Chip-Assisted Single-Cell Biomarker Profiling of Heterogeneous

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES

Article

Chip-Assisted Single-Cell Biomarker Profiling of Heterogeneous Circulating Tumor Cells Using Multifunctional Nanospheres Ling-Ling Wu, Man Tang, Zhi-Ling Zhang, Chu-Bo Qi, Jiao Hu, Xu-Yan Ma, and Dai-Wen Pang Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02585 • Publication Date (Web): 08 Aug 2018 Downloaded from http://pubs.acs.org on August 8, 2018

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 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 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.

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 29 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

Chip-Assisted Single-Cell Biomarker Profiling of Heterogeneous Circulating Tumor Cells Using Multifunctional Nanospheres Ling-Ling Wu, Man Tang, Zhi-Ling Zhang, Chu-Bo Qi, Jiao Hu, Xu-Yan Ma, Dai-Wen Pang*

Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education), College of Chemistry and Molecular Sciences, State Key Laboratory of Virology, The Institute for Advanced Studies, and Wuhan Institute of Biotechnology, Wuhan University, Wuhan, 430072, P. R. China.

* Corresponding author. Email: [email protected]. Phone: 0086-27-68756759; Fax: 0086-27-68754067

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 29

ABSTRACT: Profiling the heterogeneous phenotypes of single circulating tumor cells (CTCs) from patients is a very challenging task which paves new ways for cancer management, especially personalized anti-cancer therapy. Herein, we propose a chip-assisted multifunctional nanosphere system for efficient and reliable biomarker phenotype analysis of single

heterogeneous

CTCs.

Red

fluorescent-magnetic-biotargeting

multifunctional

nanospheres and green fluorescent-biotargeting nanospheres targeting to two kinds of CTC biomarkers, are used for convenient dual-fluorescence labeling of CTCs along with magnetic tags. By integrating magnetic enrichment with size-selective single-cell trapping microfluidic chip (SCT-chip), over 90% of CTCs, even as low as 10 CTCs per mL of blood, can be individually trapped at highly ordered micropillars, spatially separated from the minimum residual blood cells. Such single CTCs offer easy readout fluorescence signals, facilitating efficient identification and reliable phenotype analysis in accordance with their biomarker expressions. Therefore, the phenotypes of breast tumor cells at the expression level of human epidermal growth factor receptor-2, important targets of clinical anti-cancer drugs, are accurately assessed, and above 82% of them can be classified into corresponding cell subpopulations.

Furthermore,

this

system

demonstrates

successful

detection

and

subpopulation analysis of heterogeneous CTCs from 7 breast cancer patients, which provides a promising new means for single-cell profiling of CTC biomarker phenotypes and guiding personalized anti-cancer therapy. KEYWORDS:

fluorescent-magnetic

nanosphere,

single

circulating

tumor

cell,

heterogeneous phenotype, biomarker profiling

2


Page 3 of 29 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


Circulating tumor cells (CTCs), shed into vasculature from solid tumor, not only assist cancer diagnosis, prognosis, and relapse identification, but conveniently provide biological information to guide therapy.1, 2 Currently, numerous techniques have been developed for CTC isolation and detection, such as affinity separation,3-5 microfiltration,6 and dielectrophoresis.7 Beyond that, assessment of predictive biomarkers on CTCs may guide proper therapeutic intervention, particularly for personalized anti-cancer therapy.8-10 For instance, the expression level of human epidermal growth factor receptor-2 (HER2) is generally utilized to match breast cancer patients with right treatment schedules and predict the effectiveness of targeted drugs.11, 12 However, due to the heterogeneity inherent in CTCs even from the same patients, profiling biomarkers at single-cell level is highly urgent and still technically difficult.1, 13, 14 In recent years, nanomaterials have been popularly employed in characterizing cellular biomarker expressions with their prominent magnetic,15,

16

optical,17,

18

and electric

properties.19, 20 The research trend is gradually towards single-cell analysis from original bulk analysis,13, 14 especially with the rapid advances in microfluidic technique and its merits in micromanipulation.21-24 For example, magnetic nanoparticle-tagged CTCs can be spatially sorted with well-designed fluidic devices according to their biomarker expression levels.25-27 While, magnetism-based single-cell analysis suffers from drawbacks of rigid magnetic field manipulation and low throughout due to the single variable of magnetism. In contrast, fluorescence analysis receives more attention in assessing cellular biomarker expressions for its high sensitivity, high throughput, and convenient coupling with various detection systems.8, 9, 14, 20, 28 Nevertheless, there are still some key issues to be addressed to achieve 3



Page 4 of 29

fluorescence-enabled biomarker analysis of heterogeneous CTCs. Firstly, the extreme rarity of CTCs (several per billions of blood cells) makes their efficient isolation and accurate identification a prerequisite for biomarker characterization.5,

29

But isolation and

identification commonly utilizing the specific recognition of different CTC biomarkers, may hamper biomarker profiling due to the limited types of biomarkers,8 and also lead to low detection efficiency and CTC loss.30 Secondly, the sensitivity of fluorescence analysis suffers from background, as well as photobleaching (fluorescent dyes), weak intensity (up-conversion phosphors), and nonspecific staining (quantum dots (QDs)) of fluorescent probes. What’s more, fluorescence analysis of single CTCs remains challenging. Commercial fluorescence-activated cell sorting (FACS) does not possess sufficient sensitivity for analyzing the rare CTCs in patient samples.26, 27 For fluorescence microscopy, CTC signals are generally obfuscated by residual white blood cells (WBCs),31,

