Volume 15 Number 5 7 March 2015 Pages 12171396

Lab on aChip

Miniaturisation for chemistry, physics, biology, materials science and bioengineering

www.rsc.org/loc

ISSN 1473-0197

PAPER
Matthias Mehling, Tino Frank et al.
Real-time tracking, retrieval and gene expression analysis of migrating
human T cells

Lab on a Chip
View Article Online

PAPER

View Journal | View Issue

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Real-time tracking, retrieval and gene expression

analysis of migrating human T cells
Cite this: Lab Chip, 2015, 15, 1276

Matthias Mehling, Tino Frank, Cem Albayrak and Sava Tay

Dynamical analysis of single-cells allows assessment of the extent and role of cell-to-cell variability,
however traditional dish-and-pipette techniques have hindered single-cell analysis in quantitative biology.
We developed an automated microfluidic cell culture system that generates stable diffusion-based chemokine gradients, where cells can be placed in predetermined positions, monitored via single-cell time-lapse
microscopy, and subsequently be retrieved based on their migration speed and directionality for further
Received 3rd September 2014,
Accepted 4th December 2014
DOI: 10.1039/c4lc01038h

off-chip gene expression analysis, constituting a powerful platform for multiparameter quantitative studies
of single-cell chemotaxis. Using this system we studied CXCL12-directed migration of individual human
primary T cells. Spatiotemporally deterministic retrieval of T cell subsets in relation to their migration
speed, and subsequent analysis with microfluidic droplet digital-PCR showed that the expression level of

www.rsc.org/loc

CXCR4 the receptor of CXCL12 underlies enhanced human T cell chemotaxis.

Introduction
Cell migration has important roles in various physiological
processes such as embryogenesis, tissue repair, and especially
in immune responses.1,2 For protective immunity, migration
of T cells provides the basis for orchestrated homing and
positioning within lymphoid and non-lymphoid tissues.3
Tissue-specific homing and intra-parenchymal migration of T
cells is a highly regulated process at various temporal and
spatial scales.4 Specifically, exposure to chemokine gradients
and binding of chemokines to G-protein coupled receptors
induces polarization of T cells, and the formation of protrusions where focal adhesions link extracellular matrix proteins
to the actin-cytoskeleton result in directed migration towards
the gradients.5 Besides protective immunity, T cell migration
is also a key element in the pathogenesis of autoimmune
diseases such as multiple sclerosis6 or Crohn's disease.7 The
majority of the above-described insights in cell migration are
based on findings in animal models.
For human T cells, some of the migration-characteristics
have been recapitulated in vitro, mostly by the use of transwell assays. Transwell assays such as the Boyden-chamber
are robust, allow enumeration of the displacement of individual cells across a membrane and therefore provide a
quantitative measure of chemotaxis.8 However, this approach

Department of Biosystems Science and Engineering, ETH Zrich, Mattenstrasse

26, 4058 Basel, Switzerland. E-mail: savas.tay@bsse.ethz.ch
Electronic supplementary information (ESI) available: 5 videos, 1 image,
1 supplementary methods file. See DOI: 10.1039/c4lc01038h
M. Mehling and T. Frank contributed equally.

1276 | Lab Chip, 2015, 15, 12761283

is unapt to define key aspects impacting the biology of

cellular motility in vivo. Specifically, (i) no information can
be derived regarding the spatial and temporal stability of
chemotactic gradients, (ii) no complex multidirectional
gradients of multiple chemoattractants can be established
(which will typically be present in an in vivo system), and
most importantly (iii) single cells cannot be monitored and
characterized in real-time phenotypically or functionally. As a
result, most in vitro but also in vivo studies assessed various
aspects of T cell migration with a population-averaged
manner where migration characteristics at the single cell
level were not interrogated.
To characterize migration of human T cells and to aid
quantitative studies of chemotaxis, we developed an automated microfluidic cell culture system that significantly
surpasses the capabilities of traditional cell culture and
migration assays, and characterized migration of primary
human CD4+ T cells in gradients of the chemokine CXCL12
using this system. Recently, microfluidic single-cell analysis
was used to greatly improve our understanding of immune
functions from single cells up to the population level.913 Our
system was designed to address limitations of both traditional and existing microfluidic approaches, representing a
major advance in single-cell analysis of cell migration. The
microfluidic migration device we developed comprises 6
independent cell culture chambers, where in each chamber a
diffusion-based chemokine gradient can be generated and
both adherent and suspension cells can be cultured under
flow-free conditions. Our device controls the type, steepness,
mean concentration and polarity of each gradient generated
in the 6 independent chambers of the device. These gradients

