Subscriber access provided by NORTH CAROLINA STATE UNIV
Article
Whole cell quenched flow analysis Ya-Yu Chiang, Sina Haeri, Carsten Gizewski, Joanna D. Stewart, Peter Ehrhard, John Shrimpton, Dirk Janasek, and Jonathan J. West Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 31 Oct 2013 Downloaded from http://pubs.acs.org on November 4, 2013
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Whole cell quenched flow analysis Ya-Yu Chiang1,2, Sina Haeri3, Carsten Gizewski4, Joanna D. Stewart5, Peter Ehrhard4, John Shrimpton3, Dirk Janasek1 and Jonathan West1,2*
1
Leibniz-Institut für Analytische Wissenschaften - ISAS - e.V., 44227 Dortmund, Germany
2
Institute for Life Sciences, University of Southampton, SO17 1BJ, U.K.
3
Engineering and the Environment, University of Southampton, SO17 1BJ, U.K.
4
Fluid Mechanics, Biochemical and Chemical Engineering, Technische Universität, 44221
Dortmund, Germany
5
Leibniz Research Centre for Working Environment and Human Factors – IfADo, 44139
Dortmund, Germany
KEYWORDS: Microfluidics, deterministic lateral displacement, single cell analysis, signal transduction, multi-phase coupling, micromixing, autophosphorylation and insulin-like growth factor receptor.
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
ABSTRACT This paper describes a microfluidic quenched flow platform for the investigation of ligandmediated cell surface processes with unprecedented temporal resolution. A roll-slip behaviour caused by cell–wall–fluid coupling was documented and acts to minimize the compression and shear stresses experienced by the cell. This feature enables high velocity (100–400 mm/s) operation without impacting the integrity of the cell membrane. In addition, rotation generates localized convection paths. This cell-driven micromixing effect causes the cell to become rapidly enveloped with ligands to saturate the surface receptors. High speed imaging of the transport of a Janus particle and fictitious domain numerical simulations were used to predict millisecond-scale biochemical switching times. Dispersion in the incubation channel was characterised by microparticle image velocimetry and minimised by using a horizontal HeleShaw velocity profile in combination with vertical hydrodynamic focusing to achieve highly reproducible incubation times (CV = 3.6%). Microfluidic quenched flow was used to investigate the pY1311 autophosphorylation transition in the type I insulin-like growth factor receptor (IGF-1R). This pre-dimerized receptor undergoes autophosphorylation within 100 ms of stimulation. Beyond this demonstration, the extreme temporal resolution can be used to gain new insights into the mechanisms underpinning a tremendous variety of important cell surface events.
INTRODUCTION Extracellular biochemical stimuli are perceived by membrane hosted receptors which respond by initiating molecular signalling cascades. Transduction mechanisms include conformational and autocatalytic modifications, such as tyrosine autophosphorylation, and occur over millisecond timescales. However, these timescales are currently inaccessible owing to an 2 ACS Paragon Plus Environment
Page 2 of 33
Page 3 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
absence of techniques for the rapid stimulation of receptors in the context of the cell membrane (the optimal biophysical environment for receptor function). Consequently the sequences and mechanisms underlying signal transduction remain poorly characterised, thereby hindering the rational design of cell surface targeting therapeutics. New techniques are therefore required to extend the cell biology toolkit to make accessible the millisecond timescales of processes occurring at the cell surface. In recent years the spatial precision of microfluidics has begun to impact biology1 and promises to deliver the temporal precision necessary to investigate ultrafast processes in cell biology. Quenched flow technology can be used to analyse the millisecond-scale kinetics of biomolecular interactions.2 The method involves the use of high pressure pumps to drive the rapid, turbulent mixing of reactive (bio)molecular partners, followed by short incubation periods defined by the length of a delay line before rapid reaction arrest by turbulent mixing with a quenching buffer. Unlike stopped flow3 this bypasses the need for distinguishable spectral signatures that may not exist in macromolecular systems. Instead the full repertoire of analytical methods is available for off-line identification of the reactions intermediate states. The use of turbulent flows enables mixing in milliseconds for the identification of reaction intermediates and the elucidation of the underlying mechanisms. A wide variety of molecular processes are hosted within the complex microenvironment of the cell membrane. For example, ligand-mediated signal transduction by membrane-spanning receptors is critical for health, with dysfunction leading to disease. Such processes are fast, occurring over sub-second timescales. The mechanisms and dynamics of these processes should be studied in the native state of the membrane to provide an authentic biophysical environment and ideally using whole cells to accommodate the array of all possible interaction partners in both membrane and cytosolic compartments. However, the turbulent, high shear conditions necessary for rapid mixing in quenched flow analysis result in 3 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Ya-Yu Chiang et al Whole cell quenched flow analysis
the disruption of cell membranes. This prevents the meaningful investigation of membranehosted reactions and consequently the systems-level molecular dynamics of many cell surface events remain poorly resolved. To tackle this challenge a gentle method for near-instantaneous cell surface stimulation followed by near-instantaneous reaction arrest is required to preserve the reaction intermediates for post-treatment analysis. The micron-scale size of cells affords the opportunity to use microfluidic processing techniques. However, at microscopic scales flows are characteristically laminar, with mixing driven by diffusion (supporting information, Figure S1). The convergence of cell and ligand streams leads to the formation of a virtual interface with membrane-hosted ligand–receptor interactions delayed by diffusion timescales.4 Advection using either microstructured channel walls5 or circulation within plug flow reactors6 can achieve mixing in milliseconds. However, mixing is shear-driven, undesirable for processing intact cells. A conceptually more straightforward approach involves deflecting cells across the virtual interface between laminar flows and into a ligand-doped adjacent stream with the process repeated downstream for reaction arrest within an adjacent quenching buffer stream. Optical,7 acoustic8 and dielectric9 mobilisation options are available. However, only pN forces can be imparted as a result of unfavourable scaling at cell dimensions and weak material contrasts between the cell and the aqueous media. Cell-length translation for switching biochemical streams therefore requires several 10’s of milliseconds, inadequate to resolve most molecular processes. Instead, hydrodynamic methods such as deterministic lateral displacement (DLD) can be used. This approach scales favourably enabling the rapid deflection of cells into neighbouring streams.10 Here, angled micropillar arrays are used to compress laminar streams to sub-cellular dimensions for the sequential deflection of cells
4 ACS Paragon Plus Environment
Page 4 of 33
Page 5 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
from their original path. Alternatively, inertial focusing can be used for whole cell translation in milliseconds.11 In this paper we present a whole cell microfluidic quenched flow analysis system which involves a two-step DLD arrangement based on a pinched-flow approach12 using socalled stream-thinning elements (STEs).13 We previously developed a prototype and used this to discover a novel biphasic epidermal growth factor receptor (EGFR) signal transduction mechanism.13 In this contribution we extend beyond this demonstration, using experimental and numerical simulation methods to investigate the operational velocity limit and the temporal resolution of the activation and incubation reaction phases. A cell–wall–fluid coupling phenomena was observed which produces roll-slip cell transport for micromixing and ligand envelopment over millisecond time scales. The system was used to investigate the rapid autophosphorylation (pY1131) state-switching dynamics of type I insulin-like growth factor receptor (IGF-1R).
EXPERIMENTAL SECTION Design, Fabrication and Packaging The dimensions of the microfluidic quenched flow system are defined by the size of the cells. In this study, human epithelial carcinoma cells, HeLa S3, with a suspension diameter of 15.8 SD ± 2.2 µm were used. The stream thinning elements (STEs) used for the convergence of cell and ligand or quench buffer streams were 25-µm-wide and 200-µm-long. Twenty-fold channel expansion at 120° produced a 500-µm-wide incubation channel, with lengths corresponding to 100 ms to 2.00 s. Mirrored channel bifurcation was used to balance the system’s fluidic resistance and reduce cell dilution. Both paths contained a second STE for reaction arrest. A quenched flow device is shown in Figure 1. 5 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 33
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 1. Microfluidic Circuit. Whole cell quenched flow analysis system replicated in PDMS. Microchannel edges have been highlighted using black lines. A STE is used for cell stimulation, with a second used for reaction arrest at the end of the incubation channel. Inset, a flow ratio of 1:7 was used for single cell displacement into a dye-doped stream. The arrows denote the cell transport direction.
The microfluidic quenched flow circuit was prepared by replica moulding in poly(dimethyl siloxane) (PDMS, Sylgard 184, Dow Corning) using a 60-µm-high SU-8 master fabricated by standard photolithographic methods (see Frimat et al).14 A biopsy puncture (Kia Medical) was used to produce 1-mm-diameter inlet and outlet ports. A thin (~300 µm) PDMS layer prepared on a glass coverslip was used to encapsulate the microchannels and support the system. For vertical focusing a lift channel, also 60 µm in depth, was replicated in this basal PDMS layer. A pressure tolerant PDMS–PDMS seal was achieved by plasma bonding (70 W, 40 kHz (Femto, Diener Eletronic, Germany)) in a 0.2 mbar oxygen atmosphere for 45 s. Immediately after plasma bonding, a 2 µL volume of polyL-lysine-g-polyethylene glycol (PLL-g-PEG (100 µg/mL), Surface Solutions, Switzerland) was applied to the inlet ports for capillary-driven filling and the electrostatic derivatization of the microchannel surfaces. The PEG moiety resists protein adsorption15 and prevents the narrow STEs being clogged with cells.
