Content uploaded by Pascal Behr
Author content
All content in this area was uploaded by Pascal Behr on Jun 13, 2018
Content may be subject to copyright.
FluidFM: Combining Atomic Force
Microscopy and Nanofluidics in a
Universal Liquid Delivery System for
Single Cell Applications and Beyond
Andre´ Meister,†,§ Michael Gabi,‡,§ Pascal Behr,‡Philipp Studer,‡,|Ja´nos Vo¨ro¨s,
‡
Philippe Niedermann,†Joanna Bitterli,†Je´roˆ me Polesel-Maris,†,⊥Martha Liley,†
Harry Heinzelmann,†and Tomaso Zambelli*,‡
Swiss Center for Electronics and Microtechnology, CSEM SA, rue Jaquet Droz 1,
2002 Neuchaˆtel, Switzerland, and Laboratory of Biosensors and Bioelectronics,
Institute for Biomedical Engineering, ETH Zurich, Gloriastrasse 35,
8092 Zurich, Switzerland
Received April 30, 2009
ABSTRACT
We describe the fluidFM, an atomic force microscope (AFM) based on hollow cantilevers for local liquid dispensing and stimulation of single
living cells under physiological conditions. A nanofluidic channel in the cantilever allows soluble molecules to be dispensed through a
submicrometer aperture in the AFM tip. The sensitive AFM force feedback allows controlled approach of the tip to a sample for extremely local
modification of surfaces in liquid environments. It also allows reliable discrimination between gentle contact with a cell membrane or its
perforation. Using these two procedures, dyes have been introduced into individual living cells and even selected subcellular structures of
these cells. The universality and versatility of the fluidFM will stimulate original experiments at the submicrometer scale not only in biology
but also in physics, chemistry, and material science.
In the last two decades, atomic force microscope (AFM)1
has revolutionized surface science at the nanoscale, not
only as a microscope with atomic resolution2,3 but also as
a nanofabrication tool displacing atoms,4manipulating and
reshaping complex molecules,5and delivering molecules
extremely locally on a surface.6AFM is also starting to
contribute to cell biology, elucidating features of membrane
morphology,7mechanical properties,8division mechanism,9
metabolism,10 voltage-induced deformation11 or adhesion,12,13
and potentially as a cancer detection tool.14 However, in these
and similar experiments, the use of the AFM to manipulate
cells remains largely limited to (sometimes crude) mechanical
processes such as indentation, compression, stretching, and
scraping. One notable exception to this list is the transport
of molecules into a cell via reversible binding to an AFM
tip.15-17
Here we present a novel use of AFM principles to
manipulate living biological cells by stimulation through
delivery of active agents directly from a solution; the
fluidFM (for fluidic force microscope) combines the ac-
curate force-controlled positioning of AFM with the
versatility of fluidics.18 A microsized channel is integrated
in an AFM cantilever and connected via channels in the
AFM chip holder to a delivery system, thus creating a
continuous and closed fluidic channel that can be filled
with an arbitrary chosen solution and can be immersed in
a liquid environment (Figure 1a). An aperture in the AFM
tip at the end of the cantilever allows liquids to be dispensed
locally. Force feedback is ensured by a standard AFM laser
detection system that measures the deflection of the cantilever
and thus the force applied by the tip to the sample during
approach and dispensing.
While this approach is similar to microinjection using
glass pipettes, there are a number of essential differences.
Microinjection uses optical microscopy to control the
position of the glass pipet tip both in the xy plane and in
the zdirection (via image focusing). As consequence of
the limited resolution of optical microscopy, subcellular
domains cannot be addressed and tip contact with the cell
* To whom correspondence should be addressed. E-mail: zambelli@
biomed.ee.ethz.ch.
†Swiss Center for Electronics and Microtechnology, CSEM SA.
‡Institute for Biomedical Engineering, ETH Zurich.
§These authors contributed equally to the work.
|Present address: London Centre for Nanotechnology, 17-19 Gordon
Street, London WC1H 0AH, United Kingdom.
⊥Present address: CEA Saclay, DSM/IRAMIS/SPCSI, 91191 Gif-sur-
Yvette, France.