This journal is The Royal Society of Chemistry 2015

View Article Online

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Lab on a Chip

are extremely stable with minimal variation of concentration

profiles over time, but the gradient type can also be changed
when desired without disturbing the cultured cells. By
integrating computer control, microfluidic membrane valves
and automated microscopy, our system allows precise positioning, monitoring and subsequent retrieval of migrating
cells from up to 10 locations inside each nanoliter-sized
culture chamber. Real-time quantification of migration via
time-lapse microscopy, automated tracking, and subsequent
retrieval of cell subpopulations at precisely determined positions allows differential genetic analysis of migrating vs. nonmigrating or slow vs. fast cells in the observed population.
Using this system we generated flow-free diffusion-based
gradients of the chemokine CXCL12, and tracked primary
human CD4+ T cells exposed to these gradients with high
spatiotemporal resolution. Exposure to gradients of CXCL12
induced migration of CD4+ T cells towards higher concentrations in the gradient, increased the migration velocity and
track straightness of the cells. Microfluidic spatiotemporal
retrieval and subsequent droplet digital PCR analysis of the
cells in relation to their migration characteristics i.e. migration towards the gradient or not revealed that the expression levels of CXCR4 the receptor of CXCL12 is much
greater in cells with enhanced chemotaxis as compared to
cells in which no chemotaxis was induced. Taken together
we demonstrate here the potential of our microfluidic system
to induce primary human T cell migration in diffusion-based
chemokine gradients, monitor cell migration of individual
cells and retrieve cells with spatiotemporal resolution for
off-chip analysis with powerful new gene expression methods
like digital-PCR.

Material and methods

Chip design and fabrication
Transparency photomasks (Fineline Imaging, Colorado Springs,
CO, US) were generated using AutoCAD (Autodesk, Inc., San
Rafael, CA, US) outline of the designed multi-layer device.
Multi-layer PDMS soft-lithography was used for fabrication of
chips, as described previously.14,15 A more detailed description is given in the supplementary methods file.
Chip operation and control
Control channels were connected to solenoid valves (Festo,
Dietikon, Switzerland) that were controlled with a custom
LabVIEW (National Instruments, Austin, TX, US) graphical
user interface and experimental scripts program we wrote,
and were electronically controlled using an established
control box system.16 Optimal closing pressures of push-up
PDMS membrane valves were individually determined for all
used chips, and ranged between 11.5 bar. A more detailed
description is given in the supplementary methods file.
Reagents and surface functionalization
For avoiding undesired attachment of cells flow channels
were treated with the copolymer pluronic 10 mg mL1