6 ACS Paragon Plus Environment
Page 7 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Operation and Flow Imaging The microfluidic system was connected to syringe pumps (e.g. neMESYS, Centoni GmbH) via PFA tubing (OD 1/16”). 1-cm-long steel tubing (18G ×1 1/2”) sections inserted into the inlets and outlets were used to enable plug and play operation. The syringes containing the cell, ligand and quenching reagents were maintained at 37°C using a custom thermoelectric Peltier control system (Newport Corporation, Irvine, USA) involving syringes jacketed by water-cooled copper heat sinks. A 1:7 cell to ligand or quench reagent volumetric flow ratio was used for experiments. Flow patterns were imaged using 1-µm-diameter Nile red fluorescent microspheres (Invitrogen). To simulate cell trajectories 10- and 20-µm-diameter polystyrene particles (Kisker Biotech GmbH & Co. KG, SD Qcs/Qtotal). Together with the parabolic flow field a 4.0-µm-wide cell stream is produced, resulting in the mean cell centre being displaced 3.9 µm (i.e. r - wcs) across the virtual interface into the ligand stream. Displacement is amplified 20-fold in the 500-µm-wide 10 ACS Paragon Plus Environment
Page 10 of 33
Page 11 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
incubation channel, with the displacement D of each cell across the virtual interface expressed as:
D = (r − wCS ) ×
wexp wSTE
where w denotes the width of the expansion channel (exp) or the stream thinning element (STE). In theory HeLa cells are displaced 78 µm across the interface. However, in practice the HeLa cells are displaced by 46.5 µm (SD ± 15.7). We attribute this to the loss of sharp microchannel corners during SU-8 fabrication. Nevertheless all cells are fully submerged in the ligand stream. Indeed, 10-fold amplification using a 250-µm-wide incubation channel also ensures displacement (26.1 µm SD ± 16.4) of the entirety of all cells into the ligand stream.
Cellular and Microfluidic Speed Limit Displacement is velocity-independent. To effectively capture cell surface events it is necessary to operate at velocities which do not disrupt the cell membrane and which produce predictable flows and thus predictable incubation times. The peak velocity occurs in the reaction arrest STE. Here, channel bifurcation and the combination of flows at a volumetric ratio of 1:7 produces a mean velocity 4-fold higher than the stimulation STE velocity. Mean reaction arrest STE velocities from 10 to 500 mm/s were evaluated for their effect on the integrity and viability of HeLa cells. Even at the extreme velocities >90% of cells remained intact (i.e. excluded trypan blue, Figure 2(A)). Cells cultured overnight were found to be viable and had normal morphology. In contrast, the turbulent mixing character (typical Re ≥10,000) of the conventional quenched flow analysis technique exerts destructive mechanical stresses on the cells, damaging mammalian cell membrane and lysing the cell. Exceptions are rugged cells such as Paramecium cells that are protected by a rigid yet elastic pellicle.18 11 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 2. Speed Limit. The impact of STE mean flow velocity on membrane integrity as measured using the trypan blue assay (A). Expansion from the STE to the incubation path causes sudden deceleration of the flow. These inertial effects manifest as Moffatt eddies above a STE velocity of 400 mm/s. Particle trajectory image revealing the vortex generated using a mean STE velocity of 500 mm/s (B).
Flows in the microfluidic circuit are characteristically laminar. However, 120° channel expansion to 10- and especially 20-fold widths causes rapid deceleration. At sizeable Re numbers vortex behaviour or so-called Moffatt eddies are produced.19 The experimental and numerical studies by Durst et al20 and Fearn et al21 provide a detailed explanation of vortex occurrence at suddenly expanding symmetrical channels at low Re numbers. In our quenched flow circuit Moffatt eddies are observed when using ≥500 mm/s flow velocities in the STEs (channel Re = 20; Figure 2(B)). Cells can become trapped for transient or lengthier periods within the circulating flow patterns. Although this feature can be exploited for cell collection,22 in the microfluidic quenched flow system this delays biochemical switching and results in undefined incubation times. A reaction arrest STE velocity of 400 mm/s (and associated 100 mm/s stimulation STE velocity) was therefore used for subsequent experiments to avoid the formation of Moffatt eddies (supporting information, Figure S3). 12 ACS Paragon Plus Environment
Page 12 of 33
Page 13 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
The flow and particle Reynolds numbers for the different regions of the circuit are also documented in the supporting information (Table S1).
Cell Deformation and Shear Stress The membrane integrity and cell viability data demonstrates that the cells are minimally affected by high velocity transport through the STEs. To obtain an understanding of the mechanical stresses the cell experiences during stream switching in the stimulation STE highspeed imaging was used in combination with numerical simulations. The velocity field simulated in Figure 3(A) causes the cell to impact the microchannel wall upon entry to the STE. The elastic nature of the cell translates the collisional forces into deformation followed by relaxation, restoring the cell’s spherical character (see Figure 3(B-D)).23 The shear stress oscillations correspond to the interactions of the different regions of the cell with the velocity field and the surface of the channel wall. Importantly, deformation is transient such that upon exiting the STE the cells are fully displaced (on average 3.9 µm) for complete submersion in the ligand stream. In contrast, classical deterministic lateral displacement involves multiple micropillar collisions, which enable deformation and size to be used as fractionation parameters.23
13 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 14 of 33
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 3. Cell Stresses. Forces experienced during passage through the STE with a mean flow velocity of 100 mm/s. The velocity field (A), shear stress (B), cell deformation (C, D; y axis (red), x axis (blue)) and mean normal stress (i.e. compression, E).
14 ACS Paragon Plus Environment
Page 15 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
The magnitude of the shear and mean normal stresses were estimated by numerical simulation using bespoke algorithms involving an implicit fictitious domain (FD) method.24 The shear stress maxima τmax of 7.0 dyne/cm2 occurs upon entry to the STE where the high velocity ligand stream converges with the cell stream (see Fig. 3(A, B)). The mean normal stress maxima σmax of 10.3 dyne/cm2 coincides with the impact of the cell with the microchannel wall further downstream (see Fig. 3(C)). This results in the elastic deformation of the cell (see Fig. 3(D,E)). These values are below physiological shear stress values experienced during blood circulation (≥30 dyne/cm2)25 and importantly only experienced for microseconds. Both stresses rapidly decay to zero. We attribute this to the initiation of cell rolling caused by multiphase coupling, with zero cell stress during steady-state transport through the remainder of the STE.
A Roll–Slip Process Drives Micromixing In our microfluidic quenched flow analysis of EGFR activation,13 the switch time τswitch was defined as the transit time from ligand contact to submersion in the expansion channel at the exit of the STE. At 100 mm/s STE mean velocities the τswitch was 2.22 ms (SD ± 0.15 ms; supporting information, Figure S4). However, in this study a cell rolling phenomena was observed and is documented in the supporting information (Movies 2 and 3). This acts to envelop the cell for rapid cell stimulation and rapid reaction arrest. With no-slip (i.e. perfect rolling), zero diffusion time and zero interface perturbation assumptions the biochemical switch times τswitch can be approximated using a simple geometric model. In this case the
τswitch is cell-size-dependent, requiring only a fraction of a rotation (123°/360° ≈ 0.34) for a cell spanning the virtual interface to become fully coated with ligands. The method is described in the supporting information (Figure S5). For the operating conditions used in our 15 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
experiments the τswitch can conveniently be approximated as 2/3πrcell/u. For a HeLa cell (Ø = 15.8 µm SD ± 2.2; supporting information, Figure S2) travelling with a linear velocity of 100 mm/s the τswitch is estimated to be 167 µs (SD ± 15). To experimentally investigate the cell rolling phenomena, the HeLa cells were substituted with 15-µm-diameter, red-fluorescent Janus particles (produced by unidirectional gold sputtering). Fluorescent imaging requires lengthy exposure times, with our current experimental system limited to imaging particles transported with a mean STE flow velocity of 0.7 mm/s. Janus particle transport is provided as a movie in the supporting information (Movie 4) and shown as an image sequence in Figure 4(A)). A roll-slip behaviour was observed, with 3.1 rotations over a 200 µm transport length, whereas a fully rolling particle rotates 4.0 times over the same distance. The use of FD simulations to extend the analysis to relevant velocities (u =100 mm/s) indicates that there is a larger slip component, with only 1.8 rotations over the 200 µm distance. Here, a cell friction coefficient of 0.05 was used.26 The rolling τswitch model neglected the slip character and interference of the cell–ligand laminar boundary. To more accurately estimate cell–ligand mixing timescales an FD simulation involving the species transport equation was undertaken. Time course concentration maps are documented in Figure 4(B-D) and an animation is provided in the supporting material. Mixing is rapid, attaining 38% of the concentration of the ligand stream in 700 µs and 44% levels upon exit from the STE (2.2 ms, Figure 4(E)). From these investigations it is evident that micromixing dominates biochemical switching, with displacement playing a lesser role. The mixing characteristic also implies that ligand concentrations should be increased to ensure saturation of the cell surface. The species transport model is only illustrative of the cell–ligand mixing process. Importantly it neglects adsorption onto the cell surface and receptor occupancy. Nevertheless 16 ACS Paragon Plus Environment
Page 16 of 33
Page 17 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
the model provides a first approximation of the true τswitch and highlights the convection paths and the large slip component. We estimate that at the 100 mm/s flow velocity a slip-roll transport length of ~100 µm (~1 ms) is required for surface saturation. The previously published τswitch definition considered transport along the entire, 200 µm length of the STE (τstim. = 2.2 ms; τquench = 0.55 ms).13 These switch times are therefore an over estimation of the cell–biochemical mixing timescales. Our single millisecond switch times are faster than those of other cell handling microfluidic techniques,10c,11,12a and are faster than the 30 ms dead times documented for the cryofixation-based quench flow method.18 Indeed, the microfluidic mixing times are equivalent to the dead times of modern stopped- and quenched-flow analysis instruments. Our method also rivals the microfluidic molecular mixers. For example, microfluidic droplet systems have been used to investigate biochemical kinetics with mixing timescales of ~1.5 ms,6 while hydrodynamic focusing has been used to measure single molecule interactions with a temporal resolution of 0.2 ms.27 Critically, however, the tremendous value of our microfluidic approach is that it extends beyond the analysis of purified molecules in the aqueous phase to enable molecular investigations in the context of the entire cell. While patch-clamp type micromanipulation of cells can be used to translate cells between laminar streams with an exchange rate of milliseconds,28 this single cell at a time handling technique is ill-equipped for high throughput research.
17 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 18 of 33
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 4. Multiphase Coupling. The roll-slip cell transport character drives a ligand envelopment micromixing process. Image overlay sequence of a Janus particle rolling with slip through the STE (u = 0.7 mm/s). Convective-diffusive mixing at 2 µm (0.1 ms), 35 µm (1.0 ms) and 142 µm (2.0 ms) distances along the STE (B-D). In these concentration maps, red denotes 100% ligand and dark blue denotes 0% ligand concentrations. Species transport (analogous to the Langmuir isotherm) was used to estimate cell stimulation timescales (E).
