Clinical Chemistry 52: 1080-1088, 2006.
First published March 23, 2006; 10.1373/clinchem.2005.059014
(Clinical Chemistry. 2006;52:1080-1088.)
© 2006 American Association for Clinical Chemistry, Inc.
Microfluidic Techniques for Single-Cell Protein Expression Analysis
Ethan Fitzpatrick1,
Sterling McBride2,
Jonathan Yavelow1,
Saltanat Najmi3,
Peter Zanzucchi2 and
Robert Wieder3,a
1 Department of Biology, Rider University, Lawrenceville, NJ.
2 Sarnoff Corporation, Princeton, NJ.
3 Division of Medical Oncology/Hematology, University of Medicine and Dentistry of New Jersey-New Jersey Medical School, Newark, NJ.
aAddress correspondence to this author at: University of Medicine and Dentistry of New Jersey-New Jersey Medical School, Division of Medical Oncology/Hematology, 185 South Orange Ave., MSB I-596, Newark, NJ 07103. Fax 973-972-2384; e-mail wiederro{at}umdnj.edu.
 |
Abstract
|
|---|
Background: The analysis of single cells obtained from needle aspirates of tumors is constrained by the need for processing. To this end, we investigated two microfluidic approaches to measure the expression of surface proteins in single cancer cells or in small populations (<50 cells).
Methods: One approach involved indirect fluorescence labeling of cell-surface proteins and channeling of cells in a microfluidic device past a fluorescence detector for signal quantification and analysis. A second approach channeled cells in a microfluidic device over detection zones coated with ligands to surface proteins and measured rates of passage and of retardation based on transient interactions between surface proteins and ligands.
Results: The fluorescence device detected expression of integrin 
5 induced by basic fibroblast growth factor (FGF-2) treatment in MCF-7 cells and that of Her-2/neu in SK-BR-3 cells compared with controls. Experiments measuring passage retardation showed significant differences in passage rates between FGF-2–treated and untreated MCF-7 cells over reaction regions coated with fibronectin and antibody to integrin 5ß1 compared with control regions. Blocking peptides reversed the retardation, demonstrating specificity.
Conclusions: Immunofluorescence detection in a microfluidic channel demonstrates the potential for assaying surface protein expression in a few individual cells and will permit the development of future iterations not requiring cell handling. The flow retardation device represents the first application of this technology for assessing cell-surface protein expression in cancer cells and may provide a way for analyzing expression profiles of single cells without preanalytical manipulation.
|
Introduction
|
|---|
Interactions of cancer cells with their surrounding microenvironments affect growth, differentiation, dormancy, invasion, metastatic potential, and cell death. These interactions depend on the expression patterns of cell-surface proteins. Although mRNA expression profiles have value in prognosis (1)(2) and in identifying differences in pathways in specific cancers(3), the poor correlation between mRNA and protein expression signatures makes it necessary to focus on protein expression to develop links to biologic behavior.
Solid tumors contain heterogeneous cell populations with different malignant potentials partially endowed by the repertoire of surface proteins (4). Individual cells in the tumor behave differently; thus, data averaged over populations may not predict the behavior of individual cells(5). The ability to derive useful information from cells obtained by biopsy or needle aspiration of tumors or micrometastases is limited by the technology for analyzing single cells or small cell populations (<50 cells). Current assays, including fluorescence-activated cell sorting(6), confocal microscopy(7), and plasmon resonance(8)(9), have limitations in processing small cell populations and analyzing the expression and activity of multiple cell-surface proteins.
Handling and characterization of cells can be addressed by microfluidics (10). Microfluidics is potentially useful for monitoring of hormone secretion(11), motility(12), membrane potential(13), intracellular signaling(14), cell death(15), and contraction(16). Microfluidics has been also used to measure molecular events in subcellular compartments(17). Signal analysis has involved use of fluorochromes(15)(17), in-channel fluorescent binding(11)(14), light microscopy(12)(17), acoustic wave sensors(16), fluorescence resonance energy transfer(18), and surface plasmon-coupled emission(19), among others.
