In-Channel Printing-Device Opening Assay for Micropatterning

Jan 29, 2015 - A semiautomated microfluidic platform for real-time investigation of nanoparticles' cellular uptake and cancer cells' tracking. Mariana...
0 downloads 0 Views 5MB Size
Download PDF
Letter pubs.acs.org/ac

In-Channel Printing-Device Opening Assay for Micropatterning Multiple Cells and Gene Analysis Hao Zhou, Liang Zhao,* and Xueji Zhang* Research Center for Bioengineering and Sensing Technology, School of Chemistry and Biological Engineering, University of Science and Technology Beijing, Beijing, 10083, P. R. China S Supporting Information *

ABSTRACT: Herein we report an easy but versatile method for patterning different cells on a single substrate by using a microfluidic approach that allows not only spatial and temporal control of multiple microenvironments but also retrieval of specific treated cells to profile their expressed genetic information at around 10-cell resolution. By taking advantages of increased surface area of gold nanoparticles on a poly(dimethylsiloxane) (PDMS) coated substrate, cell adhesive-promotive protein, human fibronectin (hFN) can be significantly accumulated on designed regions where cells can recognize the protein and spread out. Moreover, the whole device can be easily opened by hand without any loss of patterned cells which could be retrieved by mouth-pipet. Consequently, we demonstrate the possibility of analyzing the difference of gene expression patterns between wild type MCF-7 cell and MCF/Adr (drug-resistant cell line) from less than 400 cells in total for a single comprehensive assay, including parallel experiments, controls, and multiple dose treatments. Certain genes, especially the P-glycoprotein coding gene (ABCB1), show high expression level in resistant cells compared with the wild type, suggesting a possible pathway that may contribute to the antidrug mechanism.

B

hockers.24 Some reports smartly combined these two strategies to investigate cell−cell interactions.25 However, finding a method that can fulfill the perquisites of both patterning cells within a miniaturized format and plumbing gene expression in specific preconstrained cells remains a challenge. A few reports showed genetic analysis in patterned cells by utilizing fluorescent proteins to value one or two certain genes’ expression level.26,27 Nevertheless, the fluorescence reporter based methods, which ask for the pretransfected cell-lines, can hardly realize quantification of multiple genes’ expression. Quantitative polymerase chain reaction (qPCR or RT-PCR) is the gold standard for studying gene expression in both diagnostic and basic research.28 By using qPCR, the numbers of interesting genes can be quantitatively measured in a small amount of starting material, even in a single cell.29 Unfortunately, none of the cell patterning approaches demonstrate the compatibility with qPCR in a scalable format which enables both the imaging analysis and subsequent gene transcriptional profiling. To address the above challenges, we report here the development of a novel technology called “In-channel printing-Device opening Assay (Ip-Do Assay)”, which allows

ecause of the fast growing needs of studying cell function and heterogeneity, scientists became more and more dependent on the high-content or high-throughput processing of cellular images1,2 and quantitative analysis of gene expression.3,4 There are several successful technologies capable of imaging cells in an outturn-efficient way, such as automated microscopy,5 flow cytometry,6 and microfabrication based scalable cell cultivation.7 Among these methods, cell patterning is a robust technique for the studies of tissue engineering,8,9 cell-based biosensors,10 and drug screening.11 Cell patterning has some advantages like high scalability and addressability during multiplex cell screening, along with precise temporal and spatial control of the microenvironment in which cells were involved. In the past decade, cell patterning methodologies have been widely applied to understanding cell−cell interactions,12 cellular response to extracellular environments,13 cell polarization,14,15 differentiation,16,17 and other cellular functions. Basically, approaches of cell patterning can be cataloged into two groups. One approach is surface modification by using different chemicals which can be chosen either for repelling cell adhesion on specific regions, usually known as self-assembled monolayers (SAMs) based patterning,18 or for facilitating cell adhesion, such as human fibronectin19 and Arg-Gly-Asp peptide.20 The other approach of patterning is physical constraining which encloses specific regions by microfabricated channels,21 wells,22 gaps,23 or © XXXX American Chemical Society