Various measures can be undertaken to increase the rotational rate to attain faster mixing and shorter τswitch intervals: The slip component can be reduced by widening the STE such that the cell diameter is equal to the channel half width. Owing to the parabolic velocity profile, the cell surface furthest from the microchannel wall experiences the highest hydrodynamic drag, while the surface contacting the microchannel wall (the fulcrum point) experiences the lowest drag. This imparts the highest net torque to increase the rolling component. Other approaches include increasing the difference in the viscosity of the two 18 ACS Paragon Plus Environment
Page 19 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
streams, increasing the systems friction coefficient (changing device materials) and increasing the relative stream velocities (e.g. by increasing the flow ratio or reducing the microchannel width). In addition, the spin up rate can be increased by matching the fluid density with that of the cell (ρ = 1.05 g/cm3). Further experiments are required to identify the optimal conditions for maximal cell rotation and localized micromixing. In laminar flows mixing is diffusion limited, with the objective to introduce microscopic striations of the liquid layers. Our investigation of the whole cell biochemical switch time serves as a general lesson: Particulates in confined flow streams (with widths small than the cell diameter) provide localised convection conditions for rapid surface interactions. This is applicable to a multitude of particulate-borne chemistries that can all benefit from uniform and precisely defined chemically or thermally driven reaction steps. Applications range from catalysis and synthesis to our applications in cell biology. These examples highlight the widespread academic and industrial utility of the technology.
Incubation Precision In our microfluidic quenched flow system the incubation time is defined by the length of the incubation channel and the cell transport velocity. However, the cell size distribution and the parabolic velocity flow field results in dispersion leading to differing incubation times. HeLa S3 cells have a mean diameter of 15.8 µm (SD ± 2.2 µm; supporting information, Figure S2). As demonstrated in Figure 5(A) using 10- and 20-µm-particles, displacement is size dependent. Different sized cells also have different transport trajectories (although particle displacement is significantly smaller)29 and consequently experience different incubation times. Horizontal microparticle image velocimetry (µPIV) measurements demonstrate that the high aspect ratio incubation channel produces a flat-fronted Hele-Shaw velocity profile 19 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
(Figure 5(B)). This aids management of the cell size distribution. Increasing the incubation channel width from 250 µm to 500 µm reduced the absolute deviation from the mean velocity. However, the mean velocity is also reduced, producing similar velocity coefficients of variation (CV ≈ 10%). The velocity (u) profile of the vertical, z–x plane is parabolic. This is the major source of variation in the cell transport velocity (Figure 5(C)). To overcome this problem, cells can be focused at the channel mid-height to experience the same, maximum transport velocity. Amongst several actuation methods this can be achieved using dielectrophoretic9 or acoustic8 focusing. However, device fabrication and operation becomes complicated. An elegant strategy is the use of inertial focusing,11 although in this contribution we have opted to use hydrodynamic lift. A lift channel was moulded in the basal PDMS layer and aligned upstream of the entrance to the stimulation STE. This approach requires all cells to sediment to the channel base before flow convergence and lift to the STE mid-height. In practice this requires a ≥28-mm-long cell inlet channel and a 5:3 Qcell:Qlift flow ratio to elevate all cells by 22.5 µm. The method is discussed in detail in the supporting information. Using this simple approach the mean transport velocity was increased from 6.6 to 8.5 mm/s and most significantly the velocity CV was reduced from 12.8% to 3.6% (see Figure 5(C,D)). The direct scaling of variance with velocity ensures the temporal resolution is retained as the incubation timescales decrease.
20 ACS Paragon Plus Environment
Page 20 of 33
Page 21 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 5. Incubation Flows. Size-dependent trajectories; movie frame compilation of 10- and 20-µm-diameter polystyrene particles (A). Comparison of the velocity profiles obtained by microparticle image velocimetry (µPIV) at the mid-heights of the 250-µm and 500-µm-wide expansion channels (B). Measured lateral cell positions and velocities are highlighted within rings. The wider channel has a pronounced flat-fronted Hele-Shaw profile. Cell trajectory during deflection and vertical focusing using a lift flow (C). Sedimentation times and the position of 10, 15 and 20-µm-diameter cells following lift (D). Velocity (u) profile of the z–x plane obtained from µPIV z-axis scans at the lateral cell streaming position (E). Using a lift flow to focus cells at the mid-channel height (green) the velocity is increased and the velocity co-efficient of variation is decreased. Measured velocity distribution using a narrow expansion channel without lift (blue), a wide expansion channel with (green) and without (red) lift (F). 21 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Two-phase plug flows are not associated with dispersion, producing highly reproducible incubation times.6 However, as with conventional quenched flow instruments, the internal circulatory flows represent a high shear environment that may well damage the cell membrane. In addition, high capillary number flows require the use of surfactants to stabilise the plugs, which complicates droplet fusion or emulsion-breaking processes needed to combine the quench reagent with the cell-carrying aqueous plug.
Rapid IGF-IR State Switching Cellular communication via receptor tyrosine kinases (RTK) is a feature common to all metazoans. These membrane-spanning receptors transduce extracellular signals (e.g. growth factors) into the cell by reversible tyrosine autophosphorylation of the cytosolic domain. RTK functionality is fundamental to the onset of many signalling pathways. Aberrant signalling leads to a host of diseases, notably cancer and diabetes. These receptors have therefore received great attention as therapeutic targets.30 In our pilot research we have used microfluidic quenched flow for the temporal analysis of ligand-mediated epidermal growth factor receptor (EGFR) autophosphorylation transitions. A new biphasic pY1173 signal transduction was observed, with stable activation achieved by diffusion-driven dimerization over a timescale of 5 seconds. In this research, we have chosen to investigate the signal transduction kinetics of the insulin-like growth factor 1 RTK (IGF-IR). Unlike EGFR, this receptor is pre-dimerized such that autophosphorylation is not limited by diffusion-driven collision within the membrane. This enables rapid signal transduction to the kinase activation loop in the beta chains at the C-terminus.31 Shown in Figure 6, significant autophosphorylation of the pY1131 residue occurred within 100 ms and remained stable throughout the 2 s time course (p < 9.4 x 10-8). The phosphorylation levels are beneath those of the positive control. Consistent with the literature, maximal levels are 22 ACS Paragon Plus Environment
Page 22 of 33
Page 23 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
gradually attained over 5 mins.32 The quantitative image data captures the rapid stateswitching and also the natural variations and fluctuations in the system. By truncating the incubation channel lengths the initial 100 ms can be investigated in greater detail to fully elucidate the mechanisms underlying the sequential phosphorylation of the Y1131, 1135 and 1136 tyrosine residues. This is the origin of IGF-IR signal transduction and is necessary for stabilising the kinase activation loop which drives substrate catalysis for downstream signalling.33 Figure 6. Rapid Autophoshorylation State Switching. The onset of IGF-1R signal transduction by autophosphorylation (pY1131). Overlaid bright field and fluorescent microscopy images of single cells treated for precisely defined periods with IGF (A). Cells were treated for 100, 500 and 2,000 ms. The negative control sample was processed in the microfluidic quenched flow system in the absence of ligand. Positive controls were treated for 5 mins in a Petri dish. Mean intensity single cell data (B) and sub-population averaged data (C). Significant pY1131 state switching occurs within 100 ms (p < 9.4 × 10-8), with gradual rise to maximal levels following incubation for 5 mins.[28]
23 ACS Paragon Plus Environment
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis 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
CONCLUSIONS We have developed a microfluidic quenched flow platform for the investigation of cell surface events with unprecedented temporal resolution. Cell–wall–fluid coupling produces a roll-slip behaviour which enables gentle, non-destructive cell handling at high velocities and drives the millisecond-scale envelopment of the cell with ligand. Incubation dwell time variation caused by the natural cell size distribution was managed by the use of a horizontal Hele-Shaw flow profile and vertical hydrodynamic focusing. The temporal precision was demonstrated by recording IGF-1 receptor autophosphorylation state-switching. In summary, the microfluidic technology is a valuable addition to the cell biology toolkit: The temporal dimension of whole cell systems biology research can now be extended to millisecond timescales.
ASSOCIATED CONTENT Supporting Information: Movies showing high speed cell displacement (1), cells rolling (2 and 3), a rolling Janus particle (4) and cell mediated micromixing (5). Figures: Virtual interface (S1), cell size distribution (S2), flows below the Moffatt eddy regime (S3), cell transit time (S4), rotation mixing model (S5) and hydrodynamic focusing (S6). Detailed explanations of the sedimentation and lift cell focusing method and the fictitious domain simulation method. This material is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION Corresponding Author *E-mail: [email protected]
ACS Paragon Plus Environment
24
Page 24 of 33
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis
Page 25 of 33
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
ACKNOWLEDGMENTS The authors are grateful to Ulrich Marggraf for SU-8 fabrication, Norman Ahlman for constructing the temperature control system, Sarah Waide for cell culture support, Sven Müller and Jian Hou for laser confocal microscopy assistance (MPI, Molecular Physiology, Dortmund) and Maria Becker for SEM imaging. We thank Rosemary O’Connor (University College Cork) for useful IGF-1R discussions. The research was financed by the DFG (WE3737/3-1), the Taiwanese Ministry of Education and by the Ministerium für Innovation, Wissenschaft und Forschung des Landes Nordrhein-Westfalen.
REFERENCES 1.
(a) Yeo, L. Y.; Chang, H.-C.; Chan, P. P. Y.; Friend, J. R. Small 2011, 7 (1), 12–48; (b) El-Ali, J.; Sorger, P. K.; Jensen, K. F. Nature 2006, 442 (7101), 403–411.
2.
Fersht, A. R.; Jakes, R. Biochemistry 1975, 14 (15), 3350–3356.
3.
Gibson, Q. H.; Milnes, L. Biochem. J. 1964, 91, 11.
4.
Dengjel, J.; Akimov, V.; Olsen, J. V.; Bunkenborg, J.; Mann, M.; Blagoev, B.; Andersen, J. S. Nat. Biotechnol. 2007, 25 (5), 566–568.