This journal is The Royal Society of Chemistry 2015

Paper

(Millipore, Zug, Switzerland) for 3 min, followed by washing

with PBS for 30 min. Next, migration chambers were coated
with fibronectin at 250 g mL1 concentration (Millipore,
Zug, Switzerland) for 60 min followed by blocking with
RPMI 1640 containing 10% FCS, 50 U mL1 penicillin, and
50 mg mL1 streptomycin (R10, all from Life Technologies,
Zug, Switzerland). R10 containing CXCL12 chemokine at
1 g mL1 concentration (Preprotech, London, UK) was used
for the generation of chemokine gradients. Cells were
harvested for off-chip analysis using 0.05% trypsinEDTA
(Life Technologies, Zug, Switzerland).
Generation of stable gradients using temporally modulated
sourcesink flow patterns
Stable diffusion-based chemokine gradients were generated
and maintained as previously described by using a switching
source-sink flow pattern.14 Briefly, the channels at the top
and the bottom of the cell culture chamber/migration chamber were sequentially refilled with fresh medium either with
the chemokine or without it. Therefore a local high concentration (source) and a low concentration was established
where as between the gradient is built up and maintained by
diffusion. The sink or source was replaced every 4 minutes as
reported before.
PBMCs and T-cell isolation
EDTA blood was obtained from healthy volunteers after
informed consent (study approved by the institutional review
board of both cantons of Basel). PBMCs were isolated
from EDTA blood by Ficoll gradient centrifugation using
SepMate tubes (Stemcell Technologies, Grenoble, France).
CD4+ T cells were purified by using negative selection with
immunomagnetic bead separation (Stemcell Technologies,
Grenoble, France). The purity of the isolated CD4+ T-cell
population was consistently greater than 95%.
Imaging and data analysis
Cells were tracked using an automated inverted microscope
(Nikon Ti, 10 and 20 ELWD Objective) equipped with a
stage-top incubator controlling for temperature (37 C),
CO2-concentration (5%) and humidity (90%), a digital CMOS
camera (ORCA-Flash 4.0, Hamamatsu Photonics) and the
microscope software Nikon AR. Image processing and data
analysis was carried out using Imaris with a tracking-tool
extension (Bitplane Inc.) and Matlab 2010 (MathWorks Inc.).
A more detailed description is given in the ESI.
Isolation of mRNA, cDNA generation and droplet digital
PCR (ddPCR)
Subpopulations of cells harvested from chip were lysed and
total RNA was purified from T cells using the RNeasy Mini Kit
(Qiagen, Hilden, Germany). Total RNA was used for reverse
transcription (Promega, Madison, WI). For droplet digital
PCR (ddPCR), 2.6 l cDNA (final concentration 350 ng l1)

Lab Chip, 2015, 15, 12761283 | 1277

View Article Online

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Paper

was combined with 10 l of 2 ddPCR Super Mix for Probes

(Bio-Rad), solution primers and probes. Deionized sterile
water was added to bring the total volume to 20 l. The
following primers and hydrolysis probe were used: CD3 forward: TCC GAG ATC GAG ATG ATG; CD3 reverse: GGA AGG TAC
AGT TGG TAA TG; CD3 probe: 6FAM-AGG TTC ACT TGT TCC
GAG CCC A-BHQ-1. For quantification of CXCR4-expression
a CXCR4 TaqMan Gene Expression Assay (FAM-MGB) was
used (Life Technologies, Zug, Switzerland). The resultant
20 l ddPCR solutions were transferred to DG8 cartridges,
emulsified by the QX100 Droplet Generator (Bio-Rad); and
the emulsions were placed in a Veriti thermal cycler (Life
Technologies) for PCR. The temperature schedule for PCR
was: 1, 95 C for 10 min; 40, 94 C for 30 s followed by
60 C for 1 min; 1, 98 C for 10 min; and the ramp speed
was 2.5 C s1. After, the emulsions were analyzed using the
QX100 Droplet Reader and QuantaSoft software (Bio-Rad).
Fluorescence from the emulsion droplets was quantified in
the Absolute Quantification setting, and the signal threshold
was manually set by applying to all wells the threshold

Lab on a Chip

value determined by auto-analysis of one of the most concentrated samples.

Statistical analysis
Data nested in the different groups were analyzed with the
KolmogorovSmirnov test. MannWhitney test was performed
in the case of non-normality. Data with normal distribution
were assessed by paired Student 2-sided t test. Values of
p < 0.05 were considered to be significant.