NANO
LETTERS
2009
Vol. 9, No. 6
2501-2507
10.1021/nl901384x CCC: $40.75 2009 American Chemical Society
Published on Web 05/19/2009
membrane cannot be discriminated from tip penetration
of the membrane. Cells are often lethally damaged and
skilled personal are required for microinjection.19-23 The
limited resolution of this method and the absence of
mechanical information contrast strongly with the high
resolution imaging and the direct control of applied forces
that are possible with AFM. Precise force feedback reduces
potential damage to the cell; the cantilever geometry
minimizes both the normal contact forces on the cell and
the lateral vibrations of the tip that can tear the cell membrane
during microinjection; the spatial resolution is determined
by the submicrometer aperture so that injection into subcel-
lular domains becomes easily achievable.
The inclusion of force control in the dispensing/injection
system also enables extremely local dispensing on surfaces.
With force feedback, the dispensing tip can be brought into
very close contact with a sample without risk of damage
either to the sample or to the tip so that a minimal amount
of solution can be released directly at the chosen spot. Local
dispensing will allow precise stimulation of cells, applying
biomolecules to a well-defined position on one cell that may
either be isolated or part of an ensemble or network of cells.
It will also open new possibilities in biochemical analysis,
for example, allowing analyte solutions to be dispensed
locally on an array of antibodies.
Microchanneled AFM cantilevers were produced using an
original microfabrication process based on thermal fusion
bonding of two previously etched silicon wafers to create
cavities lined with silicon dioxide within the body of the silicon.
This was followed by a selective silicon etch to produce free-
standing silicon dioxide structures.24,25 The result is a hollow
glass cantilever attached to a silicon chip (see Supporting
Information Figures S1 and S2). The closed microchannel
within the cantilever enters the silicon chip and ends in an
Figure 1
.
The fluidFM. (a) Diagram showing a microchanneled cantilever chip fixed to a drilled AFM probeholder. The system was
shown to be watertight up to 4 bar internal overpressure but overpressures of the order of 10 mbars were usually applied. The
fluidFM can be operated in air or with the whole system (probeholder and chip) immersed in a liquid. The external liquid or bath and
the liquid in the microchannel may be the same or may be different. During liquid dispensing, the substrate can be simultaneously
observed with an optical microscope either through the transparent probe holder or through the glass substrate. (b) Scanning electron
micrograph of the aperture beside the apex of the pyramidal AFM tip for the intracellular injection experiments as in Figure 3. The
tip is made of silicon nitride (Si3N4) with a wall thickness of 100 nm. The apertures are milled by focused ion beam with a diameter ranging
from 1 µm down to 100 nm. Thin Si3N4walls are compulsory for milling such nanosized holes. Before the FIB milling, a metallic layer
(gold or platinum) is deposited on both sides of the cantilever to avoid charging effects during milling and increase the reflectivity for the
AFM laser. (c) Scanning electron micrograph of the aperture at the apex of the pyramidal AFM tip for the cell staining experiments by
gentle contact as in Figure 4.
2502 Nano Lett., Vol. 9, No. 6, 2009
open reservoir. An aperture at the apex of the cantilever tip
is opened using focused ion beam milling (FIB). Profiting
from the flexibility of FIB, the relative positions of the hole
and the tip apex can be chosen to facilitate the experiments
to be carried out either at the side of the apex for injection
into a cell after perforation of its membrane (Figure 1b) or
at the apex for dispensing onto a surface or gentle contact
with the cell membrane (Figure 1c). Similar microchanneled
silicon nitride probes with an aperture located at the base of
the tip have been fabricated for Dip-Pen applications that
are limited to operation in controlled air atmosphere because
such micromachined fountain pens were not combined with
a drilled probeholder.24,26,27 Also tapered glass pipettes have
not been used for single-cell manipulation yet possibly
because of the limited control on the aperture location and
the apex sharpness.28-30
The force detection system of the fluidFM is so sensitive
that the interactions between tip and sample can be
reduced to the piconewton range enabling two different
strategies when addressing fragile samples such as living
cells. At small force set points, the hollow cantilever with
the aperture beside the apex as in Figure 1b can be brought
into gentle but close contact with cells without unwished
puncturing of the cell membrane (see Figure 2a). On the other
side, membrane perforation for intracellular injection is
simply achieved by selecting a higher force set point taking
advantage of the extremely sharp tip (radius of curvature on
the order of tens of nanometers) as indicated by the typical
indentation jump of the force distance curve of Figure 2a.31
If the aperture is milled at the tip apex as in Figure 1c, no
indentation jump is observed in the force-distance curves
(see Figure 2b) meaning that the tip with this configuration
remains in “gentle contact” with the membrane without
tearing it.