5.
Stroock, A. D.; Dertinger, S. K. W.; Ajdari, A.; Mezic, I.; Stone, H. A.; Whitesides, G. M. Science 2002, 295 (5555), 647–651.
6.
Song, H.; Ismagilov, R. F. J. Am. Chem. Soc. 2003, 125 (47), 14613–14619.
7.
(a) Eriksson, E.; Enger, J.; Nordlander, B.; Erjavec, N.; Ramser, K.; Goksor, M.; Hohmann, S.; Nystrom, T.; Hanstorp, D. Lab Chip 2007, 7 (1), 71–76; (b) Chiou, P. Y.; Ohta, A. T.; Wu, M. C. Nature 2005, 436 (7049), 370–372.
8.
(a) Evander, M.; Lenshof, A.; Laurell, T.; Nilsson, J. Anal. Chem. 2008, 80 (13), 5178– 5185; (b) Gedge, M.; Hill, M. Lab Chip 2012, 12 (17), 2998–3007.
9.
(a) Fiedler, S.; Shirley, S. G.; Schnelle, T.; Fuhr, G. Anal. Chem. 1998, 70 (9), 1909– 1915; (b) Seger, U.; Gawad, S.; Johann, R.; Bertsch, A.; Renaud, P. Lab Chip 2004, 4 (2), 148–151; (c) Pethig, R. Biomicrofluidics 2010, 4 (2); 022811; (d) Holmes, D.; Morgan, H.; Green, N. G. Biosensors & Bioelectronics 2006, 21 (8), 1621–1630.
10.
(a) Huang, L. R.; Cox, E. C.; Austin, R. H.; Sturm, J. C. Science 2004, 304 (5673), 987–90; (b) Morton, K. J.; Loutherback, K.; Inglis, D. W.; Tsui, O. K.; Sturm, J. C.; Chou, S. Y.; Austin, R. H. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (21), 7434–7438; (c) Morton, K. J.; Loutherback, K.; Inglis, D. W.; Tsui, O. K.; Sturm, J. C.; Chou, S. Y.; Austin, R. H. Lab Chip 2008, 8 (9), 1448–53.
ACS Paragon Plus Environment
25
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis 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
11.
(a) Amini, H.; Sollier, E.; Weaver, W. M.; Di Carlo, D. Proc. Natl. Acad. Sci. U.S.A. 2012, 109 (29), 11593–11598; (b) Gossett, D. R.; Tse, H. T. K.; Dudani, J. S.; Goda, K.; Woods, T.A.; Graves, S. W.; Di Carlo, D. Small 2012, 8 (17), 2757–2764.
12.
(a) Toyama, K.; Yamada, M.; Seki, M. Biomed. Microdevices 2012, 14 (4), 751–757; (b) Yamada, M.; Kobayashi, J.; Yamato, M.; Seki, M.; Okano, T. Lab Chip 2008, 8 (5), 772–778; (c) Yamada, M.; Nakashima, M.; Seki, M. Anal. Chem. 2004, 76 (18), 5465– 5471; (d) Yamada, M.; Seki, M. Lab Chip 2005, 5 (11), 1233–1239.
13.
Chiang, Y. Y.; West, J. Lab Chip 2013, 13 (6), 1031–1034.
14.
Frimat, J. P.; Becker, M.; Chiang, Y. Y.; Marggraf, U.; Janasek, D.; Hengstler, J. G.; Franzke, J.; West, J. Lab chip 2011, 11 (2), 231–237.
15.
(a) Huang, N. P.; Michel, R.; Voros, J.; Textor, M.; Hofer, R.; Rossi, A.; Elbert, D. L.; Hubbell, J. A.; Spencer, N. D. Langmuir 2001, 17 (2), 489–498; (b) Hardelauf, H.; Sisnaiske, J.; Taghipour-Anvari, A. A.; Jacob, P.; Drabiniok, E.; Marggraf, U.; Frimat, J.-P.; Hengstler, J. G.; Neyer, A.; van Thriel, C.; West J., Lab Chip 2011, 11 (16), 2763–2771.
16.
Santiago, J. G.; Wereley, S. T.; Meinhart, C. D.; Beebe, D. J.; Adrian, R. J. Exp. Fluids 1998, 25 (4), 316–319.
17.
(a) Haeri, S.; Shrimpton, J. S. Int. J. Heat Mass Trans. 2013, 59, 219-229; (b) Haeri, S.; Shrimpton, J. S., J. Comput. Phys. 2013, 237, 21–45.
18.
(a) Knoll, G.; Braun, C.; Plattner, H. J. Cell Biol. 1991, 113 (6), 1295–1304; (b) Plattner, H.; Hentschel, J. In International Review of Cytology - a Survey of Cell Biology, Vol 255, Jeon, K. W. Ed. 2006; Vol. 255, pp 133.
19.
Moffatt, H. K. J. Fluid Mech. 1964, 18, 1–18.
20.
Durst, F.; Melling, A.; Whitelaw, J. H. J. Fluid Mech.1974, 64 (JUN3), 111–128.
21.
Fearn, R. M.; Mullin, T.; Cliffe, K. A. J. Fluid Mech. 1990, 211, 595–608.
22.
Sollier, E.; Go, D. E.; Che, J.; Gossett, D. R.; O’Bryne, S.; Weaver, W. M.; Kummer, N.; Rettig. M.; Goldman, J.; Nickols, N.; McCloskey, S.; Kulkarni, R. P.; Di Carlo, D. Lab Chip, doi: 10.1039/c3lc50689d.
23.
Beech, J. P.; Holm, S. H.; Adolfsson, K.; Tegenfeldt, J. O. Lab Chip 2012, 12 (6), 1048–1051.
24.
(a) Haeri, S.; Shrimpton, J. S. Prog. Energy Combust. Sci. 2011, 37 (6), 716-740; (b) Haeri, S.; Shrimpton, J. S. Int. J. Multiphas. Flow 2012, 40, 38–55.
25
(a) Davies, P. F. Physiol. Rev. 1995, 75 (3), 519–560; (b) Resnick, N.; Yahav, H.; Shay-Salit, A.; Shushy, M.; Schubert, S.; Zilberman, L. C. M.; Wofovitz, E. Prog. Biophys. Mol. Biol. 2003, 81 (3), 177–199.
26.
(a) Dean, D.; Hemmer, J.; Vertegel, A.; Laberge, M. Materials 2010, 3 (9), 4668–4680; (b) Angelini, T. E.; Dunn, A. C.; Urueña, J. M.; Dickrell, D. J.; Burris, D. L.; Sawyer, W. G. Faraday Discuss. 2012, 156, 31–39.
27.
Gambin, Y.; VanDelinder, V.; Ferreon, A. C. M.; Lemke, E. A.; Groisman, A.; Deniz, A. A. Nat. Methods 2011, 8 (3), 239–241.
28.
Fedorov, N.; Benson, L.; Graef, J. D.; Hyman, J.; Sollenberger, J.; Pettersson, F.; Lippiello, P. M.; Bencherif, M. J. Neurosci. Meth. 2012, 206 (1), 23–33.
ACS Paragon Plus Environment
26
Page 26 of 33
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis
Page 27 of 33
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
29.
Microparticle displacement is significantly less than that predicted by theory. This could be attributed to the large friction co-efficient (~0.5) of polystyrene on polymer surfaces. This is an order of magnitude larger than the cell-glass friction co-efficient, with higher friction acting in concert with the rounded corners to drag the particles from the original STE displacement position.
30.
Levitzki, A.; Gazit, A. Science 1995, 267 (5205), 1782–1788.
31.
Lopaczynski, W.; Terry, C.; Nissley, P. Biochem. Biophys. Res. Co. 2000, 279 (3), 955–960.
32.
Vasilcanu, D.; Girnita, A.; Girnita, L.; Vasilcanu, R.; Axelson, M.; Larsson, O. Oncogene 2004, 23 (47), 7854–7862.
33.
Favelyukis, S.; Till, J. H.; Hubbard, S. R.; Miller, W. T. Nat. Struct. Biol. 2001, 8 (12), 1058–1063.
Table of Contents Graphic
Table of Contents Synopsis: Microfluidics can be harnessed for the high speed quenched flow analysis of single cells. A liquid–cell–wall coupling phenomena produces a cell rolling characteristic which minimizes the mechanical stresses experienced by the cell and drives rapid ligand envelopment. This ultrafast microfluidic switching platform enables cell surface processes to be investigated with millisecond resolution.
ACS Paragon Plus Environment
27
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
Microfluidic Circuit. Whole cell quenched flow analysis system replicated in PDMS. Microchannel edges have been highlighted using black lines. A STE is used for cell stimulation, with a second used for reaction arrest at the end of the incubation channel. Inset, a flow ratio of 1:7 was used for single cell displacement into a dye-doped stream. The arrows denote the cell transport direction. 59x47mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 28 of 33
Page 29 of 33
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
Speed Limit 85x123mm (300 x 300 DPI)
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
117x195mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 30 of 33
Page 31 of 33
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
116x196mm (300 x 300 DPI)
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Incubation Flows. Size-dependent trajectories; movie frame compilation of 10- and 20-µm-diameter polystyrene particles (A). Comparison of the velocity profiles obtained by microparticle image velocimetry (µPIV) at the mid-heights of the 250-µm and 500-µm-wide expansion channels (B). Measured lateral cell positions and velocities are highlighted within rings. The wider channel has a pronounced flat-fronted HeleShaw profile. Cell trajectory during deflection and vertical focusing using a lift flow (C). Sedimentation times and the position of 10, 15 and 20-µm-diameter cells following lift (D). Velocity (u) profile of the z–x plane obtained from µPIV z-axis scans at the lateral cell streaming position (E). Using a lift flow to focus cells at the mid-channel height (green) the velocity is increased and the velocity co-efficient of variation is decreased. Measured velocity distribution using a narrow expansion channel without lift (blue), a wide expansion channel with (green) and without (red) lift (F). 109x145mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 32 of 33
Page 33 of 33
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
100x166mm (300 x 300 DPI)
ACS Paragon Plus Environment
Article
Whole cell quenched flow analysis Ya-Yu Chiang, Sina Haeri, Carsten Gizewski, Joanna D. Stewart, Peter Ehrhard, John Shrimpton, Dirk Janasek, and Jonathan J. West Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 31 Oct 2013 Downloaded from http://pubs.acs.org on November 4, 2013
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Whole cell quenched flow analysis Ya-Yu Chiang1,2, Sina Haeri3, Carsten Gizewski4, Joanna D. Stewart5, Peter Ehrhard4, John Shrimpton3, Dirk Janasek1 and Jonathan West1,2*
1
Leibniz-Institut für Analytische Wissenschaften - ISAS - e.V., 44227 Dortmund, Germany
2
Institute for Life Sciences, University of Southampton, SO17 1BJ, U.K.