We adapted microfluidic technology in 2 ways to measure the expression of proteins on a few cells. One approach involved detection of indirect fluorescence labeling of cellular antigens and the other involved detection based on the Bell model, in which external force alters the binding equilibrium to a ligand (20).
|
Materials and Methods
|
|---|
cells and tissue culture
MCF-7, MDA-MB-231, and SK-Br-3 cells were obtained from the American Type Culture Collection. Cells were cultured in DMEM with 2 mmol/L glutamine, 100 mL/L heat-inactivated fetal calf serum (FCS),
1
50 kilounits/L penicillin, and 50 mg/L streptomycin (Gemini Bio-Products) with or without 10 µg/L recombinant human basic fibroblast growth factor (rhFGF-2; R&D Systems).
assembly of the microfluidic device
Microcapillaries were constructed by use of a PMP-102 Micropipette Puller (MicroData Instruments). The instrument heated a 5-mm capillary tube (VWR Scientific) to 427 °C (800 °F) and pulled it to achieve a constriction with an 
30 µm inner diameter. The pulled capillary tube was mounted on the stage of an Olympus IX 70 microscope, which minimized vibration and breakage. Typically, 20 µL of cell suspension containing 20 000 cells in phosphate-buffered saline (PBS; 137 mmol/L NaCl, 7 mmol/L Na2HPO4, 1.5 mmol/L KH2PO4, 2.7 mmol/L KCl, pH 7.4) was loaded into a 100-µL Hamilton syringe. Cells were evenly suspended in the syringe, pumped into the capillary tube at a flow rate of 120 nL/s, and visualized by use of the Olympus microscope. Fluorescent images were digitized and recorded on a PC using Hewlett Packard Benchmark software. The fluorescence intensity was quantified by a C31034 photomultiplier tube. Data were transferred to an Excel spreadsheet.
immunofluorescence labeling microscopy
Cells were fixed with 37 mL/L formaldehyde in PBS for 10 min. For fibrillar actin labeling, fixed cells were washed twice with PBS, incubated with 0.15 mmol/L BODIPY FL phallacidin (Molecular Probes) in PBS containing 10 g/L bovine serum albumin (BSA) for 20 min at room temperature, washed 3 times with PBS, and mounted in Slow-Fade (Molecular Probes). Antibody labeling was carried out with a primary rabbit polyclonal anti-human integrin 5 or mouse monoclonal anti-Her-2/neu antibody (Santa Cruz Biotechnology) and Texas Red– or fluorescein isothiocyanate–conjugated secondary antibodies (Oncogene Science) diluted in PBS-Tween at 1:200. Typically, 1 x 106 cells were processed for labeling. Fixed cells were blocked with 20 g/L BSA before incubation with antibodies at 37 °C and mounting with Anti-Fade. A portion of the fluorescence-tagged cells was visualized and photographed by use of an Olympus BX40 fluorescence microscope and an Olympus MagnaFire digital photographic system. Remaining cells were used in experiments with the microfluidic device.
western immunoblot detection
Cells in culture at 75% confluence were lysed in modified RIPA buffer and analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, as described previously (21). Membranes were probed with antibodies to human integrin 5 and Her-2/neu (Santa Cruz) and detected by electrochemiluminescence.
flow retardation channel
The flow cell (see Fig. 5A
) was constructed with a demountable top glass plate with "windows" drilled with two 1.59 mm (0.0625 inch) holes to allow fluid flow into and out of the 150-µm space between the top and bottom plates; this spacing was determined by the thickness of a silver spacer (Perkin-Elmer). The bottom plate was a 1 x 3-inch microscope slide coated with protein for the retardation assay. The dimensions of the flow cell were 20 mm x 10 mm x 150 µm. The device was developed with a port for a syringe that was equipped with a WPI Micro 4TM microsyringe pump able to inject a small cell population at a flow of 75 nL/s. To prevent cell–cell adhesion and nonspecific binding, the cells were suspended in standard medium containing 0.1 mL/L Tween 20. The assembled flow cell was mounted on an inverted microscope (Olympus IX 70), and an 200 µm field of view was captured by a video camera that recorded cell passage. Rates of passage were determined by manual frame-to-frame comparisons.
View larger version (10K):
[in this window]
[in a new window]
|
Figure 5. Schematic design of an etched glass channel with a demountable glass top cover with fabricated protein-binding surfaces.