Intracellular injection (see Figure 3a) was demonstrated
by using fluorescein isocyanate sodium salt (FITC) as dye
because it neither binds to the cell membrane nor does it
spontaneously diffuse across it. Figure 3b is an AFM
topography image of a myoblast cell taken in the amplitude-
modulation (AM) mode with the fluidic channel filled with
a physiological solution containing the FITC dye. The AM
mode decreases the force exerted on the cell because the
effective stiffness of the probe becomes keff )k/Qwith Q
as the quality factor of the cantilever in liquid environment
(see Supporting Information, Figure S2b). The cell is resolved
with a spatial accuracy down to 200 nm details confirming
the ability of the microchanneled cantilevers to act both as
scanning probe and as nanopipette. Upon completion of the
AFM imaging, the probe was stopped and positioned with
the AFM controller over a selected point of the cell. As we
are relying on the force-distance curve of the tip approach
to the cell in order to determine the set point necessary to
penetrate into the cell membrane, we switched to the contact
mode, introduced the extremity of the tip into the cell,
injected the fluorescent solution by hydrostatic pressure for
few seconds, and finally retracted the tip far away from the
selected cell. From simulations of the flow rates through the
aperture of such hollow cantilever (see Supporting Informa-
tion, Figure S3), we can estimate that the injected amount
of FITC solution is less than 10 fL. Figure 3c is a fluorescent
image taken with a confocal scanning laser microscope
(CLSM) of the same cell in Figure 3b. It can be seen that
that the introduction of liquid into the cell has resulted in
the appearance of a fluorescence signal without an important
change in the cell volume. To demonstrate that the dye is
inside the cell, z-stack images were taken. The corresponding
cross section shows a fluorescent intensity not only at the
cell membrane but also in the cytosol. The fact that the dye
remains in the cell indicates that the membrane recovers and
reseals tightly after the withdrawal of the tip. Figure 3d,e
shows intracellular injection of FITC into three neurons one
after the other with the same microchanneled cantilever
proving that probes can be used for consecutive manipulation
without clogging.
Figure 2. Contact regimes with the cell membrane. (a) Typical
force distance curve on a myoblast cell for a cantilever with an
aperture milled beside the pyramid apex like in Figure 1b (k)
0.3 N/m, approach of 50 nm/s); the arrow indicates the typical
discontinuity corresponding to indentation of the cell membrane.
The set point of around 3 nN thus separates the two contact regimes,
the gentle contact on the cell membrane and the membrane
perforation for intracellular injection. (b) Typical force distance
curve on a myoblast cells for a cantilever with an aperture of 1 µm
milled at pyramid apex like in Figure 1c (k)0.3 N/m, approach
of 50 nm/s); no indentation discontinuity appears with this aperture
configuration.
Nano Lett., Vol. 9, No. 6, 2009 2503
Figure 3. Intracellular injection by force-controlled perforation of the cell membrane. (a) Diagram showing the intracellular injection procedure
by membrane perforation (not to scale). The aperture is intentionally milled beside the apex in order to exploit the nanometric curvature radius of
the pyramid to facilitate the membrane indentation (set point of 3 nN at least). (b) Differential interference contrast image of a myoblast and
corresponding AFM image (amplitude of 30 nm, scan rate of 0.3 s, cantilever as in Figure S2). (c) CLSM stack image of the same cell for z
)1µm after FITC intracellular injection of 15 s and corresponding cross section along the dashed red line; the fluorescent thickness of 6
µm is in accordance with the height of the cell measured by AFM. (d,e)Differential interference contrast image of neuroblastoma and
corresponding fluorescent image upon FITC intracellular injection of the three ones in the middle.