3
Engineering and the Environment, University of Southampton, SO17 1BJ, U.K.
4
Fluid Mechanics, Biochemical and Chemical Engineering, Technische Universität, 44221
Dortmund, Germany
5
Leibniz Research Centre for Working Environment and Human Factors – IfADo, 44139
Dortmund, Germany
KEYWORDS: Microfluidics, deterministic lateral displacement, single cell analysis, signal transduction, multi-phase coupling, micromixing, autophosphorylation and insulin-like growth factor receptor.
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
ABSTRACT This paper describes a microfluidic quenched flow platform for the investigation of ligandmediated cell surface processes with unprecedented temporal resolution. A roll-slip behaviour caused by cell–wall–fluid coupling was documented and acts to minimize the compression and shear stresses experienced by the cell. This feature enables high velocity (100–400 mm/s) operation without impacting the integrity of the cell membrane. In addition, rotation generates localized convection paths. This cell-driven micromixing effect causes the cell to become rapidly enveloped with ligands to saturate the surface receptors. High speed imaging of the transport of a Janus particle and fictitious domain numerical simulations were used to predict millisecond-scale biochemical switching times. Dispersion in the incubation channel was characterised by microparticle image velocimetry and minimised by using a horizontal HeleShaw velocity profile in combination with vertical hydrodynamic focusing to achieve highly reproducible incubation times (CV = 3.6%). Microfluidic quenched flow was used to investigate the pY1311 autophosphorylation transition in the type I insulin-like growth factor receptor (IGF-1R). This pre-dimerized receptor undergoes autophosphorylation within 100 ms of stimulation. Beyond this demonstration, the extreme temporal resolution can be used to gain new insights into the mechanisms underpinning a tremendous variety of important cell surface events.
INTRODUCTION Extracellular biochemical stimuli are perceived by membrane hosted receptors which respond by initiating molecular signalling cascades. Transduction mechanisms include conformational and autocatalytic modifications, such as tyrosine autophosphorylation, and occur over millisecond timescales. However, these timescales are currently inaccessible owing to an 2 ACS Paragon Plus Environment
Page 2 of 33
Page 3 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
absence of techniques for the rapid stimulation of receptors in the context of the cell membrane (the optimal biophysical environment for receptor function). Consequently the sequences and mechanisms underlying signal transduction remain poorly characterised, thereby hindering the rational design of cell surface targeting therapeutics. New techniques are therefore required to extend the cell biology toolkit to make accessible the millisecond timescales of processes occurring at the cell surface. In recent years the spatial precision of microfluidics has begun to impact biology1 and promises to deliver the temporal precision necessary to investigate ultrafast processes in cell biology. Quenched flow technology can be used to analyse the millisecond-scale kinetics of biomolecular interactions.2 The method involves the use of high pressure pumps to drive the rapid, turbulent mixing of reactive (bio)molecular partners, followed by short incubation periods defined by the length of a delay line before rapid reaction arrest by turbulent mixing with a quenching buffer. Unlike stopped flow3 this bypasses the need for distinguishable spectral signatures that may not exist in macromolecular systems. Instead the full repertoire of analytical methods is available for off-line identification of the reactions intermediate states. The use of turbulent flows enables mixing in milliseconds for the identification of reaction intermediates and the elucidation of the underlying mechanisms. A wide variety of molecular processes are hosted within the complex microenvironment of the cell membrane. For example, ligand-mediated signal transduction by membrane-spanning receptors is critical for health, with dysfunction leading to disease. Such processes are fast, occurring over sub-second timescales. The mechanisms and dynamics of these processes should be studied in the native state of the membrane to provide an authentic biophysical environment and ideally using whole cells to accommodate the array of all possible interaction partners in both membrane and cytosolic compartments. However, the turbulent, high shear conditions necessary for rapid mixing in quenched flow analysis result in 3 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Ya-Yu Chiang et al Whole cell quenched flow analysis
the disruption of cell membranes. This prevents the meaningful investigation of membranehosted reactions and consequently the systems-level molecular dynamics of many cell surface events remain poorly resolved. To tackle this challenge a gentle method for near-instantaneous cell surface stimulation followed by near-instantaneous reaction arrest is required to preserve the reaction intermediates for post-treatment analysis. The micron-scale size of cells affords the opportunity to use microfluidic processing techniques. However, at microscopic scales flows are characteristically laminar, with mixing driven by diffusion (supporting information, Figure S1). The convergence of cell and ligand streams leads to the formation of a virtual interface with membrane-hosted ligand–receptor interactions delayed by diffusion timescales.4 Advection using either microstructured channel walls5 or circulation within plug flow reactors6 can achieve mixing in milliseconds. However, mixing is shear-driven, undesirable for processing intact cells. A conceptually more straightforward approach involves deflecting cells across the virtual interface between laminar flows and into a ligand-doped adjacent stream with the process repeated downstream for reaction arrest within an adjacent quenching buffer stream. Optical,7 acoustic8 and dielectric9 mobilisation options are available. However, only pN forces can be imparted as a result of unfavourable scaling at cell dimensions and weak material contrasts between the cell and the aqueous media. Cell-length translation for switching biochemical streams therefore requires several 10’s of milliseconds, inadequate to resolve most molecular processes. Instead, hydrodynamic methods such as deterministic lateral displacement (DLD) can be used. This approach scales favourably enabling the rapid deflection of cells into neighbouring streams.10 Here, angled micropillar arrays are used to compress laminar streams to sub-cellular dimensions for the sequential deflection of cells
4 ACS Paragon Plus Environment
Page 4 of 33
Page 5 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
from their original path. Alternatively, inertial focusing can be used for whole cell translation in milliseconds.11 In this paper we present a whole cell microfluidic quenched flow analysis system which involves a two-step DLD arrangement based on a pinched-flow approach12 using socalled stream-thinning elements (STEs).13 We previously developed a prototype and used this to discover a novel biphasic epidermal growth factor receptor (EGFR) signal transduction mechanism.13 In this contribution we extend beyond this demonstration, using experimental and numerical simulation methods to investigate the operational velocity limit and the temporal resolution of the activation and incubation reaction phases. A cell–wall–fluid coupling phenomena was observed which produces roll-slip cell transport for micromixing and ligand envelopment over millisecond time scales. The system was used to investigate the rapid autophosphorylation (pY1131) state-switching dynamics of type I insulin-like growth factor receptor (IGF-1R).
EXPERIMENTAL SECTION Design, Fabrication and Packaging The dimensions of the microfluidic quenched flow system are defined by the size of the cells. In this study, human epithelial carcinoma cells, HeLa S3, with a suspension diameter of 15.8 SD ± 2.2 µm were used. The stream thinning elements (STEs) used for the convergence of cell and ligand or quench buffer streams were 25-µm-wide and 200-µm-long. Twenty-fold channel expansion at 120° produced a 500-µm-wide incubation channel, with lengths corresponding to 100 ms to 2.00 s. Mirrored channel bifurcation was used to balance the system’s fluidic resistance and reduce cell dilution. Both paths contained a second STE for reaction arrest. A quenched flow device is shown in Figure 1. 5 ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 33
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 1. Microfluidic Circuit. Whole cell quenched flow analysis system replicated in PDMS. Microchannel edges have been highlighted using black lines. A STE is used for cell stimulation, with a second used for reaction arrest at the end of the incubation channel. Inset, a flow ratio of 1:7 was used for single cell displacement into a dye-doped stream. The arrows denote the cell transport direction.
The microfluidic quenched flow circuit was prepared by replica moulding in poly(dimethyl siloxane) (PDMS, Sylgard 184, Dow Corning) using a 60-µm-high SU-8 master fabricated by standard photolithographic methods (see Frimat et al).14 A biopsy puncture (Kia Medical) was used to produce 1-mm-diameter inlet and outlet ports. A thin (~300 µm) PDMS layer prepared on a glass coverslip was used to encapsulate the microchannels and support the system. For vertical focusing a lift channel, also 60 µm in depth, was replicated in this basal PDMS layer. A pressure tolerant PDMS–PDMS seal was achieved by plasma bonding (70 W, 40 kHz (Femto, Diener Eletronic, Germany)) in a 0.2 mbar oxygen atmosphere for 45 s. Immediately after plasma bonding, a 2 µL volume of polyL-lysine-g-polyethylene glycol (PLL-g-PEG (100 µg/mL), Surface Solutions, Switzerland) was applied to the inlet ports for capillary-driven filling and the electrostatic derivatization of the microchannel surfaces. The PEG moiety resists protein adsorption15 and prevents the narrow STEs being clogged with cells.
6 ACS Paragon Plus Environment
Page 7 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Operation and Flow Imaging The microfluidic system was connected to syringe pumps (e.g. neMESYS, Centoni GmbH) via PFA tubing (OD 1/16”). 1-cm-long steel tubing (18G ×1 1/2”) sections inserted into the inlets and outlets were used to enable plug and play operation. The syringes containing the cell, ligand and quenching reagents were maintained at 37°C using a custom thermoelectric Peltier control system (Newport Corporation, Irvine, USA) involving syringes jacketed by water-cooled copper heat sinks. A 1:7 cell to ligand or quench reagent volumetric flow ratio was used for experiments. Flow patterns were imaged using 1-µm-diameter Nile red fluorescent microspheres (Invitrogen). To simulate cell trajectories 10- and 20-µm-diameter polystyrene particles (Kisker Biotech GmbH & Co. KG, SD Qcs/Qtotal). Together with the parabolic flow field a 4.0-µm-wide cell stream is produced, resulting in the mean cell centre being displaced 3.9 µm (i.e. r - wcs) across the virtual interface into the ligand stream. Displacement is amplified 20-fold in the 500-µm-wide 10 ACS Paragon Plus Environment
Page 10 of 33
Page 11 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
incubation channel, with the displacement D of each cell across the virtual interface expressed as:
D = (r − wCS ) ×
wexp wSTE
where w denotes the width of the expansion channel (exp) or the stream thinning element (STE). In theory HeLa cells are displaced 78 µm across the interface. However, in practice the HeLa cells are displaced by 46.5 µm (SD ± 15.7). We attribute this to the loss of sharp microchannel corners during SU-8 fabrication. Nevertheless all cells are fully submerged in the ligand stream. Indeed, 10-fold amplification using a 250-µm-wide incubation channel also ensures displacement (26.1 µm SD ± 16.4) of the entirety of all cells into the ligand stream.