32

and CTC random

distribution in large area and even several planes makes analysis laborious.33, 34 Hence, it is highly desirable to develop efficient and reliable fluorescence analysis strategy for single-cell biomarker profiling of heterogeneous CTCs. Previously,

we

introduced

poly(styrene/acrylamide)

copolymer

nanospheres

(Pst-AAm-COOH) as carriers for loading QDs or/and nano-γ-Fe2O3 to construct fluorescent or/and magnetic nanospheres.35-38 Not only do these functional nanospheres possess negligible nonspecific absorption and prominent stability, they also converge the fluorescence/magnetism of hundreds of nanoparticles, generating high fluorescence intensity and quick magnetic response.3,

38-40

Hence, they have demonstrated superiorities in the

capture and detection of DNA, proteins, pathogens, and tumor cells.39, 41-44 4


Page 5 of 29 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


Herein, red fluorescent-magnetic nanospheres (RMNs) and green fluorescent nanospheres (GNs) were prepared and modified with corresponding monoclonal antibody (mAb)

to

fabricate

two

kinds

of

multifunctional

nanobioprobes:

anti-epithelial-cell-adhesion-molecule (EpCAM) mAb modified RMNs (RMNs-anti-EpCAM, IRMNs) and anti-HER2 mAb modified GNs (GNs-anti-HER2, IGNs). With the combination of IRMNs and IGNs and the integration of single-cell trapping microfluidic chip (SCT-chip), a chip-assisted multifunctional nanosphere system was developed for efficient and reliable fluorescence analysis of biomarker phenotype of single heterogeneous CTCs. The unique advantages of this approach can be listed as follows: 1) Simultaneous fluorescence and magnetic labeling of CTCs with IRMNs by targeting the same biomarkers avoids the competitions between different kinds of single-functional bioprobes. Thus, CTCs can be efficiently isolated from samples by magnetic force along with strong fluorescence for in situ identification, free of CTC loss and tedious procedures from additional identification. 2) The intensity analysis of fluorescence from IGNs binding specifically and efficiently to cellular HER2 capacitates reliable classification of CTCs into three cell subpopulations according to HER2 expression (low, medium, or high). 3) With the assistance of size-selective SCT-chip, dual-labeled CTCs are individually trapped at micropillars without fluorescence interference from WBCs and free IRMNs, allowing accurate and sensitive fluorescence analysis of single CTCs. Moreover, the highly ordered configuration of micropillars permits indexing cell position, which significantly reduces analysis time and improves analysis efficiency. Therefore, we expect that this chip-assisted multifunctional nanosphere system provides new opportunities for CTC phenotype analysis and personalized anti-cancer therapy. 5



Page 6 of 29

Scheme 1. Schematic diagram for single-cell biomarker analysis of heterogeneous CTCs using multifunctional nanospheres integrated with SCT-chip.

EXPERIMENTAL SECTION Reagents and Instruments. Anti-EpCAM mAb, tetraethyl orthosilicate (TEOS), (3-aminopropyl)triethoxysilane poly(ethyleneimine)

(PEI)

(APTES),

were

and

purchased

succinic from

anhydride,

Sigma-Aldrich.

and

branched

Allophycocyanin

(APC)-labeled anti-CD45 mAb was bought from Abcam. CdSe/ZnS QDs was obtained from Wuhan Jiayuan Quantum Dots Co. Ltd. Anti-HER2 mAb, Hoechst 33342 and magnetic scaffold (12320D) were purchased from Thermo Fisher Scientific. All cell lines were purchased from China Center for Type Culture Collection. Red blood cell (RBC) lysis buffer was bought from G-Biosciences. Dynamic light scattering data were measured on a Malvern Zetasizer Nano ZS instrument. Transmission electron microscopy (TEM) images were 6


Page 7 of 29 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


recorded with a FEI Tecnai G2 20 TWIN electron microscope. Fluorescence emission spectra were measured with a Fluorolog-3 (Horiba Jobin Yvon) fluorescence spectrometer. Magnetic hysteresis loops were acquired with a vibrating sample magnetometer (Lake Shore 7410 VSM). Fabrication of IRMNs and IGNs. RMNs were fabricated according to our published layer-by-layer (LBL) assembly method.37 As illustrated in Scheme S1, four layers of nano-γ-Fe2O3 and three layers of QDs emitting at 610 nm were successively assembled onto PEI-coated Pst-AAm-COOH through coordination between primary amines of PEI and metallic atoms of nanoparticles. For biological applications, silica shells were coated onto the surfaces of RMNs by hydrolysis of TEOS and APTES, and carboxyl groups were subsequently introduced by reacting with succinic anhydride. GNs were prepared by embedding QDs emitting at 525 nm into the hydrophobic hollow cavities of Pst-AAm-COOH.35, 36 Then, referring to our published procedure,3 RMNs-anti-EpCAM and GNs-anti-HER2 were obtained using carbodiimide chemistry to cross-link amines of antibodies with carboxyl groups on the surface of nanospheres. The resultant immunonanospheres were dispersed in 0.01M pH 7.2 PBS at 4 °C for later use. Capture and Labeling of Tumor Cells with IRMNs. Tumor cells (5.0 × 104 cells per mL) were incubated with 0.198 mg of IRMNs for 20 min at room temperature (RT) with gentle shaking. After magnetic separation, the numbers of cells captured and uncaptured were counted with a hemocytometer to calculate capture efficiency. As controls, IRMNs and unmodified RMNs were used to treat Jurkat T and SK-BR-3 cells, respectively. Before fluorescence imaging, free IRMNs were partly removed by centrifugation (600 rpm, 3 min). 7



