Neurochip technology researchers developed neurons on silicon microchips for the first time

Neurochip technology developed by Canadian team

Naweed Syed’s lab cultivated brain cells on a microchip.The University of Calgary, Faculty of Medicine scientists who proved it is possible to cultivate a network of brain cells that reconnect on a silicon chip—or the brain on a microchip—have developed new technology that monitors brain cell activity at a resolution never achieved before.

Developed with the National Research Council Canada (NRC), the new silicon chips are also simpler to use, which will help future understanding of how brain cells work under normal conditions and permit drug discoveries for a variety of neurodegenerative diseases, such as Alzheimer’s and Parkinson’s.

The new technology from the lab of Naweed Syed, in collaboration with the NRC, is published online this month in the journal, Biomedical Devices.

“This technical breakthrough means we can track subtle changes in brain activity at the level of ion channels and synaptic potentials, which are also the most suitable target sites for drug development in neurodegenerative diseases and neuropsychological disorders,” says Syed, professor and head of the Department of Cell Biology and Anatomy, member of the Hotchkiss Brain Institute and advisor to the Vice President Research on Biomedical Engineering Initiative of the

U of C.

The new neurochips are also automated, meaning that anyone can learn to place individual brain cells on them. Previously it took years of training to learn how to record ion channel activity from brain cells, and it was only possible to monitor one or two cells simultaneously. Now, larger networks of cells can be placed on a chip and observed in minute detail, allowing the analysis of several brain cells networking and performing automatic, large-scale drug screening for various brain dysfunctions.

This new technology has the potential to help scientists in a variety of fields and on a variety of research projects. Gerald Zamponi, professor and head of the Department of Physiology and Pharmacology, and member of the Hotchkiss Brain Institute, says, “This technology can likely be scaled up such that it will become a novel tool for medium throughput drug screening, in addition to its usefulness for basic biomedical research”.

The U of C is excited at the potential of this made in Canada technology.

“The University of Calgary is proud to be the home of this cutting edge Canadian work with a neurochip. The advances in research and healthcare made by possible by this technology are immense. The work and collaboration happening in the lab of Naweed Syed is another example demonstrating our leadership in the field of biomedical engineering,” says Rose Goldstein the University of Calgary’s vice-president of research.