Received: December 8, 2014 Accepted: January 29, 2015

A

DOI: 10.1021/ac504823s Anal. Chem. XXXX, XXX, XXX−XXX

Letter

Analytical Chemistry

Figure 1. Schematic of “Ip-Do Assay” method for cell patterning and gene analysis. Microfluidic chip reversibly bond with a PDMS-coated glass substrate and was used as an “Ip-Do Assay” device. To synthesize the gold nanofilm on PDMS as a strip-pattern, gold(III) chloride solution was delivered into the chip and incubated to conduct the reduction reaction, followed by human fibronectin incubation and blow-drying. Cells were introduced into channels which have been orthogonally aligned with the nanogold strips after the first layer of the chip was peeled off. Cells attached and formed patterns on hFN modified gold-islands in channels after overnight incubation. On the basis of imaging analysis, the desired pellet of patterned cells can be further retrieved by using a mouth-pipet to perform consequent biochemical reactions in the tube and qPCR.

demonstration that shows the possibility to recover microfluidic micropatterned cells and measure the gene expression by qPCR analysis. Instead of other methods immobilizing nanoparticles, we took a strategy of in situ synthesis of GNF31 on the PDMS surface to construct the nanostructures which can be used as docking spots for human fibronectin (hFN), a well-known extracellular matrix (ECM) protein, enabling cell adhering on the chip. The GNF provides a significantly increased surface area which permits accumulation of extracellular matrix proteins to promote cell adhesion and growth. To dissect the surface engineering process, we observed the substrate morphology changes of before and after synthesis and modification by utilizing atomic force microscope (AFM, nanoscope IIIa, Bruker) topography and scanning electron microscopy (Figures S4 and S5 in the Supporting Information). The AFM and SEM images showed that the initial PDMS membrane was basically clear and barely flat while the GNF consisted of a layer of 100 nm nanoparticles, offering a stable microenvironment for protein anchoring. These nanostructures were relatively stable and could provide a better binding environment for proteins or other extracellular matrix molecules. Notably, after hFN incubation, the nanoparticles became bigger and the surface roughness increased, indicating that the hFN molecules were absorbed on the surface of gold nanoparticles. We then tested the utility of printing adhesive mammalian cells on the chip with the help of the nanofilm interface and hFN accumulation. We fabricated a 6 × 6 GNF microisland array (100 μm × 100 μm for each island) by aligning the second chip layer containing six parallel channels orthogonally with the six preformed GNF strips on the substrate. We delivered the human cervical carcinoma cell, Hela cell suspension (2 × 106 cells/mL) into the device and then equalized the liquid levels between the inlet and outlet of each

not only patterning multiple types of cells inside the microfluidic chip onto a substrate feasibly but also measuring the gene expression level of specific patterned cells precisely. By performing various testing experiments, we proved that “Ip-Do Assay” technology could achieve successful cell patterning with nearly 100% cell viability and owned the ability of handling different types of cells simultaneously, thus allowing multiplex treatments on a cell array (6 × 6) at a spatial resolution of 80 μm, with permitting the picking up of small groups of cells which can been consequently used for conducting gene expression analysis. Moreover, by applying this new method to investigating the molecular biological mechanism of resistivity to the antitumor drug doxorubicin in human breast cancer cell lines, we showed the accessibility of personal medicine on-a-chip compatible with a standard qPCR protocol. The Ip-Do Assay consists of several key steps which include device setting, gold nanofilm (GNF) synthesis, cell printing, chip opening, cell retrieval by mouth-pipet manipulation, and transcriptome amplification30 for multiplex qPCR analysis, as shown in Figure 1. The cells can be easily printed on specific regions where GNF is addressed in channels. Uniquely, compared with those presealed microfluidic approaches, this technique contains two times of assembly and two times of peeling off the chip layer from the substrate, which gain at least two advantages: (1) the extreme flexibility of controlling the size of cell-patterning, which can be tuned by adjusting the size, interval, or shape of GNFs on PDMS and the dimensions of the second chip layer independently or jointly (Figures S1 and S3 in the Supporting Information). (2) After cell printing or even certain treatments, device opening allows the accessibility of picking up patterned cells from the substrate (Figure S2 and Movie SV1 in the Supporting Information). Additionally, quality-improved images can be acquired by eliminating PDMS scattering from the top and walls (Figures S6 and S7 in the Supporting Information). To our knowledge, this is the first B

DOI: 10.1021/ac504823s Anal. Chem. XXXX, XXX, XXX−XXX

Letter

Analytical Chemistry

Figure 2. Imaging and gene analysis of patterning MCF-7 and HUVECs in a single chip. (A) Bright field, cells patterned on predesigned areas in chip channels after the top channels were peeled off from substrate. (B) Fluorescence field, two types of cells were prestained by CellTracker Green dye (MCF-7) and CellTracker red dye (HUVEC), respectively. The width of channels was 100 μm while the width of GNF strips was 80 μm. The scale bar is 100 μm. (C) Amount of copy numbers of several interested genes normalized by cell amount.