Results and discussion

Microfluidic cell culture system for real-time analysis of
single-cell chemotaxis
For assessing cell migration, we developed a microfluidic
device that allows (i) localized positioning of human T cells,
(ii) the generation of diffusion-based flow-free chemokine
gradients in parallel chambers containing cells, (iii) analysis
of cell migration with video microscopy and automated tracking and (iv) retrieving of cells according to their position in

Fig. 1 Overview of microfluidic chemotaxis and cell retrieval device. (A) Schematic overview of the geometry and functionality of an individual
microfluidic migration chamber. Cells can be seeded into and be harvested from up to 10 positions inside the gradient chamber using side
channels (see ESI movie S1). (B) Supplying multiplexer architecture of the 3-side-channel device. (C) Schematic and (D) actual photograph of
actual device with 6 microfluidic migration chambers and the respective multiplexers between inlets and outlets (green structures: flow channels;
red structures: control channels). (E) Overview of the fully integrated microfluidic system with inserts showing the microfluidic chip mounted on
the automated microscope.

1278 | Lab Chip, 2015, 15, 12761283

This journal is The Royal Society of Chemistry 2015

View Article Online

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Lab on a Chip

the chamber, as illustrated in Fig. 1A. The core component of

this 2-layer polydymethylsiloxane (PDMS) device are migration chambers (l = 900 m, w = 250 m, h = 25 m)
containing 38 ports at both long ends originating from
two multiplexers localized proximally and distally of the
x-position chamber while the ports at the two short ends
of the chamber are connected to support channels (Fig. 1B).
All ports are equipped with independently addressable PDMS
membrane valves for controlled flow of fluids and cells. The
device contains 6 independent migration chambers, 12 inlets
for reagents and media, and 4 waste and cell harvesting
outlets. Support channels connect the reagent inlets with the
migration chambers, multiplexers, and the outlets where
cells can be retrieved (Fig. 1C).
The actual assembly of the individual components of
the device is given in Fig. 1D, while Fig. 1E gives an overview
of the full automated microfluidic system. As shown in
Fig. 2A/B and ESI movie S1 the device can simultaneously
generate 6 independent diffusion-based chemical gradients
using a source-sink configuration,14 and cells can be cultured
and monitored in these gradients. By flowing a different
molecule (i.e. chemokine) through the source and sink
channels that are orthogonal to the migration chamber and
allowing diffusion, we can generate flow-free chemokine
gradients in the migration chambers where the type, steepness, mean concentration and duration of each gradient can
independently be controlled. Spatially and temporally opposing gradients can be generated, and the polarity or ligand
type of the gradients can be switched when needed. As the
device relies on diffusion for mass transport and not fluid
flow, the above-mentioned operations can be performed
without disturbing the positions or migration behavior of
the cells.17 The cell culture conditions, including culture
media and gas exchange rates and humidity were optimized
to allow week-long experiments with excellent cell viability
and growth.18
To realize a complete system for cell migration studies,
we integrated this device to an automated microscope and
tracking software, where various tasks including surface
treatments, cell seeding, gradient generation and video
microscopy is computer controlled through a graphical user
interface and custom scripts written in Matlab or Labview.
The combination of automation, nanoliter-sized chambers
and controlled laminar flow conditions allows our system to
culture and carefully analyse small populations of cells if
needed, constituting a major advantage when working with
rare cell types. Use of the multiplexer localized proximally and
distally of specific migration chambers allows localized positioning of lymphocytes in distinct sections of the migration
chambers. As illustrated in Fig. 2C/D and shown in ESI
movie S2 for primary human CD4+ T cells, lymphocytes can
be positioned in specific regions of the migration chamber:
for example into an area of 250 200 micrometers in the middle section of the 250 900 micrometer sized culture chamber. Following positioning of the cells, we supply chambers
with fresh cell culture medium through diffusion from the