(A), slide covered with waterproof tape mask with thin rectangular cutouts representing the flow lanes to be coated. (B), flow lanes coated with different ligands illustrated after removal of waterproof tape. (C), flow cell with demountable slide coated on the flow retardation array zone. The channel dimensions are 50–100 µm x 50–100 µm.
|
|
protein coating of the retardation channels
Standard 1 x 3 x 0.1-inch microscope slides were prepared with at least 2 side-by-side coatings, accommodating a control coating adjacent to the binding protein of interest, and up to as many as 4 rectangular coatings (Fig. 5
). Retardation rates in the presence of multiple binding proteins or controls were monitored in the same experiment. A few monolayers of protein were sufficient for binding, especially if the linker allowed extension of the binding protein from the surface of the glass. Reproducibility in the coating thickness on the glass plates or on witness plates was achieved by monitoring with a 3-wavelength (HeNe) ellipsometer (AutoEL®-III Automatic Ellipsometer; Rudolph Research) and was critically dependent on having a contamination-free glass surface. Thus, cleaning was a critical step. Cleaning was carried out with SC1 cleaning solution, consisting of water–hydrogen peroxide–ammonium hydroxide (5:1:1 by volume) (22).
The binding layer was prepared by coating with poly-L-lysine treated with glutaraldehyde. We monitored the coating thickness separately by incorporating fluorescent material, as above. The slides were then incubated for 1 h with 5 mg/L binding protein solutions to adsorb them to this coating. The protocol for binding proteins to glass, while wearing a mask, is as follows:
- (a) Clean the glass slide with 700 mL/L ethanol and let air dry
- (b) Cover the slide with a waterproof tape (e.g., Sigma SealPlate film) to form a mask with thin rectangular cutouts representing the flow lanes to be coated, ensuring even and firm adherence in all places
- (c) Incubate the exposed cutouts with a 0.1 g/L solution of poly-L-lysine HBr at room temperature for 30 min
- (d) Wash with distilled water
- (e) Incubate the exposed glass strips with a 25 mL/L glutaraldehyde solution for 1 h
- (f) Wash with distilled water
- (g) Incubate the individual exposed glass strips for 1 h with 5 mg/L solutions of the desired proteins, making certain not to let any 2 protein solutions mix
- (h) Wash with distilled water
- (i) Immerse the slide with tape still attached in a 0.2 mol/L solution of glycine and let sit at least 12 h at 4 °C, and for storage until ready for use
Control experiments demonstrated that adsorbed proteins were not desorbed from the surface at the flow rates used. A linking protocol reported by Delamarche et al. (23) can be used if desorption becomes an issue. To block nonspecific binding, medium containing 100 mL/L FCS and 0.1 mL/L Tween 20 was pumped into the chamber before the introduction of cells. MCF-7 cells used for the assay were treated overnight with 10 µg/L FGF-2, detached by trypsinization, fixed in 37 mL/L formaldehyde, and suspended in DMEM containing 100 mL/L FCS and 0.1 mL/L Tween 20.
|
Results
|
|---|
detection of fluorescence-labeled cellular proteins
The microfluidic device consisted of a microcapillary tube, a syringe pump, a microscope, a photomultiplier tube, and a computer for data acquisition. The current level of complexity required cell labeling with a fluorochrome before analysis. The microcapillary tube had an inner diameter of 300 µm and a constricted portion with an inner diameter of 30 µm (Fig. 1A
). The constriction generated an increase in pressure and caused the cells to assume a single file as they entered this segment. A Venturi tube measured pressure differences. Because of the relationship of flow velocity to tube diameter, the rate of cell passage was too high for measurement in the constricted region but slowed considerably when the cells exited the constriction and remained constant for 1000 µm, permitting reproducible measurements. For the images in Fig. 1B
, the objective was focused on a phallacidin-labeled cell passing the widening segment of the microcapillary tube (boxed cell in Fig. 1A
), and the cell was photographed by light and fluorescence microscopy. Cell velocity through different segments of the microcapillary tube was directly proportional to fluid velocity and the area of the segment cross-section. The predicted and actual cell velocities corresponding to variable flow rates in the wide portion of the microcapillary tube are shown in Fig. 1 of the Data Supplement accompanying the online version of this article at http://www.clinchem.org/content/vol52/issue6/. The system enabled the collection of 4–5 data points in 0.5 s with a 40x objective.