channels, some cells at the margin of the channels may fall off from the substrate, and still it is because cells attach to the edge of the channels (Figure S9 in the Supporting Information). Taken together, these results indicate that a reliable protein anchoring provided by gold nanostructure is necessary for cell recognition, patterning and prevents the possible cell damage in the device-opening process in this technique. Next we tested the robustness of our Ip-Do assay. We confirmed this welfare by performing cell printing 3 weeks after the GNF generation and hFN blow-dry (Figure S10 in the Supporting Information). To our surprise, after 3 weeks storage at room temperature, the chip can still be used to pattern the Hela cells without any different operations compared with the previous ones and meanwhile shows cell viability close to 100% as well. We preliminary attribute this robustness to the GNF nanostructures wherein the hFN nested, reducing the chance of being oxidized or harmed from an improper environment. Furthermore, this robust technology permits the feasibility for commercialization in personal diagnosis or a clinical assay by eliminating the cost of storage such as “frozen before use” and inconvenient requirements like fresh device fabrication. This format enables us to perform printing of either homoor heterotypes of cells on a single chip without complex chemical modification. Since the channels are independent of each other, this approach allows patterning multiple cell types simultaneously. In order to verify this, we labeled human umbilical vein endothelial cell (HUVECs) and breast cancer cell MCF-7 with green and red fluorescent dyes (CellTracker, Life Technologies), respectively. By printing these two types of cells in the six parallel channels alternatively, we obtained a 6 ×

channel to facilitate cells settling down. After overnight incubation, on average 10−12 Hela cells would automatically pattern on each microisland in the chip (Movie SV2 in the Supporting Information). The channels were then rinsed by 1× PBS twice to clean up the unattached cells in channels, while the cells adherent on microislands were still retained. During the 24 h after seeding, the cells adhered to and spread only on the gold-islands sprouted by hFN, indicating that compared with intact PDMS, the live cells prefer attaching and growing on the ECM functional surface. In addition, we also validated that the peeling off process has no harm to the attached cells, which is one of the most important preconditions to perform genetic measurements (Figure S6 and S7 in the Supporting Information). Moreover, to assess the specificity of this printing strategy, we substantiated the indispensability of both the GNF and the human fibronectin accumulation on it by performing two sets of control experiments. First, we observed that if there were GNF strips without hFN incubation, cells would distribute randomly on the PDMS membrane. When peeling off the channels, some cells at the margin of the channels may fall off from the substrate, because during the attachment process, some cells may adhere to the edge of the channels (Figure S8 in the Supporting Information). The results suggested that, comparing with intact PDMS membrane, the GNF structure has no evident preference for cells to adhere on. We then proved that by using a PDMS substrate with hFN incubation but without GNF strips, cells would also distribute randomly in channels, indicating inertness of PDMS for hFN binding, which leads to incapability of cell capture. When peeling off the C

DOI: 10.1021/ac504823s Anal. Chem. XXXX, XXX, XXX−XXX

Letter

Analytical Chemistry

Figure 3. Imaging analysis and gene expression profiling using “Ip-Do Assay” to dissect the drug resistivity mechanism. (A) Micrography analysis of doxorubicin induced cell apoptosis. MCF7 and MCF/Adr (resistant) cell lines were patterned in the chip where they were induced by doxorubicin to investigate the apoptosis. Each cell pattern contained about 10 cells. Cells were stained by Annexin V-FITC/PI to detect apoptosis responses. The scale bar is 100 μm. Top, bright field and bottom, the fluorescent field. (B) Heatmap of gene analysis after adriamycin induction. 2Δ method was used to normalize the data. ACTB was selected as the reference gene. The color key means fold-change contrasted with the drug-free control.

6 heteropatterned HUVEC/MCF-7 array on a single chip. No obvious cross-contamination was detected based on both fluorescent and chemical checking out (Figure 2A, B, S3). Besides imaging analysis, we can genetically distinguish multiple prepatterned cells even between proximal channels. To validate this, we profiled the gene expression of HUVEC and MCF-7 cells, respectively, in the above patterning system by qPCR characterization (Figure 2C). To obtain the target cells after opening of the chip, we picked them up from the certain microislands (Figure S2 and Movie SV1 in the Supporting Information) using a mouth-pipet under the observation via a stereo microscope (745T, Nikon). On each patterned area (gold-island, 80 μm (GNF) × 100 μm (channel)), 6−10 previously printed cells were retrieved by a capillary with an open tip (40 μm in diameter). Each sample was quickly put into cell lysis buffer (Supporting Information) right after recovery from the substrate. Accordingly, the whole transcriptome of picked cells was amplified for consequent qPCR analysis by which the expression of all interested genes had been measured. The results indicated that two housekeeping genes, ACTB and GAPDH, showed a stable and comparable expression level in both MCF-7 and HUVEC cells