High-fidelity patch-clamp recordings from neurons
cultured on a polymer microchip
Dolores Martinez & Christophe Py & Mike W. Denhoff &
Marzia Martina & Robert Monette & Tanya Comas &
Collin Luk & Naweed Syed & Geoff Mealing
# Her Majesty the Queen in Right of Canada 2010
Abstract We present a polymer microchip capable of monitoring
neuronal activity with a fidelity never before obtained on
a planar patch-clamp device. Cardio-respiratory neurons Left
Pedal Dorsal 1 (LPeD1) frommollusc Lymnaea were cultured
on the microchip’s polyimide surface for 2 to 4 hours. Cultured
neurons formed high resistance seals (gigaseals) between the
cell membrane and the surface surrounding apertures etched in
the polyimide. Gigaseal formation was observed without
applying external force, such as suction, on neurons. The
formation of gigaseals, as well as the low access resistance and
shunt capacitance values of the polymer microchip resulted in
high-fidelity recordings. On-chip culture of neurons permitted,
for the first time on a polymeric patch-clamp device, the
recording of high fidelity physiological action potentials.
Microfabrication of the hybrid poly(dimethylsiloxane)—polyimide
(PDMS-PI) microchip is discussed, including a two-layer
PDMS processing technique resulting in minimized shrinking
Keywords Planar patch-clamp . Microfluidic . Neurons .
Poly(dimethylsiloxane) . Polyimide . Action potential
PDMS poly(dimethylsiloxane)
PI polyimide
LPeD1 Left Pedal Dorsal 1
R-C access resistance – shunt capacitance
SEM scanning electron microscope
AFM atomic force microscope
1 Introduction
Ion channels play an important role in virtually all aspects
of cellular physiology. Ion-channel dysfunctions have been
linked to a wide spectrum of pathophysiologies, including
neurodegenerative diseases. Consequently, their function
and regulation have been the focus of an enormous amount
of research over the last several decades and therapeutic
strategies targeting modulation of ion-channel activity has
become the focus of many drug discovery programs
(Dunlop et al. 2008; Willumsen et al. 2003).
The gold standard for ion-channel monitoring is conventional
patch-clamp, whereby a glass pipette tip filled
with electrochemically conductive physiological saline
solution is positioned over a cell. Following the application
of suction to the pipette, a high resistance seal is formed
with the cell membrane, thereby enabling quality ion-flux
recordings (Hamill et al. 1981). This technique offers both
quality and versatility: different cell types can be studied;
cells can be dissociated in suspension or, as is the case for
cell cultures or tissue slices, allowed to form connections
with other cells. Several configurations are possible,
including cell-attached, isolated patch (inside- or outsideout)
and whole-cell. Conventional patch-clamp is, however,
a laborious process requiring highly skilled personnel. This
unfortunate drawback makes the technique inappropriate
D. Martinez (*) : C. Py : M. W. Denhoff
Institute for Microstructural Sciences,
National Research Council of Canada,
1200 Montreal Road,
Ottawa, Ontario K1A0R6, Canada
M. Martina : R. Monette : T. Comas : G. Mealing
Institute for Biological Sciences,
National Research Council of Canada,
1200 Montreal Road,
Ottawa, Ontario K1A0R6, Canada
C. Luk : N. Syed
Hotchkiss Brain Institute, University of Calgary,
3330 Hospital Dr. N.W.,
Calgary, Alberta T2N 4N1, Canada
Biomed Microdevices
DOI 10.1007/s10544-010-9452-z
for high throughput drug screening and presents a bottleneck
in the drug discovery process.
In an effort to increase throughput, planar patch-clamp
systems were introduced; the glass pipette was replaced by a
membrane having microfeatures (apertures) mimicking the tip
of the pipette. Several microchip membrane types have been
studied: silicon dioxide (Hediger et al. 1999; Lehnert et al.
2002; Stett et al. 2003a; Pantoja et al. 2004; Sordel et al.
2006), silicon nitride (Schmidt et al. 2000; Fertig et al. 2000;
Mealing et al. 2005; Py et al. 2008), quartz/glass (Fertig et
al. 2002; Ong et al. 2007; Chen et al. 2009), poly
(dimethylsiloxane) (Klemic et al. 2005; Dahan et al. 2008;
Chen and Folch 2006; Lau et al. 2006) and polyimide (Kiss
et al. 2003; Stett et al. 2003b). High throughput analysis
systems have been commercialized (Schroeder et al. 2003;
Bruggemann et al. 2003; Kutchinsky et al. 2003; Dubin et al.