View larger version (63K):
[in this window]
[in a new window]
|
Figure 1. Schematic of the microfluidic device (A), and micrographs of a cell exiting the constricted portion of the microcapillary (B).
(A), fluorescently labeled cells in suspension are introduced into a microcapillary tube by a Hamilton syringe and propelled by a syringe pump. The inner diameter of the tube, 300 µm, is narrowed to 30 µm, where cells assume a single file. As the cells exit the constriction, their flow rate slows considerably, and they are detected by the 40x objective of an Olympus IX 70 microscope connected to a photomultiplier. (B), light and fluorescence micrographs, taken 0.2 s apart, of a phallacidin-stained cell as it exits the constricted portion of the microcapillary (boxed cell in panel A).
|
|
To calibrate sensitivity, we correlated the voltage generated by the photomultiplier with variable concentrations of phallacidin in a defined volume in the objective field. As shown in Fig. 2A
, the relationship between the calculated number of phallacidin molecules and voltage generated was linear, with a best fit line through the data points defined by y = 4 x 10–8x + 4.0481, with y as the relative microvoltage (µV) and x as the number of molecules (R2 = 0.9716). This approximation was applied to estimate the number of phallacidin molecules incorporated into the MCF-7 cells (Fig. 2B
). Detectable signals generated by phallacidin-labeled MCF-7 cells corresponded to 1 to 10 x 106 phallacidin molecules per cell. The photomicrographs in Fig. 2 of the online Data Supplement illustrate the relevant differences in actin staining of phallacidin-labeled MCF-7 cells.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 2. Voltage intensity corresponding to number of phallacidin molecules in the acquisition field.
(A), solutions containing variable dilutions of phallacidin were placed in an orthogonal tube with 75 µm x 75 µm edges in its cross-section (y and z dimensions), with 133.35 µm of its length (x dimension) within the field of the 40x objective. Multiple measurements of the voltage generated at each concentration were obtained and correlated with the voltage obtained with the phallacidin concentration. The graph depicts changes in voltage intensity with respect to the number of phallacidin molecules. (B), voltage peaks generated in the photomultiplier by MCF-7 cells, labeled with BODIPY FL phallacidin, passing through the microfluidic device.
|
|
We tested the capacity of the device to detect differential expression of 2 cell-surface proteins, integrin 5 and the much lower abundance tyrosine kinase receptor Her-2/neu, in 3 breast cancer cell lines. The expression of integrin 5 on MCF-7 breast cancer cells, as measured by Western blotting, was modest but was induced to high amounts by a 24-h incubation with 10 µg/L recombinant FGF-2, as demonstrated previously (24) (see Fig. 3 of the online Data Supplement). Integrin 5 expression in SK-Br-3 cells was also low. However, Her-2/neu expression was detectable in SK-Br-3 cells, which have an amplified gene copy number(25), but not in MCF-7 or MDA-MB-231 cells, which were used as negative controls. Immunofluorescence microscopy confirmed that incubation of MCF-7 cells with 10 µg/L FGF-2 for 24 h induced the expression of integrin 5 on the cell surfaces (Fig. 3
, right-hand panels) and that SK-Br-3 cells, but not MCF-7 or MDA-MB-231 cells, expressed Her-2/neu (Fig. 4
, right-hand panels).
View larger version (32K):
[in this window]
[in a new window]
|
Figure 3. Detection of indirect immunofluorescence signal from anti-integrin 5 antibody-labeled MCF-7 cells.
Cells were treated with medium with (bottom) or without (top) 10 µg/L FGF-2 for 24 h and labeled with an anti-integrin 5 rabbit polyclonal antibody and a goat anti-rabbit Texas Red–tagged secondary antibody. Immunofluorescence (top of each pair of micrographs at the right) and light microscopy (bottom of each pair of micrographs at the right) reveal labeling of FGF-2–treated MCF-7 cells but only background fluorescence in untreated MCF-7 cells. Cells assayed by the microfluidic device are recognized as peaks in the tracing (left) as they pass the objective linked to the photomultiplier.
|
|
View larger version (37K):
[in this window]
[in a new window]
|
Figure 4. Indirect immunofluorescence signal from MCF-7 (top), MDA-MB-231 (middle), and SK-Br-3 cells (bottom) labeled with anti-Her-2/neu antibody.