as expected. However, c-Myc, a gene, plays important roles in breast cancer development32 and was enriched in MCF-7 cells whereas it was not expressed in HUVECs; while another reporter gene, TM (THBD), as an endothelial signature gene,33 showed a remarkable expression only in HUVECs (Figure 2C) whereas not expressed in MCF-7. These results also proved in gene level that no cross-contamination happened in this “Ip-Do assay” system, which provides both the capability of printing multiplexed cells within a scalable format and the possibility of accessing gene expression of prepatterned cells quantitatively. Finally, we applied our Ip-Do assay to study personal medical antitumor drug screening by conducting drug-cell interaction experiments using some in vitro cultured cell lines as a testing model. In order to perform this, we seeded two homologous breast tumor cell types, MCF/Adr (breast tumor cell line with resistance on Adriamycin) and MCF-7 cells (wildtype breast cell line), in a chip alternately. This on-chip testing allows us to simulate a new-tool based clinical diagnosis, on the assumption that a cancer patient who is under a trail of some new drugs and unfortunately a drug resistance aroused from this chemicalcuring. As a drug model, adriamycin (Doxorubicin) was diluted to different concentrations in complete medium (DMEM/10% D

DOI: 10.1021/ac504823s Anal. Chem. XXXX, XXX, XXX−XXX

Letter

Analytical Chemistry FBS): low dose (10 μg/mL), high dose (100 μg/mL), and pure complete medium set as blank control. First, we printed MCF/ Adr and MCF-7 cells on a Ip-Do device, the chip was then rinsed with 1× PBS, followed by the injection of the preprepared drug solution into corresponding channels, respectively. After treatment for 6 h with corresponding drug concentration, the channel layer was peeled off. We then washed the cell array with 1× PBS and stained it with Annexin V-FITC/PI kit (KeyGen Biotech. Nanjing, China) to detect the apoptosis responses. Cells at early apoptotic stage would be stained by Annexin V-FITC, while those at the late apoptotic stage would be dual stained by AV-FITC and PI. Accordingly, the corresponding apoptosis results varied among cell types and drug concentration (Figure 3A). Under the high dose, over 50% of MCF-7 cells were in early apoptotic stage and 10% population was in late apoptotic stage, with obvious detachment from the microislands; meanwhile the viability of MCF/ Adr cells was close to 90% (Data not shown). Under the low drug concentration, only less than 10% MCF-7 cells were in initial apoptotic stage, while few MCF/Adr programmed cell deaths were observed under this concentration. Given that MCF/Adr is an adriamycin-resistant cell line, the high viability is not surprising. Both cell lines showed close to 100% viability under drug-free control, indicating the system and all manipulations had little interference on the results. Last but not least, we profiled gene expression patterns in these printed tumor cells to unveil the mechanism of drugresistance and determined a possible pathway at molecular levels during apoptosis processes (Figure 3B). In both concentrations, expression level of GAPDH and TP53 showed little difference between the two cell lines. The c-Myc gene, which is considered as a promising target for anticancer drugs,34 stably expressed in MCF/Adr cells after the drug curing; however, it showed a significant decrease in the MCF-7 cell line, suggesting adriamycin had a negative effect on tumor cell cytoactivity and proliferation. Interestingly, we found that ABCB1, which encodes the P-glycoprotein (an ATP-dependent membrane protein that can pump many foreign substances out of cells),35 showed a great rise in MCF/Adr cells, especially under high drug concentration, whereas maintained a low level in the MCF-7 cells. This result indicated that the high level of efflux pump protein P-gp plays an important role in decreasing the drug accumulation in resistant cells, showing a good agreement with the previous study.36 More notably, this gene could be aroused by a high dose of antitumor component treatment. Therefore, we found that, compared with wild type cells, the MCF/Adr cell line transcriptionally awoke its multidrug efflux pump protein P-gp, when encountered with a high dose of doxorubicin treatment. This also explained why in MCF/Adr, but not in MCF-7 cells, the c-Myc expression was unaffected under drug treatment. Decreases that varied with cell types and drug concentrations were also detected in HER2, which is believed to play an important role in the progression and development of breast cancer.37 We believe the expression changes are in normal range, given that cytoactivity has reduced in the apoptosis process. BCL2L1 acts as a potent apoptotic inhibitor, and for MCF-7, it showed an acute decrease in the apoptosis process, while under high concentration it reduced more in the process; but still, MCF/Adr was not susceptible when the drug concentration was low, similarly, a regular decrease was detected under high concentration. The result was quite identical to the fluorescence responses; thus, the BCL2L1 can be choose as a candidate marker for the detection of