2504 Nano Lett., Vol. 9, No. 6, 2009
The gentle contact procedure is demonstrated by staining
live neuroblastoma cells as illustrated in Figure 4a. A
microchannelled cantilever filled with CellTracker green, a
membrane-permeant dye, was positioned over one cell using
an optical microscope.32 It was then touched onto the cell
and left in gentle contact until the dye diffused into the
cytoplasm. CellTracker green was used because its fluores-
cence depends on the enzymatic activity of a live cell.
CellTracker green is a nonfluorescent chloromethyl derivative
that passes through the cell membrane into the cytoplasm
where enzymatic activity converts it into a fluorescent
membrane-impermeable species.
Figure 4b shows several neuroblastoma cells on a glass
slide in a physiological buffer. The fluidFM cantilever filled
with CellTracker green was touched onto the cell marked
by the red circle and left for 15 min, which is the time
necessary for the cell to metabolize the agent diffusing
through the membrane and express the fluorescent product.
The cell addressed with the tip became fluorescent whereas
neighboring cells were not, demonstrating that the dye
was indeed introduced into the selected cell, that the cell
was not lethally damaged, and that the staining procedure
does not affect nearby cells.
The precise delivery of a dye into a cellular substructure
can also be achieved thanks to the imaging capacities of
the AFM. A region is first scanned with fluidFM and the
topographical information is used to access structures too
small for optically controlled microinjection systems.
Figure 4c is a differential interference contrast image of two
neuroblastomas connected by a neurite that forms a varicos-
ity-like structure in the middle. The cantilever was positioned
on top of the varicosity using an AFM profile of the sample
(Figure 4c, bottom) and a solution containing the membrane-
permeant dye, acridine orange, was delivered directly onto
the varicosity for 1 min.32 The corresponding fluorescent
image (Figure 4d) shows a bright varicosity with decreasing
Figure 4
.
Staining living neuroblastoma cells by gentle contact on the cell membrane. (a) Diagram showing the staining procedure
by gentle contact (not to scale). The hollow tip is maintained in contact with the cell membrane thanks to the force feedback (set
point of less than 1 nN). The active agents dissolved in the solution of the microchannel spontaneously diffuse across the membrane
into the cytoplasm. (b) Superposition of a differential interference contrast image and of the corresponding fluorescent one of a cell
after staining with CellTracker green. The microchannelled cantilever filled with CellTracker green is positioned over the cell in the
red circle using the optical microscope. The tip is then brought into gentle contact with the cell membrane by the AFM force feedback
system and left there for 15 min, before taking the fluorescent and phase contrast image. (c) Optical image in differential interference
contrast method of two neuroblastoma cells connected by a neurite forming a varicosity-like structure (red arrow) and corresponding
AFM profile along the white dotted line (AM mode, amplitude of 50 nm, scan rate of 0.1 s, cantilever as in Supporting Information,
Figure S2a). The black arrow indicates the point on the profile where the tip was precisely approached with the AFM control. (d) Corresponding
fluorescence image after staining the varicosity-like structure with acridine orange for 1 min.
Nano Lett., Vol. 9, No. 6, 2009 2505
fluorescence intensity along the neurite. This demonstrates
ease with which solutes can be delivered in a controlled
manner into selected subcellular structures with the fluidFM.
Introduction of material into a cell in an AFM-based
procedure has been achieved by other authors.15-17 In their
approaches, the AFM nanoneedle was coated or function-
alized with selected molecules and then pushed through the
cell membrane into the cytoplasm where the molecules were
released. However, this approach is limited to vanishingly
small quantities of those molecules that can be reversibly
bound to an AFM tip and then spontaneously desorb from
the tip to the intracellular environment. On the contrary, the
fluidic channel of the fluidFM represents an infinite reservoir
for the active agent moreover in a form that is immediately
available for repeated delivery as it is not bound to the tip.