2005). In all planar patch-clamp systems to date, experiments
are conducted on cell lines, most commonly Chinese Hamster
Ovary (CHO) and Human Embryonic Kidney (HEK), having
specific ion channels over-expressed in their membrane by
genetic manipulation. Dissociated cells in suspension are
deposited on the planar patch-clamp recording device; individual
cells are then driven to the aperture on the chip membrane
by either suction or electrokinetic means. After the detection of
a high impedance seal (also called gigaseal), ion-channel
activity is monitored in the cell-attached or whole-cell
A more physiologically relevant approach to understanding
ion channel function requires interrogation of cells in their
native environment: the use of primary cells, rather than cell
lines, enables such studies. The ability to record from cultured
cells is also an important consideration, since extended studies
or studies involving cell interactions are often required. For
planar patch-clamp devices to be considered as an improved
alternative to the traditional pipette patch-clamp technique,
application to such relevant biology must be demonstrated.
Another impediment to the use of planar patch-clamp
devices is their low signal recording quality. The main
contributor to low signal-to-noise ratios in these devices is the
leaky seal obtained between the cell membrane and the chip
surface’s aperture. High resistance seal (gigaseal) formation
significantly reduces background current noise and permits the
recording of quality data (Hamill et al. 1981); gigaseal
formation is therefore an important parameter in patch-clamp
studies. Although a growing body of evidence points to
several parameters playing a role in gigaseal formation, its
mechanism is still poorly understood (Sordel et al. 2006; Chen
et al. 2009). Also contributing to poor signal quality is the
high R-C (access resistance—shunt capacitance) values of
some planar patch-clamp devices which reduce the chip’s
response time (Sordel et al. 2006). Silicon-based devices have
higher C due to their conductive bulk while glass-based
devices, whether in the vertical (Fertig et al. 2002) or lateral
(Ong et al. 2007) planar clamp modes, have shown high R.
Polymer-based planar patch-clamps have shown good R-C
characteristics, but the difficulty of forming a gigaseal on
these devices has resulted in low quality recordings.
This report describes a low R-C microchip for high quality
signal recording of individual neurons cultured directly on the
chip’s surface. Neurons were isolated from mollusc Lymnaea
and cultured for 2 to 4 hours over apertures to form gigaseals
(>5 GΩ) without the application of suction. We report highfidelity
recordings of action potentials from cultured neurons
on a polymer planar patch-clamp device.
The relative success of current planar patch-clamp devices
rests on the use of dissociated cells in suspension. Synaptic
connectivity, key to modeling and understanding neuronal
function, can therefore not be investigated using such devices.
This exemplifies the limits to physiological relevancy of
cellular systems investigated on current planar patch-clamp
systems. The microchip presented here contains two closelyspaced
apertures, each individually accessible through dedicated
underlying microchannels. This chip design has the
potential to permit simultaneous monitoring of individual
neurons engaged in synaptic connectivity. This report takes a
major step towards realizing this vision: the fabrication of a
low R-C hybrid poly(dimethylsiloxane)—polyimide microchip
and the first demonstration of high-quality, physiologically
relevant patch-clamp recordings from individual
neurons cultured on-chip. Studies of synaptic connectivity
between neurons cultured on the polymer microchip remains a
challenge to be addressed in future studies.
2 Polymer microchip
2.1 Design
Figure 1 presents the polymer-based microchip. Figure 1(a)
shows a schematic cross-section of the hybrid poly
(dimethylsiloxane)—polyimide (PDMS—PI) chip mounted
in a machined Plexiglas package. Polyimide was chosen
because of its high chemical and mechanical stability, its
low dielectric constant and the fact that polyimide microprocessing
techniques are standard in the microelectronics
industry. The cell is grown on-chip at the aperture location
in the culture chamber and accessed through an underlying