Page 8 of 29

Fluorescence Labeling of HER2 with IGNs. GNs-anti-HER2 were used to label cellular HER2 by incubating with tumor cells for 20 min with gentle shaking at RT. After removing excess IGNs by centrifugation, the fluorescence intensity and labeling efficiency of treated cells were determined with flow cytometry on CyAnTM ADP (Beckman coulter) by determining at least 10 000 cells. Further data analysis was performed using Flowjo software. Scanning Electron Microscopy (SEM) Observation. Samples of IGNs and IGN-labeled tumor cells were prepared for SEM imaging on Zeiss Sigma scanning electron microscope with a standard procedure. Before observation, cells were fixed with 2.5% glutaraldehyde in PBS for 8 h at 4 °C, and then dehydrated with a series of alcohol concentrations (10%, 30%, 50%, 70%, 90%, 100%) followed by drying with vacuum freeze drier. Fabrication of SCT-chip. SCT-chip was fabricated with the standard soft lithography method. With a photomask, the SU8-2015 pattern about 25-µm-thick on a silicon wafer was obtained, and then poured with polydimethylsiloxane (PDMS). After being baked at 75 °C for 4 h, the PDMS mold was peeled off from the photoresist structures. With oxygen plasma treatment, the thin PDMS film of micropillar array was carefully put onto a clean glass slide and bonded with the PDMS of upper fluid layer punched with one inlet and one outlet. The fluid flow path in SCT-chip was simulated theoretically using COMSOL Multiphysics 3.5a software. Capture and In Situ Fluorescence Analysis of Tumor Cells. As shown in Scheme 1, IRMNs and IGNs were simultaneously added into 1 mL of CTC-containing samples and incubated for 20 min at RT with gentle shaking. After magnetic separation and washing, the captured cells were dispersed in 50 µL of PBS and loaded into SCT-chip, followed by 8


Page 9 of 29 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


washing with 50 µL of PBS with a pump. The trapped cells in micropillars were analyzed individually under an inverted fluorescence microscope (TE2000-U, Nikon) mounted with CCD camera (Retiga 2000R, Qimaging Corp.). Fluorescence images were taken in separate channels to minimize potential cross-talk effect. The fluorescenc intensity was analyzed with Imaging-Pro-Plus software. Detection of CTCs from Mimic Samples and Cancer Patients. Blood samples from healthy people and cancer patients were respectively supplied by Hospital of Stomatology of Wuhan University and Hubei Cancer Hospital into EDTA-coated vacutainer tubes for use within 24 h. Mimic clinical samples were prepared by spiking rare Hoechst 33342-stained SK-BR-3 cells into healthy human blood, and then processed with RBC lysis buffer followed by centrifugation to remove the supernatant. The remaining cells were re-suspended in 1 mL of PBS containing 50 µg/mL APC-labeled anti-CD45 mAb, and then treated with chip-assisted multifunctional nanosphere system as described in the part Capture and In Situ Fluorescence Analysis of Tumor Cells. The residual WBCs after magnetic separation, the trapped tumor cells and WBCs in SCT-chip, and the lost tumor cells during the whole procedure were all counted to evaluate the performance of this method. Blood samples from 7 cancer patients and 3 healthy normal controls were treated with above-described procedure except that the remaining cells after RBC lysis were re-suspended in 1 mL of PBS containing both 50 µg/mL APC-labeled anti-CD45 mAb and 1 µg/mL Hoechst 333342. Only the cells labeled by IRMNs and Hoechst 33342 but CD45 negative were enumerated as CTCs, and their HER2 expressions were determinded based on the fluorescence intensity of the bound IGNs. 9



Page 10 of 29

RESULTS AND DISCUSSION Characterization of the IRMNs and IGNs. The TEM image showed that RMNs were monodispersed without aggregation (Figure 1A). The mean diameter of the RMNs was 349.2 ± 20.4 nm (for 200 RMNs) with the size coefficient of variation (CV) of 5.8% (Figure 1B), indicating their high uniformity in size distribution. Magnetic hysteresis loop (Figure 1C) demonstrated the excellent superparamagnetic property of the RMNs at RT with nearly zero coercivity and a large magnetic saturation value (19.1 emu/g). Above 93% of the RMNs could be attracted within 2 min with an ordinary magnetic scaffold (Figure 1D), suggesting the RMNs to be good magnetic separation vehicles. Besides magnetic property, RMNs also inherited the prominent optical properties of QDs. Fluorescence spectra of RMNs and QDs (Figure 1E) indicated that photoluminescence properties of QDs essentially unchanged with stepwise assemblies, except for a slight red-shift owing to the change of QD environments from n-hexane to nanosphere surface.44,

45

Meanwhile, fluorescence intensity of RMNs

gradually increased with increasing assembly layer of QDs, indicating the high controllability and repeatability of the assembly processes. Theoretically, a series of fluorescent-magnetic nanospheres with different emission peaks and fluorescence intensities could be obtained by changing the colors of QDs and assembly layers. RMNs with three layers of QDs were chosen in our experiment, because their fluorescence was strong enough to be observed under fluorescence microscope (Figure 1F), and maintained extraordinary stable during 7 months’ storage or under continuous illumination for 10 min with light from a 100 W mercury lamp under a 100 × oil-immersion objective (Figure 1G). Additionally, when four batches of RMNs were fabricated in parallel, their Zeta potentials during the process of silanization and 10


Page 11 of 29 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


carboxyl modification and their final fluorescence intensities were comparable (Figure 1H), proving the good reproducibility of our methods. GNs were well dispersed with smooth surface and uniform size (Figure S1A-B), and possessed strong, homogeneous, and photostable fluorescence (Figure S1C-E). Moreover, the fluorescence intensity of GNs was proportional to their concentrations over a range from 0.2 to 30 µg/mL (Figure S1F), which further confirmed the homogeneity of the GNs and made quantitative analysis possible. RMNs and GNs were then modified with anti-EpCAM and anti-HER2 mAb, respectively, which were confirmed by the fact that immunonanospheres could specifically react with dye-labeled anti-mouse IgG (Fab specific) Ab (Fig. S2 and S3A-C). Besides, there was no nonspecific interaction between RMNs-anti-EpCAM and GNs-anti-HER2 (Figure S3D), laying a solid foundation for subsequent fluorescence analysis of dual-labeled CTCs.