Cellular and Microfluidic Speed Limit Displacement is velocity-independent. To effectively capture cell surface events it is necessary to operate at velocities which do not disrupt the cell membrane and which produce predictable flows and thus predictable incubation times. The peak velocity occurs in the reaction arrest STE. Here, channel bifurcation and the combination of flows at a volumetric ratio of 1:7 produces a mean velocity 4-fold higher than the stimulation STE velocity. Mean reaction arrest STE velocities from 10 to 500 mm/s were evaluated for their effect on the integrity and viability of HeLa cells. Even at the extreme velocities >90% of cells remained intact (i.e. excluded trypan blue, Figure 2(A)). Cells cultured overnight were found to be viable and had normal morphology. In contrast, the turbulent mixing character (typical Re ≥10,000) of the conventional quenched flow analysis technique exerts destructive mechanical stresses on the cells, damaging mammalian cell membrane and lysing the cell. Exceptions are rugged cells such as Paramecium cells that are protected by a rigid yet elastic pellicle.18 11 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 2. Speed Limit. The impact of STE mean flow velocity on membrane integrity as measured using the trypan blue assay (A). Expansion from the STE to the incubation path causes sudden deceleration of the flow. These inertial effects manifest as Moffatt eddies above a STE velocity of 400 mm/s. Particle trajectory image revealing the vortex generated using a mean STE velocity of 500 mm/s (B).
Flows in the microfluidic circuit are characteristically laminar. However, 120° channel expansion to 10- and especially 20-fold widths causes rapid deceleration. At sizeable Re numbers vortex behaviour or so-called Moffatt eddies are produced.19 The experimental and numerical studies by Durst et al20 and Fearn et al21 provide a detailed explanation of vortex occurrence at suddenly expanding symmetrical channels at low Re numbers. In our quenched flow circuit Moffatt eddies are observed when using ≥500 mm/s flow velocities in the STEs (channel Re = 20; Figure 2(B)). Cells can become trapped for transient or lengthier periods within the circulating flow patterns. Although this feature can be exploited for cell collection,22 in the microfluidic quenched flow system this delays biochemical switching and results in undefined incubation times. A reaction arrest STE velocity of 400 mm/s (and associated 100 mm/s stimulation STE velocity) was therefore used for subsequent experiments to avoid the formation of Moffatt eddies (supporting information, Figure S3). 12 ACS Paragon Plus Environment
Page 12 of 33
Page 13 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
The flow and particle Reynolds numbers for the different regions of the circuit are also documented in the supporting information (Table S1).
Cell Deformation and Shear Stress The membrane integrity and cell viability data demonstrates that the cells are minimally affected by high velocity transport through the STEs. To obtain an understanding of the mechanical stresses the cell experiences during stream switching in the stimulation STE highspeed imaging was used in combination with numerical simulations. The velocity field simulated in Figure 3(A) causes the cell to impact the microchannel wall upon entry to the STE. The elastic nature of the cell translates the collisional forces into deformation followed by relaxation, restoring the cell’s spherical character (see Figure 3(B-D)).23 The shear stress oscillations correspond to the interactions of the different regions of the cell with the velocity field and the surface of the channel wall. Importantly, deformation is transient such that upon exiting the STE the cells are fully displaced (on average 3.9 µm) for complete submersion in the ligand stream. In contrast, classical deterministic lateral displacement involves multiple micropillar collisions, which enable deformation and size to be used as fractionation parameters.23
13 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 14 of 33
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 3. Cell Stresses. Forces experienced during passage through the STE with a mean flow velocity of 100 mm/s. The velocity field (A), shear stress (B), cell deformation (C, D; y axis (red), x axis (blue)) and mean normal stress (i.e. compression, E).
14 ACS Paragon Plus Environment
Page 15 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
The magnitude of the shear and mean normal stresses were estimated by numerical simulation using bespoke algorithms involving an implicit fictitious domain (FD) method.24 The shear stress maxima τmax of 7.0 dyne/cm2 occurs upon entry to the STE where the high velocity ligand stream converges with the cell stream (see Fig. 3(A, B)). The mean normal stress maxima σmax of 10.3 dyne/cm2 coincides with the impact of the cell with the microchannel wall further downstream (see Fig. 3(C)). This results in the elastic deformation of the cell (see Fig. 3(D,E)). These values are below physiological shear stress values experienced during blood circulation (≥30 dyne/cm2)25 and importantly only experienced for microseconds. Both stresses rapidly decay to zero. We attribute this to the initiation of cell rolling caused by multiphase coupling, with zero cell stress during steady-state transport through the remainder of the STE.
A Roll–Slip Process Drives Micromixing In our microfluidic quenched flow analysis of EGFR activation,13 the switch time τswitch was defined as the transit time from ligand contact to submersion in the expansion channel at the exit of the STE. At 100 mm/s STE mean velocities the τswitch was 2.22 ms (SD ± 0.15 ms; supporting information, Figure S4). However, in this study a cell rolling phenomena was observed and is documented in the supporting information (Movies 2 and 3). This acts to envelop the cell for rapid cell stimulation and rapid reaction arrest. With no-slip (i.e. perfect rolling), zero diffusion time and zero interface perturbation assumptions the biochemical switch times τswitch can be approximated using a simple geometric model. In this case the
τswitch is cell-size-dependent, requiring only a fraction of a rotation (123°/360° ≈ 0.34) for a cell spanning the virtual interface to become fully coated with ligands. The method is described in the supporting information (Figure S5). For the operating conditions used in our 15 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
experiments the τswitch can conveniently be approximated as 2/3πrcell/u. For a HeLa cell (Ø = 15.8 µm SD ± 2.2; supporting information, Figure S2) travelling with a linear velocity of 100 mm/s the τswitch is estimated to be 167 µs (SD ± 15). To experimentally investigate the cell rolling phenomena, the HeLa cells were substituted with 15-µm-diameter, red-fluorescent Janus particles (produced by unidirectional gold sputtering). Fluorescent imaging requires lengthy exposure times, with our current experimental system limited to imaging particles transported with a mean STE flow velocity of 0.7 mm/s. Janus particle transport is provided as a movie in the supporting information (Movie 4) and shown as an image sequence in Figure 4(A)). A roll-slip behaviour was observed, with 3.1 rotations over a 200 µm transport length, whereas a fully rolling particle rotates 4.0 times over the same distance. The use of FD simulations to extend the analysis to relevant velocities (u =100 mm/s) indicates that there is a larger slip component, with only 1.8 rotations over the 200 µm distance. Here, a cell friction coefficient of 0.05 was used.26 The rolling τswitch model neglected the slip character and interference of the cell–ligand laminar boundary. To more accurately estimate cell–ligand mixing timescales an FD simulation involving the species transport equation was undertaken. Time course concentration maps are documented in Figure 4(B-D) and an animation is provided in the supporting material. Mixing is rapid, attaining 38% of the concentration of the ligand stream in 700 µs and 44% levels upon exit from the STE (2.2 ms, Figure 4(E)). From these investigations it is evident that micromixing dominates biochemical switching, with displacement playing a lesser role. The mixing characteristic also implies that ligand concentrations should be increased to ensure saturation of the cell surface. The species transport model is only illustrative of the cell–ligand mixing process. Importantly it neglects adsorption onto the cell surface and receptor occupancy. Nevertheless 16 ACS Paragon Plus Environment
Page 16 of 33
Page 17 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
the model provides a first approximation of the true τswitch and highlights the convection paths and the large slip component. We estimate that at the 100 mm/s flow velocity a slip-roll transport length of ~100 µm (~1 ms) is required for surface saturation. The previously published τswitch definition considered transport along the entire, 200 µm length of the STE (τstim. = 2.2 ms; τquench = 0.55 ms).13 These switch times are therefore an over estimation of the cell–biochemical mixing timescales. Our single millisecond switch times are faster than those of other cell handling microfluidic techniques,10c,11,12a and are faster than the 30 ms dead times documented for the cryofixation-based quench flow method.18 Indeed, the microfluidic mixing times are equivalent to the dead times of modern stopped- and quenched-flow analysis instruments. Our method also rivals the microfluidic molecular mixers. For example, microfluidic droplet systems have been used to investigate biochemical kinetics with mixing timescales of ~1.5 ms,6 while hydrodynamic focusing has been used to measure single molecule interactions with a temporal resolution of 0.2 ms.27 Critically, however, the tremendous value of our microfluidic approach is that it extends beyond the analysis of purified molecules in the aqueous phase to enable molecular investigations in the context of the entire cell. While patch-clamp type micromanipulation of cells can be used to translate cells between laminar streams with an exchange rate of milliseconds,28 this single cell at a time handling technique is ill-equipped for high throughput research.
17 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 18 of 33
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 4. Multiphase Coupling. The roll-slip cell transport character drives a ligand envelopment micromixing process. Image overlay sequence of a Janus particle rolling with slip through the STE (u = 0.7 mm/s). Convective-diffusive mixing at 2 µm (0.1 ms), 35 µm (1.0 ms) and 142 µm (2.0 ms) distances along the STE (B-D). In these concentration maps, red denotes 100% ligand and dark blue denotes 0% ligand concentrations. Species transport (analogous to the Langmuir isotherm) was used to estimate cell stimulation timescales (E).