microchannel in the PDMS layer. Figure 1(b) presents a
micrograph of the chip: the top PI membrane (3 μm thick)
contains two apertures (4 μm diameter) for patch-clamping.
The apertures are closely-spaced (110 μm apart) and
individually accessible through the bottom PDMS microchannels
(each 200 μm wide and 10 μm deep). Four via
holes, 150–200 μm in diameter, are punched through the
PDMS at the extremities of the microchannels allowing
perfusion of chemicals in the underlying fluidics. Pillars
Biomed Microdevices
(20×20 μm2) are integrated in the microchannels as support
for the thin PI film. The chip is packaged in a Plexiglas
culture chamber with aligned inlets and outlets, as shown in
Fig. 1(c). Via holes in the PDMS are aligned with machined
holes in the Plexiglas to allow for perfusion through
silicone tubing. The tubing also houses Ag/AgCl electrodes
during patch-clamp experiments.
A patch-clamp recording’s signal-to-noise ratio and time
response following a voltage (or current) pulse depend on the
microchip’s R-C characteristic: low chip capacitance will
improve noise and a low R-C constant allow for recording of
fast kinetic events. The chip’s shunt capacitance is proportional
to: the area of self-supported membrane, i.e. where the
culture medium and solution in the underlying fluidics are
superimposed across the PI membrane; the membrane’s
dielectric constant and; the inverse of the membrane’s
thickness. To minimize capacitance, the choice of a polymer
membrane with a low dielectric constant, 3.4 in the case of PI,
is obviously advantageous. A thick PI membrane and the
integration of microfluidic channels in the planar patch-clamp
system are the other two means of reducing capacitance
values. The dimensions described above result in 7.7 pF shunt
capacitance as measured at 100 kHz with channels filled with
physiological saline. The access resistance of the aperture,
however, is directly proportional to the PI membrane’s
thickness and the access resistance of the microfluidic channel
is inversely proportional to its section. Although the membrane’s
thickness and microchannels’ width (through area of
self-supported surface) are tied to overall capacitance values,
the access resistance can be lowered by increasing the depth of
the microfluidic channels. An access resistance of 1.43 MΩ
was calculated assuming a 52 Ω.cm resistivity solution
(150mMphosphate buffered saline solution). For the aperture
itself, assuming a cylinder of 3.3 μm height and 4.1 μm
diameter, the resistance is 0.13 MΩ. While the spreading
resistance (Rs), due to the current spreading out from the
aperture into both the top and bottom fluid reservoirs, is
found from (Denhoff 2006):
Rs ¼ r
where ρ is the solution resistivity and d, the aperture diameter.
In our case, Rs is 0.13 MΩ, for a total aperture resistance of
0.26 MΩ. Finally, the resistance of the microfluidic channels
(10 μm thickness, 200 μm width and 3 mm length) is
1.17 MΩ, taking the pillars into account, for an overall
calculated access resistance of 1.43 MΩ. The access
resistance, using Ag/AgCl electrodes in phosphate buffer
saline solution, was measured to be 1.6 MΩ, in good
agreement with the calculated value. The measured capacitance
is comparable to that of a glass pipette and the measured
access resistance, two to three times smaller than typical glass
pipette values, making the microchip capable of high fidelity
patch-clamp recording (Sigworth and Klemic 2005).
2.2 Fabrication
Figure 2 presents a schematic of the polymer microchip
fabrication: the two layers forming the chip, PDMS and PI,
were processed independently starting from 2 inch Si (PDMS,
steps 2a to 2d) and Si/SiO2 (PI, steps 1a to 1c) wafers and
bonded with alignment (steps 3a and 3b), producing nine 1×
Culture chamber polyimide
to output 1
input 1
to output 2
input 2
output 1
input 1
output 2
Input 2
50 μm
Fig. 1 Polymer microchip for high quality patch-clamp recordings.
(a) Schematic of the polymer microchip: a polyimide membrane
containing an aperture sits on top of a poly(dimethylsiloxane)
microchannels. The chip is mounted in a Plexiglas package having
an upper 2 mm opening for cell culture. Neurons are grown on-chip
over the aperture. A gigaseal is formed between the cell and the
apertures’ outer rim, allowing for high quality patch-clamp recordings