This journal is The Royal Society of Chemistry 2015

Paper

sides for 45 minutes to allow attachment of the cells to the

ECM-substrate. The cells can be monitored with time-lapse
microscopy or stained for immunohistochemistry (Fig. S3).
When needed the cells can be retrieved from various positions inside the migration chamber using the side-ports and
the multiplexer directed towards the cell outlets. Fig. 2E/F
and ESI movie S4 illustrate the controlled retrieval of
densely seeded T cells from the migration chamber based on
their horizontal position. Following retrieval from the migration chamber into the distal multiplexer, cells can be transferred to one of the outlets for retrieval with a pipette and
off-chip analysis. If needed, retrieved cells can also be positioned back into the migration chambers without taking
them off the chip. For retrieving attached cells from specific
regions of the migration chambers, trypsin is gently flown
into the migration chamber, and the chamber is then sealed.
The cells incubated in trypsin detach from the PDMS
substrate within a few minutes, but they remain fixed in their
original positions, as there is no convective mixing in the
used microfluidic conditions. The detached cells can then
be retrieved based on their horizontal positions. To illustrate
the specific retrieval of cells, unlabeled cells and cells labeled
with calcein red or calcein green were positioned in the
middle, the left and the right section of the chamber, respectively (Fig. 2G). Following attachment, cells were detached
and sequentially retrieved from defined regions of the migration chamber by slowly flowing trypsin via the corresponding
channels of the multiplexers into harvesting outlets. By doing
so, 8590% of labeled cells from specific regions of the
migration chamber were harvested sequentially based on
their position in the chamber into different harvesting outlets, as shown exemplarily for an outlet that contains calcein
green stained cells from the right section of the chamber
(Fig. 2G, right panel). Purity of the retrieved cells in the
harvesting outlets ranged between 82% and 90%. Taken
together, our system comprises a significant step in analysis
of cell migration because it allows precise and if necessary
confined positioning of cells in our migration chamber, the
establishment and control of diffusion-based chemokine
gradients, automated live cell microscopy and cell tracking,
and retrieving of individual cells based on their position
(i.e. migration speed) in the chamber.

Microfluidic chemotaxis of primary human CD4+ T cells

Migration of human T cells plays a central role in protective
immunity but also in the pathogenesis of autoimmune diseases such as MS. The latter is supported by the fact that two
highly efficacious drugs for the treatment of MS impact on
T cell migration: Fingolimod acts as a functional antagonist
on the S1P receptor (S1PR), hereby preventing recirculation
of T cells from SLT to peripheral blood;19 natalizumab blocks
adhesion of blood T cells to endothelial cells and as a consequence migration across the bloodbrain barrier.20
To assess migration characteristics of human T cells,
CD4+ T cells were enriched by negative isolation with

Lab Chip, 2015, 15, 12761283 | 1279

View Article Online

Lab on a Chip

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Paper

Fig. 2 Functionality of microfluidic chemotaxis and cell retrieval device. (A) Generation of diffusion-based gradients of a dye in microfluidic
migration chambers (upper panels) and quantification of the respective relative concentrations in the chambers (lower panels). (B) Example of
simultaneous generation of various diffusion-based gradients with food dye (chamber 14 and 6) and fluorescent FITC-dextran molecules
(chamber 5) with actual pictures of the individual chambers (left column) and measured relative concentrations in the chambers (right column).
(C and D) Localized seeding of primary human T cells in microfluidic migration chamber (see also ESI movie S2). (E and F) Retrieving of
unattached human T cells from microfluidic migration device for off-chip analysis (see ESI movie S4). (G) Specific retrieving of attached
primary human T cells according to position in the chamber and transfer to harvesting outlet for off-chip analysis. Only green dyed cells can
be seen in the harvesting outlet.

1280 | Lab Chip, 2015, 15, 12761283

This journal is The Royal Society of Chemistry 2015

View Article Online

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Lab on a Chip

magnetic beads from healthy donor EDTA-blood and placed

in fibronectin-coated microfluidic migration chambers. Following attachment, T cells were either exposed to gradients
of CXCL12 by providing CXCL12 containing culture medium
from one side of chamber and only medium from other side,
or cultured under control conditions by providing cell culture
medium from both sides at the same rate. Most T cells
cultured under control conditions migrated spontaneously in a
non-directional fashion (Fig. 3A/B; and ESI movie S5) with a
mean track speed (track length [m]/time[min]) of 2.6 m min1,
resulting in a mean track length of 266 m during a 2 h
observation (Fig. 3C/D). Exposure to a CXCL12 gradient
induced migration of T cells (Fig. 3A/B; and ESI movie S6)
and resulted in significantly increased displacement towards

Paper

higher concentrations of the gradient (Fig. 3F). Also, the

mean track speed was increased resulting in a significantly
longer mean track length of 414 m (Fig. 3C/D). Directionality
(distance from starting position to final position [m]/track
length [m]) of cell migration was also significantly increased
in cells exposed to a CXCL12 gradient (Fig. 3E).

