M.; Cho, Y.K. Sensors and Actuators B: Chemical 2013, 178, 648-655. (29) Park, J.; Sunkara, V.; Kim, T.-H.; Hwang, H.; Cho, Y.-K. Analytical Chemistry 2012, 84, 2133-2140. (30) Coumans, F. A. W.; van Dalum, G.; Beck, M.; Terstappen, L. W. M. M. PLoS ONE 2013, 8, e61774. (31) Park, J.-M.; Kim, M. S.; Moon, H.-S.; Yoo, C. E.; Park, D.; Kim, Y. J.; Han, K.-Y.; Lee, J.-Y.; Oh, J. H.; Kim, S. S.; Park, W.-Y.; Lee, W.-Y.; Huh, N. Analytical Chemistry 2014, 86, 3735-3742. (32) Lin, B. K.; McFaul, S. M.; Jin, C.; Black, P. C.; Ma, H. Biomicrofluidics 2013, 7, -. (33) Zheng, S.; Lin, H.; Liu, J.-Q.; Balic, M.; Datar, R.; Cote, R. J.; Tai, Y.-C. Journal of Chromatography A 2007, 1162, 154-161. (34) Lim, L. S.; Hu, M.; Huang, M. C.; Cheong, W. C.; Gan, A. T. L.; Looi, X. L.; Leong, S. M.; Koay, E. S.-C.; Li, M.-H. Lab on a Chip 2012, 12, 4388-4396. (35) Zheng, S.; Lin, H.; Lu, B.; Williams, A.; Datar, R.; Cote, R.; Tai, Y.-C. Biomed Microdevices 2011, 13, 203-213.

9

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 10 of 12

For TOC only

10 ACS Paragon Plus Environment

Page 11 of 12

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

Analytical Chemistry

231x155mm (144 x 144 DPI)

ACS Paragon Plus Environment

Analytical Chemistry

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

Page 12 of 12