Cells were labeled with mouse monoclonal antibody to Her-2/neu and with a rabbit anti-mouse Texas Red–tagged secondary antibody. Immunofluorescence (top of each pair of micrographs at the right) and light microscopy (bottom of each pair of micrographs at the right) demonstrates labeling of Sk-Br-3 cells but only background fluorescence in MCF-7 and MDA-MB-231 cells. Cells assayed by the microfluidic device are recognized as peaks in the tracing (left) as they pass the objective linked to the photomultiplier.
|
|
In the microfluidic device, the fluorescent antibody-tagged cells appeared as voltage peaks above the background tracing of the carrier solution in the flow channel. We used 1000 background points between peaks in the segments shown to obtain the mean and SD for the baseline. Signals for integrin 5 from untreated MCF-7 cells (Fig. 3
) and for Her-2/neu in MCF-7 and MDA-MB-231 cells (Fig. 4
) that were 1 SD above baseline were counted. Because only positive peaks were captured, differences between baseline and peak means, as determined by heteroscedastic t-test, were highly significant for all cells (Table 1
). However, because the device in its current inception did not account physically for every cell whose fluorescence was captured by the photomultiplier, the distributions of the integrin 5 peaks in the untreated MCF-7 cells (Fig. 3
) and of the Her-2/neu peaks in MCF-7 and MDA-MB-231 cells (Fig. 4
) were not gaussian, but were skewed to the higher values detectable above background, whereas most of the peaks were hidden within the background tracing. In contrast, the distributions of the integrin 5 peaks in the FGF-2–treated MCF-7 cells (Fig. 3
) and of the Her-2/neu peaks in Sk-Br-3 cells (Fig. 4
) represented the majority of cells in these normally distributed populations and captured them as fluorescence peaks above background. Nevertheless, the absolute differences between the integrin 5 signal and baseline were 1.2 µV for FGF-2–treated cells and 0.46 for untreated cells, values that were significantly different by t-test. Similarly, the differences between the mean values for the Her-2/neu peaks and baseline were 0.602 for MCF-7, 0.441 for MDA-MB-231, and 1.271 µV for SK-Br-3 cells, with the latter being significantly greater than either of the other two. These data demonstrate the principle that such a microfluidic device can assay the expression of cell-surface proteins.
tagless detection of cell surface proteins
We developed a model for flow retardation of cancer cells in a channel based on reversible interactions between cell-surface proteins and immobilized ligands in the channel. The force of flow was used to reduce the formation of a binding complex, AB, and increase the formation of an interaction complex (AB)*. The use of external force to alter cell-binding equilibrium conditions was first described by George Bell in 1978 (20), wherein the equilibrium of binding can be shifted to favor the encounter, or interaction complex (AB)*, by the force imparted to the cell by flow given by Stokes’ law:
where F is the force (dynes), 
is the fluid viscosity (g/cm · s), r is the cell radius (µm), and v is the fluid stream velocity (cm/s). Experimentally measured retardation of cells by binding proteins is thus a measure of the formation of the encounter, or interaction complex. Appropriate controls with either soluble ligand or irrelevant antibody adsorbed to the device can be used to determine the presence of a protein on the surface of the cell. Coating of individual channels is described in the Materials and Methods section (Fig. 5A
). An example of a slide coated with different proteins in 4 zones is shown in Fig. 5B
, and a schematic of the demountable system is illustrated in Fig. 5C
.