Enhanced chemotaxis of CD4+ human T cells in CXCL12

gradients is linked to CXCR4-expression
Migration of human T cells towards chemoattractants has been
linked to expression levels of the respective receptors.21,22
In one study using a transwell migration system, levels of
CXCR5-expression in human CD8+ T cells were correlated to

Fig. 3 CXCL12-directed migration of human CD4+ T cells in microfluidic migration device. (A) Overview of microfluidic migration chambers
with plotted migration trajectories of CD4+ T cells in the absence (left panel) and the presence of CXCL12 gradient (right panel; color code
indicates timepoint during 2 h tracking). (B) Trajectories of migrating T cells plotted on a common starting point in the absence (left panel) and
presence (right panel) of CXCL12 gradient. (C) Mean migration speed ( SEM, standard error of the mean), (D) mean track length ( SEM) and
(E) mean straightness ( SEM) of CD4+ T cells cultured under control conditions or when exposed to the CXCL12-gradient. (F) Mean
x-displacement ( SEM) of primary human T cells in the absence (control) or presence of CXCL12 gradient.

This journal is The Royal Society of Chemistry 2015

Lab Chip, 2015, 15, 12761283 | 1281

View Article Online

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

Paper

Fig. 4 mRNA expression of chemokine receptor CXR4 in weakly

or strongly migrating CD4+ T cells in gradients of the chemokine
CXCL12. The cells were harvested from different positions in the
migration chamber after chemotaxis, observed under video microscopy.
(A) Illustrative plots of digital droplet PCR analysis of CXCR4-mRNA
expression. (B) Mean mRNA-expression levels of CD3 (control) and
CXCR4 ( SEM; standard error of the mean) in primary human CD4+
T cells in correlation to their migration in gradients of CXCL12
(weak vs. strong chemotaxis).

chemotactic response towards CCL5, the ligand of

CXCR5.22 Despite the fact that transwell migration systems
are unapt to define migration characteristics of lymphocytes on a single cell level, initial sorting of the cells with
anti-CXCR5 potentially impacts on migration in gradients
with CCL5.
To overcome these limitations we determined levels of
CXCR4-expression in primary human T cells following
migration experiments in gradients of CXCL12. Specifically,
cells were placed at one end of our microfluidic migration
device and exposed to a gradient of CXCL12. Following
migration of cells in the gradient with video microscopy,
we harvested the cells via the side channels of the migration chamber in correlation to their position in the chamber, which corresponds to their chemotactic response to
CXCL12. By doing so we specifically harvested on the one
hand a pooled subpopulation of primary human CD4+ T cells
that had migrated in gradients of CXCL12 and on the
other hand a subpopulation of the same type of cells that
had not shown any chemotactic response. We analysed the
expression of CXCR4 in both group of pooled cells (~20 cells)
using microfluidic droplet-based digital-PCR, which allows
absolute copy-number quantitation of nucleic acids.23,24 Interestingly, we observed significantly higher levels of CXCR4expression in cells that had migrated strongly in gradients
of CXCL12 as compared to cells that had not shown a chemotactic response (Fig. 4A/B). Taken together these findings
indicate that in primary human T cells migration in gradients of CXCL12 is linked to levels of CXCR4-expression.