Since the biomolecules to be introduced into the cell
no longer have to be grafted or adsorbed to the AFM tip
but are simply dissolved in a delivery solution, a much
wider range of applications of intracellular injection
becomes available than with the methods of Chen,
Cuerrier, or Han. This range now spans from the insertion
of particles to study how they move inside the cell
according to their chemical functionalization, to the
insertion of proteins, genes (eventually into the nucleus),
enzymes, ligands for fundamental research of cell biology
but also for applications like tissue engineering from single
ad hoc infected cells. In addition, the gentle contact
approach opens the door for studies such as the behavior
of viruses at the cell membrane or the interaction of drugs
with transmembrane proteins.
In conclusion, we have shown novel applications using
hollow force-controlled AFM cantilevers. The microchan-
neled cantilevers were successfully connected to a delivery
system via a modified AFM probeholder thus enabling
force-controlled dispensing of a solution containing
selected molecules into individual cells in a physiological
environment. The cantilever geometry and the AFM force
feedback allow very local as well as facile delivery of
molecular species to cells either by gentle contact with
their membrane or by perforation of the latter.
The experiments described here demonstrate the po-
tential of the fluidFM in the field of single-cell biology
through precise stimulation of selected cell domains with
whatever soluble agents at a well-defined time. We
confidently expect that the inclusion of an electrode in
the microfluidics circuit will allow a similar approach
toward patch-clamping with force controlled gigaseal
formation. Other strategies at the single-cell level, such
as the controlled perforation of the cell membrane for local
extraction of cytoplasm will also be explored.
Finally, the universality and versatility of the fluidFM
opens the way for original experiments in the following
wide spectrum of disciplines: in physics (e.g., droplet
interactions), materials sciences (e.g., submicrometer
etching), chemistry (e.g., local functionalization of poly-
mer layers), and molecular electronics (e.g., deposition
of conductive polymers onto nanoelectrodes).
Acknowledgment. The development of the fluidFM was
supported by the Swiss innovation promotion agency CTI-
KTI through a “feasibility project” (Contract No. 9191.1
PFNM-NM). The partial support of the Swiss Federal
Office for Education and Science (OFES) in the framework
of the EC funded project NaPa (Contract No. NMP4-CT-
2003-500120) for the development and microfabrication
of the hollow cantilevers is gratefully acknowledged. We
are indebted to Paul Lu¨thi and Stephen Wheeler (ETH
LBB workshop) for their technical help and to Daniel
Schaffhauser (ETH LBB) for the simulations of the flow
rates.
Supporting Information Available: Description of Ma-
terials and Methods. Figure S1: Micrographs of the
microchanneled cantilevers. Figure S2: Resonance peaks
of a microchanneled cantilever in different environments.
Figure 3S: Water flow rates through the microchanneled
cantilevers for different values of the aperture diameter
at the tip apex. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Binnig, G.; Quate, C. F.; Gerber, C. Phys. ReV. Lett. 1986,56, 930–
933.
(2) Giessibl, F. J. Science 1995,267, 68–71.
(3) Barth, C.; Reichling, M. Nature 2001,414, 54–57.
(4) Sugimoto, Y.; Abe, M.; Hirayama, S.; Oyabu, N.; Custance, O.; Morita,
S. Nat. Mater. 2005,4, 156–159.
(5) Thalhammer, S.; Stark, R. W.; Muller, S.; Wienberg, J.; Heckl, W. M.
J. Struct. Biol. 1997,119, 232–237.
(6) Piner, R. D.; Zhu, J.; Xu, F.; Hong, S. H.; Mirkin, C. A. Science 1999,
283, 661–663.
(7) Henderson, E.; Haydon, P. G.; Sakaguchi, D. S. Science 1992,257,
1944–1946.
(8) Simon, A.; Cohen-Bouhacina, T.; Aime, J. P.; Porte, M. C.; Amedee,
J.; Baquey, C. Cell. Mol. Biol. 2004,50, 255–266.
(9) Matzke, R.; Jacobson, K.; Radmacher, M. Nat. Cell Biol. 2001,3,
607–610.