To test the system, we directed the flow of FGF-2-treated MCF-7 cells over a slide with parallel zones variably coated with fibronectin or BSA, or an uncoated zone (reagents only), and the rate of flow was measured with a video recorder. A total of 50 cells were sampled flowing over each zone of the slide. Assuming gaussian distributions, the mean (SD) passage rates were 12.84 (1.92) µm/s over BSA and 13.00 (1.76) µm/s over the uncoated region, compared with 10.01 (1.54) µm/s over fibronectin, a value that was significantly different from either control (P <0.001 for both controls, Student t-test; Fig. 6
). This experiment was repeated several times with the same results. The experiments were repeated concurrently with cells not treated with FGF-2 to demonstrate differences in adhesion based on induction of integrins 5ß1. Passage rate measurements over a patch coated with fibronectin and one left blank with only a polylysine coating demonstrated statistically significant retardation of cells treated with FGF-2 (P <0.001) as well as of the untreated cells (P <0.01). However, the ratio of passage rates on polylysine compared with those on fibronectin were significantly greater with FGF-2–treated cells than controls (P <0.001), corresponding to the increase in cell-surface expression of integrin 5ß1 demonstrated previously (Table 2A
). This is an encouraging effect that suggests potential quantitative applications of the system. In addition to 2 adjacent coatings, up to 4 large square or rectangular coatings have been prepared such that retardation in the presence of multiple binding proteins, or controls, may be achieved in the same experiment, as shown by the data in Table 2B
. In this experiment, FGF-2–treated cells demonstrated highly significant retardation by fibronectin and an anti-integrin 5ß1 antibody but not by BSA or control IgG. The interaction between fibronectin and integrin 5ß1 leading to retardation of FGF-2–treated cells was specific; a fibronectin-blocking peptide, GRGDSP, reversed the effect, whereas a nonspecific peptide did not (Table 2C
). These experiments demonstrate that the device is functional.
View larger version (11K):
[in this window]
[in a new window]
|
Figure 6. Flow rate distributions of MCF-7 cells treated with 10 µg/L FGF-2 for 24 h over channels coated with fibronectin, BSA, and poly-L-lysine (blank).
|
|
|
Discussion
|
|---|
We investigated the development and use of microfluidic approaches to analyze the expression of surface proteins in cancer cells. In one approach, we adapted a methodology for handling cells labeled by direct or indirect fluorescence, introduction into a microcapillary tube, and optimization of delivery rates and detection sensitivity. The sensitivity was sufficient to detect fluorescence signals in cells expressing even very low amounts of a surface antigen. However, in these control cells, the signal intensity distribution curve was skewed to higher values, indicating that a significant fraction of cells in a population with presumably normally distributed signal intensities had a signal below the baseline noise. In contrast, cells expressing high concentrations of surface antigen had a gaussian distribution of fluorescence peak intensities, indicating that most of these cells were captured by the assay. The statistical relationship between the expression of surface antigens in 2 populations of cells was determined. Video tracking software developed at Sarnoff Corporation allows tracking of up to hundreds of individual cells as they traverse different zones of the device. Future iterations of the device could permit on-chip labeling, phase-contrast light microscopic covisualization, quenching of fluorophores, and recycling and relabeling of cells with additional fluorescent ligands of cellular antigens to extensively characterize individual cells within a small population without prior handling, making it useful for analyzing a few cells.
We used this cell-handling and delivery technology to develop a novel concept of tagless detection of cell surface antigens. This approach was based on quantitative flow retardation of cells by transient binding of surface proteins to ligands immobilized to patches of the channel. This approach, in its ultimate iteration, may be able to query the expression of multiple proteins in the same cell with virtually no handling after aspiration from a tumor. The methodology had to strike a balance between nonspecific cell–cell and cell–channel binding and appropriate reversible binding to immobilized ligands. Greenberg et al. (26)(27) and Pierres et al.(28) showed that at relatively low flow (4–6 µL/s) in flow cells of 17 x 6 x 0.2 mm3 at a calculated wall shear rate of 125 s–1(29), complex retardation events occur in a time frame of seconds to milliseconds. The shear rate G was calculated by the formula: G = 6Q/wh2, where Q is the flow rate, and w and h are the width and height of the chamber(29). Our data show that retardation of fixed MCF-7 cells occurs in microfluidic cells of similar dimensions, 20 mm x 10 mm x 150 µm (length x width x height), but with lower flow rates of 75 nL/s and shear rates in the range of 2 s–1. These flow rates translate to rolling of cells over the coated patches on the order of 13–15 µm/s, at which retardation by 20%–25% is sufficient to detect differences in surface protein expression. Adding 0.1 mL/L Tween 20 to the complete medium containing 100 mL/L FCS overcame issues with cell clumping, channel clogging, and nonspecific and irreversible adhesion to the flow surfaces while maintaining surface antigen integrity and improving the fraction of cells scored.