Figure 1. Characterization of the RMNs. TEM image (A) and size distribution histogram (B) of the RMNs. (C) Magnetic hysteresis loop of the RMNs measured at RT. (D) Capture efficiencies of RMNs at different attraction times. (E) Fluorescence (FL) spectra of RMNs coated with one layer (purple), double layers (blue), and triple layers (green) of QDs, compared with those of magnetic nanospheres (black) and QDs dispersed in n-hexane (red). 11



Page 12 of 29

(F) Fluorescence microscopic image of the RMNs. (G) FL intensity of the RMNs at different storage times or different illumination times. (H) Zeta potentials of four batches of RMNs at different stages of silanization and carboxyl modification and their FL intensities. Visual Recognition and Efficient Isolation of Tumor Cells with IRMNs. The feasibility of IRMNs for simultaneous magnetic capture and fluorescence labeling of tumor cells was investigated. As shown in Figure 2A, more than 94% of SK-BR-3 cells were captured with IRMNs. At the same condition, IRMNs hardly capture any Jurkat T cells (a kind of human peripheral blood leukemia T cells as control cells), and unmodified RMNs failed to capture SK-BR-3 cells either, suggesting the binding of IRMNs to SK-BR-3 cells was specific and effective. Also, IRMNs could isolate two other kinds of breast tumor cells with efficiencies of above 98%, indicating their general applicability. Apart from IRMN-enabled magnetic capture of tumor cells, IRMNs offered great potential for direct fluorescence identification. As shown in Figure 2C, the concurrent Hoechst 33342 and IRMN fluorescence signals could be observed in all captured tumor cells, and the inset image showed plenty of IRMNs on cell surface, producing strong and homogeneous red fluorescence. In contrast, no fluorescence of IRMNs was found on the surfaces of residual Jurkat T cells. Hence, although some control cells might be magnetically captured due to the nonspecific binding of a few IRMNs and the large magnetic saturation value of IRMNs, the negligible fluorescence on cell surfaces could be used to conveniently distinguish them from tumor cells. Collectively, IRMNs were capable of efficient capture and in situ identification of tumor cells.

12


Page 13 of 29 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


Figure 2. Simultaneous capture and fluorescence labeling of tumor cells with IRMNs. (A) Capture efficiencies of IRMNs to SK-BR-3 and Jurkat T cells, and RMNs to SK-BR-3 cells. (B) Efficiencies to capture SK-BR-3, MDA-MB-453, and MCF-7 cells. (C) Fluorescence microscope images of Hoechst 33342 stained-cells (blue) captured by IRMNs (red) (inset: merged images of individual cells). Reliable Discrimination of Expression Level of Cellular HER2 with IGN-Labeling. Cellular HER2 expressions are heterogeneous, and assessing HER2 phenotype of CTCs might guide treatment choice,11, 46 which is generally ignored with existing CTC detection approaches. For instance, although EpCAM-positive breast cancer cell lines (Figure S4A-B) all could be detected with IRMNs (Figure 2), the HER2 expressions in SK-BR-3, MDA-MB-453, and MCF-7 cells decreased sequentially (Figure S4C-D). GNs-anti-HER2, as fluorescence tags, exhibited specific labeling to tumor cells with over 98% efficiency (Figure S5). The SEM images confirmed the successful attachment of IGNs to cell surfaces (Figure 3A-C). Moreover, the numbers of binding IGNs greatly relied on cellular HER2

13



Page 14 of 29

expression levels. HER2-overexpressing SK-BR-3 cell was bound with a high abundance of IGNs (Figure 3A); while MDA-MB-453 cell with a ca. 6-fold lower HER2 level was attached with less IGNs (Figure 3B) compared with that of SK-BR-3 cell, whose number was obviously more than that of the tagged IGNs on MCF-7 cell with HER2 level reduced by another 6-fold (Figure 3C). As a result, these cell lines with different numbers of fluorescent tags exhibited distinctly decreased fluorescence intensities in the order of SK-BR-3, MDA-MB-453, and MCF-7 cells using fluorescence microscopy and flow cytometry (Figure 3D-E), which were quite consistent with their HER2 expression levels.47 These results confirmed reliable fluorescence assessment of the expression level of HER2 on tumor cells with IGN-labeling. By combining IRMN-enabled magnetic separation, purification, and in situ identification of tumor cells with IGN-enabled HER2 expression profiling, heterogeneous tumor cells could be detected and qualitatively analyzed (Figure S6-7). While, the approach to remove free IRMNs with low-speed centrifugation could not eliminate IRMN fluorescence interferences completely and also caused cell loss (Figure S8), which is unfit for CTC detection and analysis.

14


Page 15 of 29 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


Figure 3. IGN-based expression analysis of cellular HER2. (A-C) SEM images of SK-BR-3 (IGNs (inset a) and IGNs tethered on cell surface (inset b)) (A), MDA-MB-453 (B), and MCF-7 cells (C) labeled with IGNs (white arrows: IGNs). (D) Fluorescence microscopic images of Hoechst 33342 stained-tumor cells (blue) labeled with IGNs (green). (E) Flow cytometric analysis of IGN-labeled tumor cells. Design of SCT-chip and Optimization of the Fluidic Condition. To achieve reliable, sensitive, and efficient fluorescence analysis of CTCs at single-cell level, the size-selective SCT-chip was designed. As shown in Figure 4A, SCT-chip incorporated two layers of PDMS structure: fluid channel layer for loading samples and micropillar array layer for trapping tumor cells. The cell-trapping structure was a horseshoe-shaped micropillar composed of a semi-cylinder (radium: ca. 40 µm; height: ca. 25 µm) with a slit in the middle, and was arranged in gap distance of 40 µm and 20 µm in the direction and the vertical direction of fluid flow, respectively. The slit with an entrance about 20 µm and a gradually decreasing width to 8 µm was specifically designed to trap tumor cells from residual WBCs and free

15



IRMNs based on the relatively larger size of tumor cells (>8 µm).