Various measures can be undertaken to increase the rotational rate to attain faster mixing and shorter τswitch intervals: The slip component can be reduced by widening the STE such that the cell diameter is equal to the channel half width. Owing to the parabolic velocity profile, the cell surface furthest from the microchannel wall experiences the highest hydrodynamic drag, while the surface contacting the microchannel wall (the fulcrum point) experiences the lowest drag. This imparts the highest net torque to increase the rolling component. Other approaches include increasing the difference in the viscosity of the two 18 ACS Paragon Plus Environment
Page 19 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
streams, increasing the systems friction coefficient (changing device materials) and increasing the relative stream velocities (e.g. by increasing the flow ratio or reducing the microchannel width). In addition, the spin up rate can be increased by matching the fluid density with that of the cell (ρ = 1.05 g/cm3). Further experiments are required to identify the optimal conditions for maximal cell rotation and localized micromixing. In laminar flows mixing is diffusion limited, with the objective to introduce microscopic striations of the liquid layers. Our investigation of the whole cell biochemical switch time serves as a general lesson: Particulates in confined flow streams (with widths small than the cell diameter) provide localised convection conditions for rapid surface interactions. This is applicable to a multitude of particulate-borne chemistries that can all benefit from uniform and precisely defined chemically or thermally driven reaction steps. Applications range from catalysis and synthesis to our applications in cell biology. These examples highlight the widespread academic and industrial utility of the technology.
Incubation Precision In our microfluidic quenched flow system the incubation time is defined by the length of the incubation channel and the cell transport velocity. However, the cell size distribution and the parabolic velocity flow field results in dispersion leading to differing incubation times. HeLa S3 cells have a mean diameter of 15.8 µm (SD ± 2.2 µm; supporting information, Figure S2). As demonstrated in Figure 5(A) using 10- and 20-µm-particles, displacement is size dependent. Different sized cells also have different transport trajectories (although particle displacement is significantly smaller)29 and consequently experience different incubation times. Horizontal microparticle image velocimetry (µPIV) measurements demonstrate that the high aspect ratio incubation channel produces a flat-fronted Hele-Shaw velocity profile 19 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
(Figure 5(B)). This aids management of the cell size distribution. Increasing the incubation channel width from 250 µm to 500 µm reduced the absolute deviation from the mean velocity. However, the mean velocity is also reduced, producing similar velocity coefficients of variation (CV ≈ 10%). The velocity (u) profile of the vertical, z–x plane is parabolic. This is the major source of variation in the cell transport velocity (Figure 5(C)). To overcome this problem, cells can be focused at the channel mid-height to experience the same, maximum transport velocity. Amongst several actuation methods this can be achieved using dielectrophoretic9 or acoustic8 focusing. However, device fabrication and operation becomes complicated. An elegant strategy is the use of inertial focusing,11 although in this contribution we have opted to use hydrodynamic lift. A lift channel was moulded in the basal PDMS layer and aligned upstream of the entrance to the stimulation STE. This approach requires all cells to sediment to the channel base before flow convergence and lift to the STE mid-height. In practice this requires a ≥28-mm-long cell inlet channel and a 5:3 Qcell:Qlift flow ratio to elevate all cells by 22.5 µm. The method is discussed in detail in the supporting information. Using this simple approach the mean transport velocity was increased from 6.6 to 8.5 mm/s and most significantly the velocity CV was reduced from 12.8% to 3.6% (see Figure 5(C,D)). The direct scaling of variance with velocity ensures the temporal resolution is retained as the incubation timescales decrease.
20 ACS Paragon Plus Environment
Page 20 of 33
Page 21 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Figure 5. Incubation Flows. Size-dependent trajectories; movie frame compilation of 10- and 20-µm-diameter polystyrene particles (A). Comparison of the velocity profiles obtained by microparticle image velocimetry (µPIV) at the mid-heights of the 250-µm and 500-µm-wide expansion channels (B). Measured lateral cell positions and velocities are highlighted within rings. The wider channel has a pronounced flat-fronted Hele-Shaw profile. Cell trajectory during deflection and vertical focusing using a lift flow (C). Sedimentation times and the position of 10, 15 and 20-µm-diameter cells following lift (D). Velocity (u) profile of the z–x plane obtained from µPIV z-axis scans at the lateral cell streaming position (E). Using a lift flow to focus cells at the mid-channel height (green) the velocity is increased and the velocity co-efficient of variation is decreased. Measured velocity distribution using a narrow expansion channel without lift (blue), a wide expansion channel with (green) and without (red) lift (F). 21 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
Ya-Yu Chiang et al Whole cell quenched flow analysis
Two-phase plug flows are not associated with dispersion, producing highly reproducible incubation times.6 However, as with conventional quenched flow instruments, the internal circulatory flows represent a high shear environment that may well damage the cell membrane. In addition, high capillary number flows require the use of surfactants to stabilise the plugs, which complicates droplet fusion or emulsion-breaking processes needed to combine the quench reagent with the cell-carrying aqueous plug.
Rapid IGF-IR State Switching Cellular communication via receptor tyrosine kinases (RTK) is a feature common to all metazoans. These membrane-spanning receptors transduce extracellular signals (e.g. growth factors) into the cell by reversible tyrosine autophosphorylation of the cytosolic domain. RTK functionality is fundamental to the onset of many signalling pathways. Aberrant signalling leads to a host of diseases, notably cancer and diabetes. These receptors have therefore received great attention as therapeutic targets.30 In our pilot research we have used microfluidic quenched flow for the temporal analysis of ligand-mediated epidermal growth factor receptor (EGFR) autophosphorylation transitions. A new biphasic pY1173 signal transduction was observed, with stable activation achieved by diffusion-driven dimerization over a timescale of 5 seconds. In this research, we have chosen to investigate the signal transduction kinetics of the insulin-like growth factor 1 RTK (IGF-IR). Unlike EGFR, this receptor is pre-dimerized such that autophosphorylation is not limited by diffusion-driven collision within the membrane. This enables rapid signal transduction to the kinase activation loop in the beta chains at the C-terminus.31 Shown in Figure 6, significant autophosphorylation of the pY1131 residue occurred within 100 ms and remained stable throughout the 2 s time course (p < 9.4 x 10-8). The phosphorylation levels are beneath those of the positive control. Consistent with the literature, maximal levels are 22 ACS Paragon Plus Environment
Page 22 of 33
Page 23 of 33
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
Ya-Yu Chiang et al Whole cell quenched flow analysis
gradually attained over 5 mins.32 The quantitative image data captures the rapid stateswitching and also the natural variations and fluctuations in the system. By truncating the incubation channel lengths the initial 100 ms can be investigated in greater detail to fully elucidate the mechanisms underlying the sequential phosphorylation of the Y1131, 1135 and 1136 tyrosine residues. This is the origin of IGF-IR signal transduction and is necessary for stabilising the kinase activation loop which drives substrate catalysis for downstream signalling.33 Figure 6. Rapid Autophoshorylation State Switching. The onset of IGF-1R signal transduction by autophosphorylation (pY1131). Overlaid bright field and fluorescent microscopy images of single cells treated for precisely defined periods with IGF (A). Cells were treated for 100, 500 and 2,000 ms. The negative control sample was processed in the microfluidic quenched flow system in the absence of ligand. Positive controls were treated for 5 mins in a Petri dish. Mean intensity single cell data (B) and sub-population averaged data (C). Significant pY1131 state switching occurs within 100 ms (p < 9.4 × 10-8), with gradual rise to maximal levels following incubation for 5 mins.[28]
23 ACS Paragon Plus Environment
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis 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
CONCLUSIONS We have developed a microfluidic quenched flow platform for the investigation of cell surface events with unprecedented temporal resolution. Cell–wall–fluid coupling produces a roll-slip behaviour which enables gentle, non-destructive cell handling at high velocities and drives the millisecond-scale envelopment of the cell with ligand. Incubation dwell time variation caused by the natural cell size distribution was managed by the use of a horizontal Hele-Shaw flow profile and vertical hydrodynamic focusing. The temporal precision was demonstrated by recording IGF-1 receptor autophosphorylation state-switching. In summary, the microfluidic technology is a valuable addition to the cell biology toolkit: The temporal dimension of whole cell systems biology research can now be extended to millisecond timescales.
ASSOCIATED CONTENT Supporting Information: Movies showing high speed cell displacement (1), cells rolling (2 and 3), a rolling Janus particle (4) and cell mediated micromixing (5). Figures: Virtual interface (S1), cell size distribution (S2), flows below the Moffatt eddy regime (S3), cell transit time (S4), rotation mixing model (S5) and hydrodynamic focusing (S6). Detailed explanations of the sedimentation and lift cell focusing method and the fictitious domain simulation method. This material is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION Corresponding Author *E-mail: [email protected]
ACS Paragon Plus Environment
24
Page 24 of 33
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis
Page 25 of 33
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
ACKNOWLEDGMENTS The authors are grateful to Ulrich Marggraf for SU-8 fabrication, Norman Ahlman for constructing the temperature control system, Sarah Waide for cell culture support, Sven Müller and Jian Hou for laser confocal microscopy assistance (MPI, Molecular Physiology, Dortmund) and Maria Becker for SEM imaging. We thank Rosemary O’Connor (University College Cork) for useful IGF-1R discussions. The research was financed by the DFG (WE3737/3-1), the Taiwanese Ministry of Education and by the Ministerium für Innovation, Wissenschaft und Forschung des Landes Nordrhein-Westfalen.
REFERENCES 1.
(a) Yeo, L. Y.; Chang, H.-C.; Chan, P. P. Y.; Friend, J. R. Small 2011, 7 (1), 12–48; (b) El-Ali, J.; Sorger, P. K.; Jensen, K. F. Nature 2006, 442 (7101), 403–411.
2.
Fersht, A. R.; Jakes, R. Biochemistry 1975, 14 (15), 3350–3356.
3.
Gibson, Q. H.; Milnes, L. Biochem. J. 1964, 91, 11.
4.
Dengjel, J.; Akimov, V.; Olsen, J. V.; Bunkenborg, J.; Mann, M.; Blagoev, B.; Andersen, J. S. Nat. Biotechnol. 2007, 25 (5), 566–568.