of neural activity. (b) 20× micrograph of the polymer microchip,
showing two apertures distant by 110 μm, each opening to a 200 μm
underlying PDMS microchannels (with 20 μm square support pillars.
(c) Microchip in its Plexiglass package
Biomed Microdevices
1 cm2 chips. Si and Si/SiO2 wafers were pre-cleaned in
acetone, isopropyl alcohol (80°C) and hydrofluoric acid (1%
in water, all CMOS grade from J.T. Baker, Phillipsburg, NJ)
successively, rinsed in deionized water and dried for
5 minutes on a hot plate at 110°C.
PDMS microchannels were fabricated by replicamolding
from a SU8-on-Si master mold. SU8 10 photosensitive
resin (Microchem Corp., Newton, MA) was spun
at 3,000 rpm for 30 s on a pre-cleaned Si wafer, for a final
SU-8 master thickness of approximately 10 μm. A pre-bake
step (2 minutes at 65°C and 5 minutes at 95°C) was
followed by slow cooling to room temperature over 1 hour
to avoid stress release. After a 15 s UV exposure (10 mW.
cm−2), the master was post-baked for 1 minute at 65°C and
2 minutes at 95°C, followed by a slow cooling to room
temperature over 1 hour. The final master was obtained
after developing in SU8 developer (Microchem Corp.) for 6
minutes with agitation. Replica-molding results in the
deformation of the replicated PDMS: upon curing, the
PDMS shrinks and the original master dimensions are lost,
making wafer level assembly over several centimeters
impossible. A reproducible shrinkage can be easily compensated
by the design of the master; however, shrinkage is
not reproducible. Because it is dependant on curing
temperatures, master mold design and base-to-curing agent
ratios, shrinking variations are present within the same
replicated PDMS substrate. Lower curing temperatures and
different base-to-curing agent ratios have been successfully
tested to minimize shrinking (Lee and Lee 2008); these
however did not reduce shrinking variations. To reduce
PDMS shrinking variations, we developed the two-layer
PDMS processing technique depicted in steps 2a to 2d of
Fig. 2. The first PDMS layer was made on a blank Si wafer
by pouring ~3.3 g (10:1 curing ratio) of Sylgard 184 resin
(Dow Corning, Midland, MI), producing a thick support
structure after a 3 hour cure at 120°C. The thick PDMS
layer was then released from its mold (step 2a). For the
second layer, ~0.5 g (10:1 curing ratio) of PDMS was spin
coated on the SU8-on-silicon master at 3,700 rpm, producing
a thin structure containing the microchannels after a
5 min cure on a hot plate at 95°C (step 2b). Thin and thick
PDMS layers were covalently bonded following plasma
treatment of both surfaces (step 2c). After 5 minutes, the
PDMS was peeled from the SU8-on-Si mold (step 2d). The
thick PDMS layer acted as a support for the thin PDMS
containing the microchannels, resulting in microstructures
with minimized shrinking variations. Nine microchips were
produced per wafer: chip-to-chip distance was set to 1 cm
on the mask. Using the two-layer approach, PDMS
shrinking and shrinking variation were both reduced, to 6
±4 μm over 1 cm versus 138±13 μm over 1 cm for PDMS
cured in the typical fashion (90°C for two hours). This
improvement allowed alignment-bonding to the upper PI
layer over the full wafer. Fluidics access via holes were
punched in the PDMS at the beginning and end of both
channels using a machined syringe needle (step 2d),
producing small 150 to 200 μm diameter via holes,
minimizing capacitance as discussed in the Design section.
To prepare for bonding with the upper PI membrane, the
PDMS surface was given an air plasma treatment.
A pre-cleaned Si/SiO2 wafer was used as a template for
processing apertures in the PI. PI 2610 (HD Microsystems,
Si or Si/SiO2 Al/PR 1813 PI
1a 1b 1c
2c 2d
Fig. 2 Polymer microchip fabrication. PI and PDMS are processed
independently on two inch wafers; alignment-bonding forms the final
microchip. A 50 nm Al mask and photoresist layer are deposited on a
Si/SiO2—supported PI layer (step 1a). Lithography and wet etching of
the exposed Al define the PI to be processed (step 1b). Reactive ion
etching of the exposed PI results in the apertures (step 1c). PDMS
shrinking variations are reduced over the two inches mould by a twolayer
technique. PDMS is poured on a blank silicon wafer, cured and
released from the wafer, producing a thick support structure (step 2a).
PDMS is spin-coated on the SU8-on-Si mould at 3,700 rpm and cured