Page 16 of 29

6, 48

Both microscopic

images and SEM images (Figure 4B-C) showed that micropillars were uniform in shape, and orderly staggered to maximize collisions with cells. The SCT-chip consisted of 4 parallel channels respectively with 8 array patterns (12 × 14 micropillars were considered as one unit of pattern), so there were 5376 micropillars in the area of ca. 48 mm2 for each SCT-chip. The capability of SCT-chip to trap tumor cells was first investigated, and MDA-MB-453 cells (ca. 5 × 103 cells mL-1) were chosen to optimize flow rate for their comparatively smaller size (Figure S9A). As shown in Figure 4D, trapping efficiencies of SCT-chip to target cells gradually increased with the decreasing flow rate. In view of sample throughput, the optimal flow rate was chosen as 3 µL min-1, at which 87% of MDA-MB-453 cells were trapped. Under the optimized flow rate, the other two kinds of tumor cells were all trapped with about 90% efficiencies, and only 4.4% of WBCs were trapped. Additionally, the trajectories of fluorescent microspheres (ca. 7 µm) revealed their free passages through slits (Figure S9B), suggesting that the interferents with size smaller than 8 µm, such as WBCs and free IRMNs, could be discarded with these microstructures. On the other hand, the theoretical analysis of fluid flow path in SCT-chip demonstrated that the flow rates in the cell-trapped slits were nearly zero (Figure 4F), namely that fluid bypassed them, indicating that single cells could be trapped. The microscopic images also showed that micropillars trapped single cells or not (Figure 4G). From all these results, it could be concluded that the SCT-chip was able to efficiently trap single tumor cells from the interferents, facilitating subsequent detection and analysis.

16


Page 17 of 29 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


Figure 4. Design and cell-trapping of SCT-chip. (A) The schematic diagram for designing and cell-trapping of SCT-chip. The microscopic image (B) and the SEM image (C) of micropillar array (inset: an individual micropillar). (D) Trapping efficiencies to MDA-MB-453 cells at different flow rates. (E) Trapping efficiencies to different cells under the optimized conditions. (F) Numerical flow simulation data showing the distribution of velocities within SCT-chip as a contour plot. (G) Microscopic image of Hoechst 33342-stained MDA-MB-453 cells trapped in SCT-chip. Chip-Assisted Single-Cell Analysis of Cellular HER2 Expression with Multifunctional Nanospheres. By integrating magnetic separation with SCT-chip, dual-labeled cells with IRMNs and IGNs were individually trapped at the entrance or inside the slits of ordered microstructures depending on their sizes (Figure 5A). The fluorescence images of IRMNs channel showed that fluorescence background from free IRMNs had been completely 17



Page 18 of 29

eliminatedand red fluorescence could be observed on cell surfaces, enabling some SK-BR-3 cells with relatively weak fluorescence to be easily identified. Therefore, the size-selectivity and ordered configuration of SCT-chip eliminated fluorescence interferences, and enhanced sensitivity and efficiency of fluorescence analysis. On the other hand, the green fluorescence from IGNs binding to cell surfaces clearly reflected the cellular HER2 expression levels (Figure 5A), which were further determined by taking the average fluorescence intensity within whole region of a single cell and subtracting the average background intensity in non-cell region from this value to yield a numerical value, named as IGN-enabled HER2 score. To quantitatively analyze heterogeneous CTCs, we proposed a classification standard on HER2 expression level using three breast tumor cell lines with well-known HER2 expression levels. IGN-enabled HER2 scores for each cell type were measured from at least 500 cells, and with their distributions, appropriate thresholds were determined to categorize individual cellular HER2 expression into three groups (Low: HER2 score < 4.08; Medium: 4.08 ≤ HER2 score ≤ 10.57; High: HER2 score > 10.57) (Figure 5B). Under this standard, above 82% of MCF-7, MDA-MB-453 and SK-BR-3 cells were classified to cell subpopulations of low, medium and high HER2 expression, respectively (Figure 5C). Meanwhile, MCF-7 cells were hardly distributed in high HER2 expression subpopulation, and neither were SK-BR-3 cells in low HER2 expression subpopulation. Moreover, the mixed cells from different cell lines also showed the distinctly different and highly reproducible pattern of cell subpopulation distributions (Figure S10). Therefore, the great consistency of subpopulation distributions of these three cell lines with their HER2 expression levels confirmed the validity of category standard of HER2 expression, 18


Page 19 of 29 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


and verified the feasibility and reliability of chip-assisted single-cell profiling of HER2 expression with multifunctional nanospheres.

Figure 5. Single-cell analysis of HER2 expression with chip-assisted multifunctional nanosphere system. (A) Microscopic images of trapped tumor cells dual-labeled with IRMNs and IGNs on chip. Merge 1: merged images of IRMNs and IGNs; Merge 2: merged images of IRMNs, IGNs and bright field. (B) HER2 scores generated from IGNs labeled three breast tumor cell lines to categorize HER2 expression into three different levels: low, medium, and high. (C) Percentage distributions of MCF-7 (red), MDA-MB-453 (orange) and SK-BR-3 (green) cells in cell subpopulations of different HER2 expression. Capture and Analysis of Heterogeneous CTCs from Mimic Clinical Samples and Cancer Patients. To investigate the performance of chip-assisted multifunctional nanosphere system to detect rare tumor cells in mimic clinical samples, about 20-1000 Hoechst 33342-stained SK-BR-3 cells were spiked into 1 mL of blood. Regression analysis of