Conclusion
Migration of primary human T cells is classically studied
in vitro by the use of transwell migration assays.8 These
techniques helped recapitulating important insights into

1282 | Lab Chip, 2015, 15, 12761283

Lab on a Chip

T cell migration gained from animal studies also in human

T cells. However, transwell migration assays are inherently
limited in controllability of experimental conditions, allow
assessment of lymphocytes only on a population level, and
do not allow real-time longitudinal analysis of individual
cells or populations. In this report we addressed these limitations and developed a microfluidic device specifically
engineered to study migration of cells, such as primary
human T cells. We combined this device with computer
control, automated cell tracking and off-chip genetic analysis to realize a complete system and pipeline of protocols
for quantitative studies of cell migration. This approach
enabled us to (i) generate stable chemokine gradients,
(ii) track individual cells in such gradients in real time and
(iii) retrieve cells as a function of their position in the
device i.e. their migration characteristics for further
off-chip analysis such as digital gene expression profiling.
Few previous studies addressed T cell migration using
non-traditional techniques like microfluidics. One study
found increased directed migration towards the chemokine
CCL21 compared to CCL19, and no additive effects of these
two CCR7-ligands.25 In contrary, the presence of background
CCL21 induced repulsive chemotaxis away from gradients of
CCL19, illustrating the significance of studying T cell migration in controllable chemokine gradients. Further, characteristics of T cell migration in competing chemokine gradients
of CCL19 and CXCL12 appeared to be related to the specific
position of individual cells in the gradients, a finding that
would have been masked in bulk assays.26 These highlighted
the importance of studying T cell migration under controlled
conditions with high spatio-temporal resolution and on a
single cell level. However, these studies are based on microfluidic devices in which the gradients are generated using
parallel streams of laminar flow across the culture chamber,
hereby significantly contrasting the flow-free nature of in vivo
chemokine-gradients. Further, spatially specific retrieval of
migrating cells or different subpopulations based on migration speeds has not been possible from microfluidic devices
before this study.
As a proof of concept of these functionalities, we assessed
expression levels of the CXCL12-receptor CXCR4 in primary
human CD4+ T cells as function of their migration towards
higher concentrations of CXCL12. Cells responding to
CXCL12 expressed significantly higher levels of CXCR4 mRNA
as compared to cells that did not migrate in gradients of
CXCL12. These observations are in line with a previous study
that linked expression levels of CXCR5 on T cells with migratory responses towards the CXCR5-ligand CCL5.22 In this
study CD4+ and CD8+ T cells were sorted and subjected to
CCL5-directed migration exclusively on the basis of CXCR5expression, which was not assessed in T cell subpopulations.
Taken together, our microfluidic migration device allows
linking expression levels of given molecules as a function of
migration characteristics of specific individual cells, which
has the potential to significantly add to our understanding of
how migration of cells is regulated on a single cell level.

This journal is The Royal Society of Chemistry 2015

View Article Online

Lab on a Chip

Acknowledgements

Published on 05 December 2014. Downloaded by ETH-Zurich on 06/05/2015 09:23:34.

M. M. is supported by the University of Basel Research Grant,

the Swiss Multiple Sclerosis Society and the Novartis Foundation. S. T. acknowledges support from the ERC Starting Grant
337986 (SingleCellDynamics), Swiss National Science Foundation, SNF Systems X Grant NeuroStemX and Swiss NCCR
Molecular Systems Engineering. We thank Th. Horn and
A. Ponti from the Single Cell Facility of the DBSSE of the ETH
Zrich for support with imaging and image analysis.