adriamycin induced apoptosis procedures in the human breast cancer trail. In summary, the development of microscale cell patterning technologies has been plagued by the lack of compatibility for analyzing gene expression with widely adopted methods, such as qPCR. Herein, we address this limitation by presenting an “Ip-Do assay” microfluidic approach to print different types of cells with multiplexed niches that allow high-throughput study of both image processing and corresponding gene expression profiling. We characterized the performance of this novel method and employed it to reproduce a clinical diagnosis for studying drug-cell interactions. We have demonstrated that this robust technology provides a feasible strategy for multiple cellprinting, scalable condition screening, imaging processing, and even dissecting transcriptional expression of the single patterned cell-pellet base on qPCR, which permits multiple gene analysis simultaneously. We therefore envision that this “Ip-Do assay” method will find various applications for future biological and clinical research, wherein high-throughput cellular imaging and multiplex gene quantification are both required.

■

ASSOCIATED CONTENT

S Supporting Information *

Experimental section, additional figures, tables, and videos. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Prof. Hongwu Du in Department of Biotechnology, University of Science and Technology Beijing, for providing Hela, HUVECs, and MCF-7 cell lines in this research. Mr. Yibiao Liu is acknowledged for help in obtaining highresolution AFM images and Mr. Yang Liu for helping on maintaining the micro fabrication room. We also acknowledge funding support from the National Natural Science Foundation of China (Grant 21305007, 21275017), The Fundamental Research Funds for the Central Universities (Grant FRF-TP13-040A), The Chinese Government Scholarship from China Scholarship Council (Grant 201406465024), and the Beijing Higher Education Young Elite Teacher Project (Grant YETP0423).

■

REFERENCES

(1) Pushkarsky, I.; Liu, Y.; Weaver, W.; Su, T.; Mudanyali, O.; Ozcan, A.; Di Carlo, D. Sci. Rep. 2014, 4, 4717. (2) Battich, N.; Stoeger, T.; Pelkmans, L. Nat. Methods 2013, 10, 1127−1133. (3) Dimov, I.; Lu, R.; Lee, E.; Seita, J.; Sahoo, D.; Park, S.; Weissman, I.; Lee, L. Nat. Commun. 2014, 5, 3451. (4) Streets, A. M.; Zhang, X.; Cao, C.; Pang, Y.; Wu, X.; Xiong, L.; Yang, L.; Fu, Y.; Zhao, L.; Tang, F.; Huang, Y. Proc. Natl. Acad. Sci. U.S.A. 2014, 111, 7048−7053. (5) Lubeck, E.; Cai, L. Nat. Methods 2012, 9, 743−748. (6) Edwards, B.; Oprea, T.; Prossnitz, E.; Sklar, L. Curr. Opin. Chem. Biol. 2004, 8, 392−398.