19



Page 20 of 29

captured cell number versus spiked cell number obtained y=0.901x (R2=0.998) (Figure 6A), indicating the efficient capture of the system to rare tumor cells. Besides, 95.8 ± 3.4% of residual WBCs during magnetic separation were further removed with SCT-chip which only trapped 89-942 WBCs (Figure 6B). Trapped cells were identified using a common immunocytochemistry method including APC-labeled anti-CD45 mAb (a marker for leukocytes), RMNs-anti-EpCAM, GNs-anti-HER2, and Hoechst 33342 nuclear staining. As shown in Figure 6C, tumor cells were identified as Hoechst 33342+/EpCAM+/CD45− together with the signals of IGNs for HER2 analysis; WBCs were spatially separated from tumor cells and distinguished by Hoechst 33342+/EpCAM-/CD45+, which had negligible effects on CTC identification and analysis. We then applied this system to detect clinical samples from 7 breast cancer patients and 3 healthy people, with 3-6 CTCs in per mL of patient bloods and no CTC in healthy individuals (Table S1). HER2 expressions of single CTCs were evaluated with IGN-enabled HER2 scores, and were compared with corresponding HER2 statuses of their primary tumor tissues (details in S11, Supporting Information.). 78% overall concordance was observed between CTC HER2 statuses and primary tumors, which was consistent with other reports.9, 13, 46 Meanwhile, HER2 expression of isolated CTCs was quite heterogeneous, with 4 of 7 patients containing different cell subpopulation. For Patient 1# with HER2-negative tumor for whom oncologists may not consider HER2-targeted therapy, the existence of CTCs with high and medium HER2 expressions might affect therapeutic decisions, allowing the opportunity of adopting HER2-targeted therapies.10,

46, 49

Therefore, the chip-assisted multifunctional nanosphere

system was capable of detection and analysis of heterogeneous CTCs, showing potential for 20


Page 21 of 29 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


guiding and monitoring personalized therapy.

Figure 6. (A) Recovery of spiked SK-BR-3 cells with chip-assisted multifunctional nanosphere system at different cell concentrations (20-1000 cells mL-1) in mimic clinical samples. (B) Scatter diagram for the numbers of residual WBCs before and after passing through chip (n = 15). (C) Microscopic images of cells captured from mimic clinical samples and identified with immunocytochemistry. Merge: merged images of nucleus, EpCAM, HER2, CD45, and bright field. (D) Enumeration and HER2 expression analysis of CTCs isolated from breast cancer patients along with HER2 statuses in primary tumors. CONCLUSIONS In summary, we have demonstrated chip-assisted multifunctional nanosphere system for single-cell

biomarker

profiling

of

heterogeneous

CTCs.

IRMNs

possessing

fluorescence-traceability, magneto-manipulability, and bio-recognizability, coupled with IGNs, enable convenient dual-fluorescence labeling of CTCs along with magnetic tags

21



Page 22 of 29

through the specific targeting to EpCAM and HER2. With a combination of magnetic isolation and size-selective SCT-chip, CTCs can be individually and efficiently trapped in micropillars, free of fluorescence interference. Thus, the sensitive and reliable fluorescence signals facilitate CTC identification and biomarker profiling at single-cell level, capacitating classification of heterogeneous CTCs into different subpopulations. Besides, the highly ordered micropillar array permits indexing cell position, which significantly reduces analysis time and improves analysis efficiency. The strategy proposed herein can be expended to assess other CTC biomarkers and even to profile multiple biomarkers with multicolor fluorescence probes, providing new opportunities for cancer diagnosis and personalized anti-cancer therapy.

ASSOCIATED CONTENT

The Supporting Information is available free of charge via the Internet at http://pubs.acs.org. Schematic diagram for the fabrication of IRMNs; characterization of GNs, IRMNs and IGNs; analysis of cellular biomarker expression; labeling performance of IGNs; dosage optimization of immunonanospheres for dual-labeling of tumor cells; free IRMN interference; cell diameters and microsphere trajectories in chip; cellular HER2 analysis of mixed cells, HER2 assay on tumor tissues; detection results from clinical samples (PDF).

AUTHOR INFORMATION Corresponding Author * Email: [email protected]

22


Page 23 of 29 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


Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT This work was supported by the National Natural Science Foundation of China (Nos. 21535005, 21475099, and 21275111), the National Science and Technology Major Project of China (2018ZX10301405), the 111 Project (No. 111-2-10), the China Scholarship Council, and Collaborative Innovation Centre for Chemistry and Molecular Medicine.

REFERENCES (1) Millner, L. M.; Linder, M. W.; Valdes, R. Ann. Clin. Lab. Sci. 2013, 43, 295-304. (2) Esmaeilsabzali, H.; Beischlag, T. V.; Cox, M. E.; Parameswaran, A. M.; Park, E. J. Biotechnol. Adv. 2013, 31, 1063-1084. (3) Wu, L. L.; Wen, C. Y.; Hu, J.; Tang, M.; Qi, C. B.; Li, N.; Liu, C.; Chen, L.; Pang, D. W.; Zhang, Z. L. Biosens. Bioelectron. 2017, 94, 219-226. (4) Ahmed, M. G.; Abate, M. F.; Song, Y. L.; Zhu, Z.; Yan, F.; Xu, Y.; Wang, X. M.; Li, Q. B.; Yang, C. Y. Angew. Chem., Int. Ed. 2017, 56, 10681-10685. (5) Xiong, K.; Wei, W.; Jin, Y.; Wang, S.; Zhao, D.; Wang, S.; Gao, X.; Qiang, C.; Yue, H.; Ma, G.; Xie, H. Y. Adv. Mater. 2016, 28, 7929-7935. (6) Hosokawa, M.; Hayata, T.; Fukuda, Y.; Arakaki, A.; Yoshino, T.; Tanaka, T.; Matsunaga, T. Anal. Chem. 2010, 82, 6629-6635. (7) Gupta, V.; Jafferji, I.; Garza, M.; Melnikova, V. O.; Hasegawa, D. K.; Pethig, R.; Davis, D. W. Biomicrofluidics 2012, 6, 24133. 23