The flow cell is mounted on an inverted microscope with a video camera that records cell passage past an 200-µm field of view. Rates are determined by frame-to-frame comparisons, currently by manual means. However, automated digital methods have been reported (28)(30), and digital data-processing software has also been developed at one of our sites (Sarnoff Corporation). In future studies, the images will be digitized, and up to hundreds of individual cells will be tracked as they traverse the multiple interaction zones.
The standard deviations of the flow rate distributions reflect the heterogeneity of surface protein expression in cell populations. Although the mean cell velocity over fibronectin is statistically different from that in the control region, only a fraction of the cells have a velocity slower than every cell in the control population (Fig. 6
). This is likely attributable to the normal biological variability in the expression of integrin 5ß1 in this population, which endows only some of the cells with the threshold of expression needed to retard flow. Modulation of solvents, surfaces, and flow rates may narrow the width of the distribution, but ultimately, an alternative detection technology that measures instantaneous interaction between immobilized ligands and cellular receptors could permit us to resolve retardation in a shorter time scale than is attainable with the present system, i.e., would detect weak binding events. Future studies will be aimed at correlating quantitative retardation with receptor number measured by physical means. For now, the observed, reproducible, specific retardation of FGF-2–treated MCF-7 cells with ligands of integrin 5ß1 represents a breakthrough in a novel, tagless detection method and will allow development of the technology as a research and diagnostic tool.
|
Acknowledgments
|
|---|
This work was supported by the New Jersey State Commission on Cancer Research (02-1140-CCR-E0 to R.W. and J.Y.) and the Department of Defense (DAMD17-03-1-0524 to R.W.).
|
Footnotes
|
|---|
1 Nonstandard abbreviations: FCS, fetal calf serum; FGF-2, basic fibroblast growth factor; PBS, phosphate-buffered saline; and BSA, bovine serum albumin. 
|
References
|
|---|
- van de Vijver MJ, He YD, van’t Veer LJ, Dai H, Hart AAM, Voskiul DW, et al. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 2002;347:1999-2009.[Abstract/Free Full Text]
- van’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AAM, Mao M, et al. Gene expression profiling predicts clinical outcomes of breast cancer. Science 2002;415:530-536.
- Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffery SS, Rees CA, et al. Molecular portraits of human breast tumours. Science 2000;406:747-752.
- Ruoslahti E, Rajotte D. An address system in the vasculature of normal tissues and tumors. Annu Rev Immunol 2000;18:813-827.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Levsky JM, Shenosy SM, Pezo RC, Singer RH. Single-cell gene expressionprofiling. Science 2002;297:836-840.[Abstract/Free Full Text]
- Herzenberg LA, Parks D, Sahaf B, Perez O, Roederer M, Herzenberg LA. The history and future of the fluorescent activated cell sorter and flow cytometer: a view from Stanford. Clin Chem 2002;48:1819-1827.[Abstract/Free Full Text]
- Klonis N, Rug M, Harper I, Wickham M, Cowman A, Tilley L. Fluorescence photobleaching analysis for the study of cellular dynamics. Eur Biophys 2002;31:36-51.