References
1 T. Lmmermann and M. Sixt, The microanatomy of T-cell
responses, Immunol. Rev., 2008, 221, 2643.
2 R. Horwitz and D. Webb, Cell migration, Curr. Biol., 2003,
13, R756R759.
3 U. H. Andrian, von and T. R. Mempel, Homing and cellular
traffic in lymph nodes, Nat. Rev. Immunol., 2003, 3, 867878.
4 P. Friedl and B. Weigelin, Interstitial leukocyte migration
and immune function, Nat. Immunol., 2008, 9, 960969.
5 K. F. Swaney, C.-H. Huang and P. N. Devreotes, Eukaryotic
Chemotaxis: A Network of Signaling Pathways Controls
Motility, Directional Sensing, and Polarity, Annu. Rev.
Biophys., 2010, 39, 265289.
6 F. Sallusto, et al., T-cell trafficking in the central nervous
system, Immunol. Rev., 2012, 248, 216227.
7 J. Marsal and W. W. Agace, Targeting T-cell migration in
inflammatory bowel disease, J. Intern. Med., 2012, 272,
411429.
8 M. Falasca, C. Raimondi and T. Maffucci, Boyden chamber,
Methods Mol. Biol., 2011, 769, 8795.
9 S. Vedel, S. Tay, D. M. Johnston, H. Bruus and S. R. Quake,
Migration of cells in a social context, Proc. Natl. Acad. Sci.
U. S. A., 2013, 110, 129134.
10 L. C. Kam, K. Shen and M. L. Dustin, Micro- and Nanoscale
Engineering of Cell Signaling, Annu. Rev. Biomed. Eng.,
2013, 15, 305326.
11 J. Melin and S. R. Quake, Microfluidic Large-Scale Integration:
The Evolution of Design Rules for Biological Automation,
Annu. Rev. Biophys. Biomol. Struct., 2013, 36, 213231.
12 S. Tay, et al., Single-cell NF-B dynamics reveal digital
activation and analogue information processing, Nature, 2010,
466, 267271.

This journal is The Royal Society of Chemistry 2015

Paper

13 L. Mazutis, et al., Single-cell analysis and sorting using

droplet-based microfluidics, Nat. Protoc., 2013, 8, 870891.
14 T. Frank and S. Tay, Flow-switching allows independently
programmable, extremely stable, high-throughput diffusionbased gradients, Lab Chip, 2013, 13, 12731281.
15 M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and
S. R. Quake, Monolithic microfabricated valves and pumps
by multilayer soft lithography, Science, 2000, 288, 113116.
16 R. Gmez-Sjberg, A. A. Leyrat, D. M. Pirone, C. S. Chen and
S. R. Quake, Versatile, fully automated, microfluidic cell
culture system, Anal. Chem., 2007, 79, 85578563.
17 G. M. Walker, et al., Effects of flow and diffusion on
chemotaxis studies in a microfabricated gradient generator,
Lab Chip, 2005, 5, 611618.
18 R. A. Kellogg, R. Gmez-Sjberg, A. A. Leyrat and S. Tay,
High-throughput microfluidic single-cell analysis pipeline
for studies of signaling dynamics, Nat. Protoc., 2014, 9,
17131726.
19 M. Mehling, T. A. Johnson, J. Antel, L. Kappos and A. Bar-Or,
Clinical immunology of the sphingosine 1-phosphate receptor modulator fingolimod (FTY720) in multiple sclerosis,
Neurology, 2011, 76, S20S27.
20 T. Derfuss, J. Kuhle, R. Lindberg and L. Kappos,
Natalizumab Therapy for Multiple Sclerosis, Semin. Neurol.,
2013, 33, 026036.
21 S. Qin, et al., Expression of monocyte chemoattractant protein-1
and interleukin-8 receptors on subsets of T cells: correlation
with transendothelial chemotactic potential, Eur. J. Immunol.,
1996, 26, 640647.
22 C. Desmetz, et al., The strength of the chemotactic response
to a CCR5 binding chemokine is determined by the level of
cell surface CCR5 density, Immunology, 2006, 119, 551561.
23 D. S. Tawfik and A. D. Griffiths, Man-made cell-like compartments for molecular evolution, Nat. Biotechnol., 1998,
16, 652656.
24 D. Dressman, H. Yan, G. Traverso, K. W. Kinzler and
B. Vogelstein, Transforming single DNA molecules into
fluorescent magnetic particles for detection and enumeration
of genetic variations, Proc. Natl. Acad. Sci. U. S. A., 2003, 100,
88178822.
25 S. Nandagopal, D. Wu and F. Lin, Combinatorial Guidance
by CCR7 Ligands for T Lymphocytes Migration in
Co-Existing Chemokine Fields, PLoS One, 2011, 6, e18183.
26 F. Lin and E. C. Butcher, T cell chemotaxis in a simple
microfluidic device, Lab Chip, 2006, 6, 14621469.

Lab Chip, 2015, 15, 12761283 | 1283