E

DOI: 10.1021/ac504823s Anal. Chem. XXXX, XXX, XXX−XXX

Letter

Analytical Chemistry (7) Zheng, C.; Zhao, L.; Chen, G.; Zhou, Y.; Pang, Y.; Huang, Y. Anal. Chem. 2012, 84, 2088−2093. (8) Gong, P.; Zheng, W.; Huang, Z.; Zhang, W.; Xiao, D.; Jiang, X. Adv. Funct. Mater. 2013, 23, 42−46. (9) Zervantonakis, I. K.; Hughes-Alford, S. K.; Charest, J. L.; Condeelis, J. S.; Gertler, F. B.; Kamm, R. D. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 13515−13520. (10) Figueroa, X. A.; Cooksey, G. A.; Votaw, S. V.; Horowitz, L. F.; Folch, A. Lab Chip 2010, 10, 1120−1127. (11) Hamon, M.; Jambovane, S.; Bradley, L.; Khademhosseini, A.; Hong, J. W. Anal. Chem. 2013, 85, 5249−5254. (12) Hong, S.; Pan, Q.; Lee, L. P. Integr. Biol. 2012, 4, 374−380. (13) Zheng, W.; Xie, Y.; Zhang, W.; Wang, D.; Ma, W.; Wang, Z.; Jiang, X. Integr. Biol. 2012, 4, 1102−1111. (14) Scott, M. A.; Wissner-Gross, Z. D.; Yanik, M. F. Lab Chip 2012, 12, 2265−2276. (15) Jiang, X.; Bruzewicz, D. A.; Wong, A. P.; Piel, M.; Whitesides, G. M. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 975−978. (16) Sip, C. G.; Folch, A. Biomicrofluidics 2014, 8, 036504. (17) Myers, F. B.; Silver, J. S.; Zhuge, Y.; Beygui, R. E.; Zarins, C. K.; Lee, L. P.; Abilez, O. J. Integr. Biol. 2013, 5, 1495−1506. (18) Sun, K.; Song, L.; Xie, Y.; Liu, D.; Wang, D.; Wang, Z.; Ma, W.; Zhu, J.; Jiang, X. Langmuir 2011, 27, 5709−5712. (19) Zhang, S. S.; Hong, S.; Kleber, A. G.; Lee, L. P.; Shaw, R. M. FEBS Lett. 2014, 588, 1439−1445. (20) Bai, H. J.; Gou, H. L.; Xu, J. J.; Chen, H. Y. Langmuir 2010, 26, 2924−2929. (21) Chen, Z.; Dai, Y.; Dong, Z.; Li, M.; Mu, X.; Zhang, R.; Wang, Z.; Zhang, W.; Lang, J.; Leng, J.; Jiang, X. Integr. Biol. 2012, 4, 1090−1095. (22) Wang, Y.; Shah, P.; Phillips, C.; Sims, C. E.; Allbritton, N. L. Anal. Bioanal. Chem. 2012, 402, 1065−1072. (23) Javaherian, S.; O’Donnell, K. A.; McGuigan, A. P. PLoS One 2011, 6, e20909. (24) Zhang, K.; Chou, C. K.; Xia, X.; Hung, M. C.; Qin, L. Proc. Natl. Acad. Sci. U.S.A. 2014, 111, 2948−2953. (25) Chen, Z.; Li, Y.; Liu, W.; Zhang, D.; Zhao, Y.; Yuan, B.; Jiang, X. Angew. Chem. 2009, 48, 8303−8305. (26) Thompson, D.; King, K.; Wieder, K.; Toner, M.; Yarmush, M.; Jayaraman, A. Anal. Chem. 2004, 76, 4098−4103. (27) Yarmush, M.; King, K. Annu. Rev. Biomed. Eng. 2009, 11, 235− 257. (28) VanGuilder, H.; Vrana, K.; Freeman, W. Biotechniques 2008, 619−626. (29) Sanchez-Freire, V.; Ebert, A.; Kalisky, T.; Quake, S.; Wu, J. Nat. Protoc. 2012, 7, 829−838. (30) Tang, F.; Lao, K.; Barbacioru, C.; Wang, Y.; Nordman, E.; Lee, C.; Xu, N.; Wang, X.; Tuch, B.; Bodeau, J.; Siddiqui, A.; Surani, M. Nat. Methods 2009, 6, 377−382. (31) Zhang, Q.; Xu, J.; Liu, Y.; Chen, H. Lab Chip 2008, 8, 352−357. (32) Janghorban, M.; Farrell, A.; Allen-Petersen, B.; Pelz, C.; Daniel, C.; Oddo, J.; Langer, E.; Christensen, D.; Sears, R. Proc. Natl. Acad. Sci. U.S.A. 2014, 111, 9157−9162. (33) Dekker, R.; Boon, R.; Rondaij, M.; Kragt, A.; Volger, O.; Elderkamp, Y.; Meijers, J.; Voorberg, J.; Pannekoek, H.; Horrevoets, A. Blood 2007, 107, 4354−4363. (34) Dobbelstein, M.; Moll, U. Nat. Rev. Drug Discovery 2014, 13, 179−196. (35) Gottesman, M.; Fojo, T.; Bates, S. E. Nat. Rev. Cancer 2002, 2, 48−58. (36) AbuHammad, S.; Zihlif, M. Genomics 2013, 101, 213−220. (37) Cooke, T.; Reeves, J.; Lanigan, A.; Stanton, P. Ann. Oncol. 2001, 12 (Suppl1), S23−S28.

F

DOI: 10.1021/ac504823s Anal. Chem. XXXX, XXX, XXX−XXX