Page 24 of 29

(8) Lee, H. J.; Cho, H. Y.; Oh, J. H.; Namkoong, K.; Lee, J. G.; Park, J. M.; Lee, S. S.; Huh, N.; Choi, J. W. Biosens Bioelectron 2013, 47, 508-514. (9) Lee, H. J.; Oh, J. H.; Oh, J. M.; Park, J. M.; Lee, J. G.; Kim, M. S.; Kim, Y. J.; Kang, H. J.; Jeong, J.; Kim, S. I. Lee, S. S.; Choi, J. W.; Huh, N. Angew. Chem., Int. Ed. 2013, 52, 8337-8340. (10)Riethdorf, S.; Müller, V.; Zhang, L.; Rau, T.; Loibl, S.; Komor, M.; Roller, M.; Huober, J.; Fehm, T.; Schrader, I.; Hilfrich, J.; Holms, F.; Tesch, H.; Eidtmann, H.; Untch, M.; von Minckwitz, G.; Pantel, K. Clin. Cancer Res. 2010, 16, 2634-2645. (11) Wolff, A. C.; Hammond, M. E. H.; Hicks, D. G.; Dowsett, M.; McShane, L. M.; Allison, K. H.; Allred, D. C.; Bartlett, J. M. S.; Bilous, M.; Fitzgibbons, P.; Hanna, W.; Jenkins, R. B.; Mangu, P. B.; Paik, S.; Perez, E. A.; Press, M. F.; Spears, P. A.; Vance, G. H.; Viale, G.; Hayes, D. F. J. Clin. Oncol. 2013, 31, 3997-4013. (12) Jordan, N. V.; Bardia, A.; Wittner, B. S.; Benes, C.; Ligorio, M.; Zheng, Y.; Yu, M.; Sundaresan, T. K.; Licausi, J. A.; Desai, R.; O'Keefe, R. M.; Ebright, R. Y.; Boukhali, M.; Sil, S.; Onozato, M. L.; Iafrate, A. J.; Kapur, R.; Sgroi, D.; Ting, D. T.; Toner, M.; Ramaswamy, S.; Haas, W.; Maheswaran, S.; Haber, D. A. Nature 2016, 537, 102-106. (13) Alix-Panabières, C.; Pantel, K. FEBS Lett 2017, 591, 2241-2250. (14) Ohno, M.; Karagiannis, P.; Taniguchi, Y. Molecules 2014, 19, 13932-13947. (15) Kwak, B.; Lee, J.; Lee, D.; Lee, K.; Kwon, O.; Kang, S.; Kim, Y. Biosen. Bioelectron. 2017, 88, 153-158. (16) Bamrungsap, S.; Chen, T.; Shukoor, M. I.; Chen, Z.; Sefah, K.; Chen, Y.; Tan, W. ACS Nano 2012, 6, 3974-3981. 24


Page 25 of 29 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


(17) Lu, W.; Arumugam, S. R.; Senapati, D.; Singh, A. K.; Arbneshi, T.; Khan, S. A.; Yu, H.; Ray, P. C. ACS Nano 2010, 4, 1739-1749. (18) Wu, C. H.; Huang, Y. Y.; Chen, P.; Hoshino, K.; Liu, H.; Frenkel, E. P.; Zhang, J. X.; Sokolov, K. V. ACS Nano 2013, 7, 8816-8823. (19) Wu, Y. F.; Xue, P.; Kang, Y. J.; Hui, K. M. Anal. Chem. 2013, 85, 3166-3173. (20) Han, S. I.; Han, K. H. Anal. Chem. 2015, 87, 10585-10592. (21) Konry, T.; Smolina, I.; Yarmush, J. M.; Irimia, D.; Yarmush, M. L. Small 2011, 7, 395-400. (22) Kang, D. K.; Ali, M. M.; Zhang, K.; Pone, E. J.; Zhao, W. TrAC, Trends Anal. Chem. 2014, 58, 145-153. (23) Joensson, H. N.; Samuels, M. L.; Brouzes, E. R.; Medkova, M.; Uhlén, M.; Link, D. R.; Andersson-Svahn, H. Angew. Chem., Int. Ed. 2009, 48, 2518-2521. (24) Issadore, D.; Chung, J.; Shao, H.; Liong, M.; Ghazani, A. A.; Castro, C. M.; Weissleder, R.; Lee, H. Sci. Transl. Med. 2012, 4, 141ra92. (25) Mohamadi, R. M.; Besant, J. D.; Mepham, A.; Green, B.; Mahmoudian, L.; Gibbs, T.; Ivanov, I.; Malvea, A.; Stojcic, J.; Allan, A. L.; Lowes, L. E.; Sargent, E. H.; Nam, R. K.; Kelley, S. O. Angew. Chem., Int. Ed. 2015, 54, 139-143. (26) Labib, M.; Green, B.; Mohamadi, R. M.; Mepham, A.; Ahmed, S. U.; Mahmoudian, L.; Chang, I. H.; Sargent, E. H.; Kelley, S. O. J. Am. Chem. Soc. 2016, 138, 2476-2479. (27) Poudineh, M.; Aldridge, P. M.; Ahmed, S.; Green, B. J.; Kermanshah L.; Nguyen, V.; Tu, C.; Mohamadi, R. M.; Nam, R. K.; Hansen, A.; Sridhar, S. S.; Finelli, A.; Fleshner, N. E.; Joshua, A. M.; Sargent, E. H.; Kelley, S. O. Nat. nanotechnol. 2017, 12, 274-281. 25