5.
Stroock, A. D.; Dertinger, S. K. W.; Ajdari, A.; Mezic, I.; Stone, H. A.; Whitesides, G. M. Science 2002, 295 (5555), 647–651.
6.
Song, H.; Ismagilov, R. F. J. Am. Chem. Soc. 2003, 125 (47), 14613–14619.
7.
(a) Eriksson, E.; Enger, J.; Nordlander, B.; Erjavec, N.; Ramser, K.; Goksor, M.; Hohmann, S.; Nystrom, T.; Hanstorp, D. Lab Chip 2007, 7 (1), 71–76; (b) Chiou, P. Y.; Ohta, A. T.; Wu, M. C. Nature 2005, 436 (7049), 370–372.
8.
(a) Evander, M.; Lenshof, A.; Laurell, T.; Nilsson, J. Anal. Chem. 2008, 80 (13), 5178– 5185; (b) Gedge, M.; Hill, M. Lab Chip 2012, 12 (17), 2998–3007.
9.
(a) Fiedler, S.; Shirley, S. G.; Schnelle, T.; Fuhr, G. Anal. Chem. 1998, 70 (9), 1909– 1915; (b) Seger, U.; Gawad, S.; Johann, R.; Bertsch, A.; Renaud, P. Lab Chip 2004, 4 (2), 148–151; (c) Pethig, R. Biomicrofluidics 2010, 4 (2); 022811; (d) Holmes, D.; Morgan, H.; Green, N. G. Biosensors & Bioelectronics 2006, 21 (8), 1621–1630.
10.
(a) Huang, L. R.; Cox, E. C.; Austin, R. H.; Sturm, J. C. Science 2004, 304 (5673), 987–90; (b) Morton, K. J.; Loutherback, K.; Inglis, D. W.; Tsui, O. K.; Sturm, J. C.; Chou, S. Y.; Austin, R. H. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (21), 7434–7438; (c) Morton, K. J.; Loutherback, K.; Inglis, D. W.; Tsui, O. K.; Sturm, J. C.; Chou, S. Y.; Austin, R. H. Lab Chip 2008, 8 (9), 1448–53.
ACS Paragon Plus Environment
25
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis 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
11.
(a) Amini, H.; Sollier, E.; Weaver, W. M.; Di Carlo, D. Proc. Natl. Acad. Sci. U.S.A. 2012, 109 (29), 11593–11598; (b) Gossett, D. R.; Tse, H. T. K.; Dudani, J. S.; Goda, K.; Woods, T.A.; Graves, S. W.; Di Carlo, D. Small 2012, 8 (17), 2757–2764.
12.
(a) Toyama, K.; Yamada, M.; Seki, M. Biomed. Microdevices 2012, 14 (4), 751–757; (b) Yamada, M.; Kobayashi, J.; Yamato, M.; Seki, M.; Okano, T. Lab Chip 2008, 8 (5), 772–778; (c) Yamada, M.; Nakashima, M.; Seki, M. Anal. Chem. 2004, 76 (18), 5465– 5471; (d) Yamada, M.; Seki, M. Lab Chip 2005, 5 (11), 1233–1239.
13.
Chiang, Y. Y.; West, J. Lab Chip 2013, 13 (6), 1031–1034.
14.
Frimat, J. P.; Becker, M.; Chiang, Y. Y.; Marggraf, U.; Janasek, D.; Hengstler, J. G.; Franzke, J.; West, J. Lab chip 2011, 11 (2), 231–237.
15.
(a) Huang, N. P.; Michel, R.; Voros, J.; Textor, M.; Hofer, R.; Rossi, A.; Elbert, D. L.; Hubbell, J. A.; Spencer, N. D. Langmuir 2001, 17 (2), 489–498; (b) Hardelauf, H.; Sisnaiske, J.; Taghipour-Anvari, A. A.; Jacob, P.; Drabiniok, E.; Marggraf, U.; Frimat, J.-P.; Hengstler, J. G.; Neyer, A.; van Thriel, C.; West J., Lab Chip 2011, 11 (16), 2763–2771.
16.
Santiago, J. G.; Wereley, S. T.; Meinhart, C. D.; Beebe, D. J.; Adrian, R. J. Exp. Fluids 1998, 25 (4), 316–319.
17.
(a) Haeri, S.; Shrimpton, J. S. Int. J. Heat Mass Trans. 2013, 59, 219-229; (b) Haeri, S.; Shrimpton, J. S., J. Comput. Phys. 2013, 237, 21–45.
18.
(a) Knoll, G.; Braun, C.; Plattner, H. J. Cell Biol. 1991, 113 (6), 1295–1304; (b) Plattner, H.; Hentschel, J. In International Review of Cytology - a Survey of Cell Biology, Vol 255, Jeon, K. W. Ed. 2006; Vol. 255, pp 133.
19.
Moffatt, H. K. J. Fluid Mech. 1964, 18, 1–18.
20.
Durst, F.; Melling, A.; Whitelaw, J. H. J. Fluid Mech.1974, 64 (JUN3), 111–128.
21.
Fearn, R. M.; Mullin, T.; Cliffe, K. A. J. Fluid Mech. 1990, 211, 595–608.
22.
Sollier, E.; Go, D. E.; Che, J.; Gossett, D. R.; O’Bryne, S.; Weaver, W. M.; Kummer, N.; Rettig. M.; Goldman, J.; Nickols, N.; McCloskey, S.; Kulkarni, R. P.; Di Carlo, D. Lab Chip, doi: 10.1039/c3lc50689d.
23.
Beech, J. P.; Holm, S. H.; Adolfsson, K.; Tegenfeldt, J. O. Lab Chip 2012, 12 (6), 1048–1051.
24.
(a) Haeri, S.; Shrimpton, J. S. Prog. Energy Combust. Sci. 2011, 37 (6), 716-740; (b) Haeri, S.; Shrimpton, J. S. Int. J. Multiphas. Flow 2012, 40, 38–55.
25
(a) Davies, P. F. Physiol. Rev. 1995, 75 (3), 519–560; (b) Resnick, N.; Yahav, H.; Shay-Salit, A.; Shushy, M.; Schubert, S.; Zilberman, L. C. M.; Wofovitz, E. Prog. Biophys. Mol. Biol. 2003, 81 (3), 177–199.
26.
(a) Dean, D.; Hemmer, J.; Vertegel, A.; Laberge, M. Materials 2010, 3 (9), 4668–4680; (b) Angelini, T. E.; Dunn, A. C.; Urueña, J. M.; Dickrell, D. J.; Burris, D. L.; Sawyer, W. G. Faraday Discuss. 2012, 156, 31–39.
27.
Gambin, Y.; VanDelinder, V.; Ferreon, A. C. M.; Lemke, E. A.; Groisman, A.; Deniz, A. A. Nat. Methods 2011, 8 (3), 239–241.
28.
Fedorov, N.; Benson, L.; Graef, J. D.; Hyman, J.; Sollenberger, J.; Pettersson, F.; Lippiello, P. M.; Bencherif, M. J. Neurosci. Meth. 2012, 206 (1), 23–33.
ACS Paragon Plus Environment
26
Page 26 of 33
Ya-Yu Chiang et al Analytical Whole cell quenchedChemistry flow analysis
Page 27 of 33
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
29.
Microparticle displacement is significantly less than that predicted by theory. This could be attributed to the large friction co-efficient (~0.5) of polystyrene on polymer surfaces. This is an order of magnitude larger than the cell-glass friction co-efficient, with higher friction acting in concert with the rounded corners to drag the particles from the original STE displacement position.
30.
Levitzki, A.; Gazit, A. Science 1995, 267 (5205), 1782–1788.
31.
Lopaczynski, W.; Terry, C.; Nissley, P. Biochem. Biophys. Res. Co. 2000, 279 (3), 955–960.
32.
Vasilcanu, D.; Girnita, A.; Girnita, L.; Vasilcanu, R.; Axelson, M.; Larsson, O. Oncogene 2004, 23 (47), 7854–7862.
33.
Favelyukis, S.; Till, J. H.; Hubbard, S. R.; Miller, W. T. Nat. Struct. Biol. 2001, 8 (12), 1058–1063.
Table of Contents Graphic
Table of Contents Synopsis: Microfluidics can be harnessed for the high speed quenched flow analysis of single cells. A liquid–cell–wall coupling phenomena produces a cell rolling characteristic which minimizes the mechanical stresses experienced by the cell and drives rapid ligand envelopment. This ultrafast microfluidic switching platform enables cell surface processes to be investigated with millisecond resolution.
ACS Paragon Plus Environment
27
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
Microfluidic Circuit. Whole cell quenched flow analysis system replicated in PDMS. Microchannel edges have been highlighted using black lines. A STE is used for cell stimulation, with a second used for reaction arrest at the end of the incubation channel. Inset, a flow ratio of 1:7 was used for single cell displacement into a dye-doped stream. The arrows denote the cell transport direction. 59x47mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 28 of 33
Page 29 of 33
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
Speed Limit 85x123mm (300 x 300 DPI)
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
117x195mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 30 of 33
Page 31 of 33
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
116x196mm (300 x 300 DPI)
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Incubation Flows. Size-dependent trajectories; movie frame compilation of 10- and 20-µm-diameter polystyrene particles (A). Comparison of the velocity profiles obtained by microparticle image velocimetry (µPIV) at the mid-heights of the 250-µm and 500-µm-wide expansion channels (B). Measured lateral cell positions and velocities are highlighted within rings. The wider channel has a pronounced flat-fronted HeleShaw profile. Cell trajectory during deflection and vertical focusing using a lift flow (C). Sedimentation times and the position of 10, 15 and 20-µm-diameter cells following lift (D). Velocity (u) profile of the z–x plane obtained from µPIV z-axis scans at the lateral cell streaming position (E). Using a lift flow to focus cells at the mid-channel height (green) the velocity is increased and the velocity co-efficient of variation is decreased. Measured velocity distribution using a narrow expansion channel without lift (blue), a wide expansion channel with (green) and without (red) lift (F). 109x145mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 32 of 33
Page 33 of 33
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
100x166mm (300 x 300 DPI)
ACS Paragon Plus Environment