- Liparoto SF, Ciardelli TL. Biosensor analysis of the interleukin-2 receptor complex. J Mol Recognit 1999;12:316-321.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Weinberger SR, Morris TS, Pawlak M. Recent trends in protein biochip technology. Pharmacogenomics 2000;1:395-416.[CrossRef][Medline]
[Order article via Infotrieve]
- Huh D, Gu W, Kamotani Y, Grotberg JB, Takayama S. Microfluidics for flow cytometric analysis of cells and particles. Physiol Meas 2005;26:R73-R98.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Roper MG, Shackman JG, Dahlgren GM, Kennedy RT. Microfluidic chip for continuous monitoring of hormone secretion from live cells using an electrophoresis-based immunoassay. Anal Chem 2003;75:4711-4717.[Medline]
[Order article via Infotrieve]
- Jeon LN, Baskaran H, Dertinger SK, Whitesides GM, Van de Water L, Toner M. Neutrophil chemotaxis in linear and complex gradients of interleukin-8 formed in a microfabricated device. Nat Biotechnol 2002;20:826-830.[ISI][Medline]
[Order article via Infotrieve]
- Farinas J, Chow AW, Wada HG. A microfluidic device for measuring cellular membrane potential. Anal Biochem 2001;295:138-142.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Yang M, Li C. W, Yang J. Cell docking and on-chip monitoring of cellular reactions with a controlled concentration gradient on a microfluidic device. Anal Chem 2002;74:3991-4001.[Medline]
[Order article via Infotrieve]
- Chan SDH, Luedke G, Valer M, Buhlmann C, Preckel T. Cytometric analysis of protein expression and apoptosis in human primary cells with a novel microfluidic chip-based system. Cytometry 2003;55A:119-125.[Medline]
[Order article via Infotrieve]
- Li PC, Wang W, Parameswaran Ash M. An acoustic wave sensor incorporated with a microfluidic chip for analyzing muscle cell contraction. Analyst 2003;128:225-231.[CrossRef][Medline]
[Order article via Infotrieve]
- Takayama S, Ostuni E, LeDuc P, Naruse K, Ingber DE, Whitesides GM. Selective chemical treatment of cellular microdomains using multiple laminar streams. Chem Biol 2003;10:123-130.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Buranda T, Huang J, Perez-Luna VH, Schreyer B, Sklar LA, Lopez GP. Biomolecular recognition on well-characterized beads packed in microfluidic channels. Anal Chem 2002;74:1149-1156.[Medline]
[Order article via Infotrieve]
- Lakowicz JR, Malicka J, Gryczynski I, Gryczynski Z. Directional surface plasmon-coupled emission: a new method for high sensitivity detection. Biochem Biophys Res Commun 2003;307:435-439.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Bell GI. Models for the specific adhesion of cells to cells. Science 1978;200:618-627.[Abstract/Free Full Text]
- Fenig E, Wieder R, Paglin S, Wang H, Persaud R, Haimovitz-Friedman A, et al. Basic fibroblast growth factor confers growth inhibition and mitogen-activated protein kinase activation in human breast cancer cells. Clin Cancer Res 1997;3:135-142.[Abstract]
- Kern W, Puotinen DA. Cleaning solutions based on hydrogen peroxide for use in silicon semiconductor technology. RCA Rev 1970;31:187-205.
- Delamarche E, Bernard A, Schmid H, Michel B, Biebuych H. Patterned delivery of immunoglobulins to surfaces using microfluidic networks. Science 1997;276:779-781.[Abstract/Free Full Text]
- Korah R, Boots M, Wieder R. Integrin 5ß1 promotes survival of growth-arrested breast cancer cells: an in vitro paradigm for breast cancer dormancy in bone marrow. Cancer Res 2004;64:4514-4522.[Abstract/Free Full Text]
- Szollosi J, Balazs M, Feuerstein BG, Benz CC, Waldman FM. ERBB-2 (HER2/neu) gene copy number, p185HER-2 overexpression, and intratumor heterogeneity in human breast cancer. Cancer Res 1995;55:5400-5407.[Abstract/Free Full Text]
- Greenberg AW, Kerr WG, Hammer DA. Relationship between selectin-mediated rolling of hematopoietic stem and progenitor cells and progression in hematopoietic development. Blood 2000;95:478-486.[Abstract/Free Full Text]
- Greenberg AW, Hammer DA. Cell separation mediated by differential rolling adhesion. Biotechnol Bioeng 2001;73:111-124.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Pierres A, Benoliel A-M, Bongrand P. Measuring the lifetime of bonds made between surface-linked molecules. J Biol Chem 1995;270:26586-26592.[Abstract/Free Full Text]
- Pierres A, Benoliel AM, Zhu C, Bongrand P. Diffusion of microspheres in shear flow near a wall: use to measure binding rates between attached molecules. Biophys J 2001;81:25-42.[Medline]
[Order article via Infotrieve]
- Perret E, Benoliel A-M, Nassoy P, Pierres A, Delman V, Thiery J-P, et al. Fast dissociation kinetics between individual E-cadherin fragments revealed by flow chamber analysis. EMBO J 2002;21:2537-2546.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
Top
Abstract
Introduction