Page 26 of 29

(28)Pramanik, A.; Vangara, A.; Nellore, B. P. V.; Sinha, S. S.; Chavva, S. R.; Jones, S.; Ray, P. C. ACS Appl. Mater. Interfaces 2016, 8, 15076-15085. (29) Arya, S. K.; Lim, B.; Rahman, A. R. A., Lab Chip 2013, 13, 1995-2027. (30) Kim, J. H.; Chung, H. H.; Jeong, M. S.; Song, M. R.; Kang, K. W.; Kim, J. S. Int. J. Nanomed. 2013, 8, 2247-2257. (31) Zhao, L.; Lu, Y. T.; Li, F.; Wu, K.; Hou, S.; Yu, J.; Shen, Q.; Wu, D.; Song, M.; OuYang, W. H.; Luo, Z.; Lee, T.; Fang, X.; Shao, C.; Xu, X.; Garcia, M. A.; Chung, L. W.; Rettig, M.; Tseng, H. R.; Posadas, E. M. Adv. Mater. 2013, 25, 2897-2902. (32) Lv, S. W.; Liu, Y.; Xie, M.; Wang, J.; Yan, X. W.; Li, Z.; Dong, W. G.; Huang, W. H. ACS Nano 2016, 10, 6201-6210. (33) Kamande, J. W.; Hupert, M. L.; Witek, M. A.; Wang, H.; Torphy, R. J.; Dharmasiri, U.; Njoroge, S. K.; Jackson, J. M.; Aufforth, R. D.; Snavely, A.; Yeh, J. J.; Soper, S. A. Anal. Chem. 2013. 85, 9092-9100. (34) Schiro, P. G.; Zhao, M.; Kuo, J. S.; Koehler, K. M.; Sabath, D. E.; Chiu, D. T. Angew. Chem., Int. Ed. 2012, 51, 4618-4622. (35) Xie, H. Y.; Zuo, C.; Liu, Y.; Zhang, Z. L.; Pang, D. W.; Li, X. L.; Gong, J. P.; Dickinson, C.; Zhou, W. Small 2005, 1, 506-509. (36)Xie, H. Y.; Xie, M.; Zhang, Z. L.; Long, Y. M.; Liu, X.; Tang, M. L.; Pang, D. W.; Tan, Z.; Dickinson, C.; Zhou, W. Bioconjug. Chem. 2007, 18, 1749-1755. (37) Xie, M.; Hu, J.; Wen, C. Y.; Zhang, Z. L.; Xie, H. Y.; Pang, D. W. Nanotechnology 2012, 23, 035602. (38) Wen, C. Y.; Xie, H. Y.; Zhang, Z. L.; Wu, L. L.; Hu, J.; Tang, M.; Wu, M.; Pang, D. W. 26


Page 27 of 29 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


Nanoscale 2016, 8, 12406-12429. (39) Hu, J.; Zhang, Z. L.; Wen, C. Y.; Tang, M.; Wu, L. L.; Liu, C.; Zhu, L.; Pang, D. W. Anal. Chem. 2016, 88, 6577-6584. (40) Chen, L.; Wu, L. L.; Zhang, Z. L.; Hu, J.; Tang, M.; Qi, C. B.; Li, N.; Pang, D. W. Biosens. Bioelectron. 2016, 85, 633-640. (41) Hu, J.; Wen, C. Y.; Zhang, Z. L.; Xie, M.; Hu, J.; Wu, M.; Pang, D. W. Anal. Chem. 2013, 85, 11929-11935. (42) Wen, C. Y.; Wu, L. L.; Zhang, Z. L.; Liu, Y. L.; Wei, S. Z.; Hu, J.; Tang, M.; Sun, E. Z.; Gong, Y. P.; Yu, J.; Pang, D. W. ACS Nano 2014, 8, 941-949. (43) Song, E. Q.; Hu, J.; Wen, C. Y.; Tian, Z. Q.; Yu, X.; Zhang, Z. L.; Shi, Y. B.; Pang, D. W. ACS Nano 2011, 5, 761-770. (44) Wen, C. Y.; Hu, J.; Zhang, Z. L.; Tian, Z. Q.; Ou, G. P.; Liao, Y. L.; Li, Y.; Xie, M.; Sun, Z. Y.; Pang, D. W. Anal. Chem. 2013, 85, 1223-1230. (45) Song, E. Q.; Wang, G. P.; Xie, H. Y.; Zhang, Z. L.; Hu, J.; Peng, J.; Wu, D. C.; Shi, Y. B.; Pang, D. W. Clin. Chem. 2007, 53, 2177-2185. (46) Munzone, E.; Nolé, F.; Goldhirsch, A.; Botteri, E.; Esposito, A.; Zorzino, L.; Curigliano, G.; Minchella, I.; Adamoli, L.; Cassatella, M. C.; Casadio, C.; Sandri, M. T. Clin. Breast Cancer 2010, 10, 392-397. (47) Prang, N.; Preithner, S.; Brischwein, K.; Göster, P.; Wöppel, A.; Müller, J.; Steiger, C.; Peters, M.; Baeuerle, P. A.; da Silva, A. J. Br. J. Cancer 2005, 92, 342-349. (48) Sun, J. S.; Liu, C.; Li, M. M.; Wang, J. D.; Xianyu, Y. L.; Hu, G. Q.; Jiang, X. Y. Biomicrofluidics 2013, 7. 11802. 27



Page 28 of 29

(49) Meng, S. D.; Tripathy, D.; Shete, S.; Ashfaq, R.; Haley, B.; Perkins, S.; Beitsch, P.; Khan, A.; Euhus, D.; Osborne, C.; Frenkel, E.; Hoover, S.; Leitch, M.; Clifford, E.; Vitetta, E.; Morrison, L.; Herlyn, D.; Terstappen, L. W. M. M.; Fleming, T.; Fehm, T.; Tucker, T.; Lane, N.; Wang, J. Q.; Uhr, J. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 9393-9398

28


Page 29 of 29 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


For TOC only:

29


Chip-Assisted Single-Cell Biomarker Profiling of Heterogeneous

Recommend Documents