(10) Pelling, A. E.; Sehati, S.; Gralla, E. B.; Valentine, J. S.; Gimzewski,
J. K. Science 2004,305, 1147–1150.
(11) Zhang, P. C.; Keleshian, A. M.; Sachs, F. Nature 2001,413, 428–
432.
(12) Benoit, M.; Gabriel, D.; Gerisch, G.; Gaub, H. E. Nat. Cell Biol. 2000,
2, 313–317.
(13) Helenius, J.; Heisenberg, C. P.; Gaub, H. E.; Muller, D. J. J. Cell Sci.
2008,121, 1785–1791.
(14) Cross, S. E.; Jin, Y. S.; Rao, J.; Gimzewski, J. K. Nat. Nanotechnol.
2007,2, 780–783.
(15) Chen, X.; Kis, A.; Zettl, A.; Bertozzi, C. R. Proc. Natl. Acad. Sci.
U.S.A. 2007,104, 8218–8222.
(16) Cuerrier, C. M.; Lebel, R.; Grandbois, M. Biochem. Biophys. Res.
Commun. 2007,355, 632–636.
(17) Han, S.-W. Nanomedicine 2008,4, 215–225.
(18) Whitesides, G. M. Nature 2006,442, 368–373.
(19) Davis, B. R.; Yannariello-Brown, J.; Prokopishyn, N. L.; Luo, Z. J.;
Smith, M. R.; Wang, J.; Carsrud, N. D. V.; Brown, D. B. Blood 2000,
95, 437–444.
(20) Stephens, D. J.; Pepperkok, R. Proc. Natl. Acad. Sci. U.S.A. 2001,
98, 4295–4298.
(21) Laforge, F. O.; Carpino, J.; Rotenberg, S. A.; Mirkin, M. V. Proc.
Natl. Acad. Sci. U.S.A. 2007,104, 11895–11900.
(22) Lu, Z.; Chen, P. C. Y.; Nam, J.; Ge, R. W.; Lin, W. J. Micromech.
Microeng. 2007,17, 314–321.
(23) Schrlau, M. G.; Falls, E. M.; Ziober, B. L.; Bau, H. H. Nanotechnology
2008,19, 015101.
(24) Hug, T. S.; Biss, T.; de Rooij, N. F.; Staufer, Q. Dig. Tech. Pap.
Transducers ′05, 2005,2, 1191–1194.
(25) Meister, A.; Przybylska, J.; Niedermann, P.; Santschi, C.; Heinzelmann,
H. NSTI-Nanotech 2008 2008,3, 273–276.
2506 Nano Lett., Vol. 9, No. 6, 2009
(26) Deladi, S.; Tas, N. R.; Berenschot, J. W.; Krijnen, G. J. M.; de Boer,
M. J.; de Boer, J. H.; Peter, M.; Elwenspoek, M. C. Appl. Phys. Lett.
2004,85, 5361–5363.
(27) Deladi, S.; Berenschot, J. W.; Tas, N. R.; Krijnen, G. J. M.; de Boer,
J. H.; de Boer, M. J.; Elwenspoek, M. C. J. Micromech. Microeng.
2005,15, 528–534.
(28) Lewis, A.; Kheifetz, Y.; Shambrodt, E.; Radko, A.; Khatchatryan, E.;
Sukenik, C. Appl. Phys. Lett. 1999,75, 2689–2691.
(29) Hong, M. H.; Kim, K. H.; Bae, J.; Jhe, W. Appl. Phys. Lett. 2000,77,
2604–2606.
(30) Taha, H.; Marks, R. S.; Gheber, L. A.; Rousso, I.; Newman, J.;
Sukenik, C.; Lewis, A. Appl. Phys. Lett. 2003,83, 1041–1043.
(31) Hategan, A.; Law, R.; Kahn, S.; Discher, D. E. Biophys. J. 2003,85,
2746–2759.
(32) The membrane-permeable CellTracker and acridin orange were used
in combination with ethidium bromide to differentiate between living
and apoptotic cells.
NL901384X
Nano Lett., Vol. 9, No. 6, 2009 2507