on a hot plate (step 2b). Both layers are surface-treated and covalently
bonded (step 2c); after a few minutes, the structure is released (step
2d). Following surface treatments, Si/SiO2—supported PI was
aligned-bonded to the PDMS structure (step 3a) followed by release
of the polymer chip
Biomed Microdevices
Parlin, NJ) was spun at 2,700 rpm for 30 s, for a final
thickness of approximately 3 μm. The wafer was heated on
a hot plate at 100°C for 90 s and 150°C for 90 s, followed
by a high temperature cure in inert atmosphere (350°C for
30 minutes in 200 sccm N2/H2 flow). A 50 nm aluminum
mask layer was then evaporated on the PI film (step 1a).
Lithography of the apertures followed a standard recipe for
Shipley 1813 (Rohm & Haas, Marlborough, MA) positive
resist. After a 1 minute hexamethyldisilazane (HMDS)
wafer drying cycle, Shipley 1813 was spun at 5,000 rpm for
30 s, followed by a 5 minutes rest and 110°C cure for 5
minutes (sep 1a). After a 6.25 s UV exposure (10 mW.
cm−2), both the resist and Al were patterned by the
developer solution (Microposit MF-321, Rohm & Haas),
exposing the PI to be fabricated (step 1b). Apertures were
etched in a reactive ion etcher (step 1c, plasma parameters:
35 min, 10 mTorr, 100 W, 10 sccm O2). Residual aluminum
was removed in 60°C metal etchant (E6, Fujifilm Electronics
Materials, East Providence, RI). The patterned PI was
thoroughly rinsed in deionized water and dried on a hot
plate at 110°C for 5 minutes. To prepare the PI for bonding
with the bottom PDMS layer, VM-651 solution (0.1% in
water, HD Microsystems) was spin-coated on the substrate
(3,000 rpm).
Aligned bonding of the PI and PDMS layers was done in
a M9 flip-chip bonder (step 3a, Besi, Londonderry, NH) at
2,000 g force for 30 s. Following bonding, the wafers were
oven-heated at 65°C for two hours and released from the Si/
SiO2 substrate by simply peeling the structure off (step 3b).
Figure 3 presents scanning electron microscopy (SEM)
images of the polyimide apertures, both top (Fig. 3(a)) and
side (Fig. 3(b), after focused ion beam sectioning) views.
Reactive ion etching of polyimide produced a nearly
anisotropic etch, with an aperture diameter of 3.9 μm at
the start of the etch (bottom) and 4.1 μm at its end (top).
2.3 Assembly and loading
Wafers were diced in nine 1×1 cm2 chips. Individual chips
were bonded to a lower glass support wafer and glued
(RTV silicone 3140, Dow Corning) to a machined Plexiglas
package fitted with connectors (Valco Instruments, Brockville,
ON, Canada) for easy microchannel access (see Fig. 1
(c)). After a 5 min sterilization in an air plasma cleaner
(Harrick Plasma, Ithaca, NY), microchannels were filled
with the electrophysiological recording solution which
contained in mM: 50 KCl, 5 MgCl2, 5 EGTA and 5
HEPES (pH 7.4; 130 mOsm). Publications describing
whole-cell patch-clamp recordings in Lymnaea neurons
report the use of solutions which are designed to isolate
well-defined currents such as Ca2+ (McCamphill et al.
2008). To study the electrophysiological behavior of
Lymnaea neurons we used a solution in the microchannels
which would maintain a physiological potassium gradient
(KCl 50 mM) across the cell membrane once the whole-cell
configuration was established. HEPES (5 mM) was added
to the pipette solution to buffer the pH, and EGTA (5 mM)
was added to chelate Ca2+. In addition Mg2+ (5 mM) was
included to facilitate seal formation (Priel et al. 2007). The
upper culture chamber was filled with poly-L-lysine
solution and left to stand for two hours. Following a
distilled water rinse, normal Lymnaea saline (see below)
was loaded in the culture chamber. Neurons were then
placed over the apertures and cultured for a minimum of
two hours at room temperature prior to recording. Microchips
were connected to an Axopatch 200B amplifier
(Axon Instruments, Foster City, CA) by placing recording
electrodes in the silicone tubes connected to the microfluidic
channels and a reference electrode in the culture
dish. Signals were filtered at 2 kHz and acquired at a
sampling rate of 20 kHz. Analyses were carried out off-line
with the software IGOR (Wavemetrics, Lake Oswego, OR)
and software running on a personal computer. Test voltagesteps
(5 mV) were then applied to measure the total
resistance (the sum of the cell membrane resistance and
total access resistance). All chemicals were from Sigma-
Aldrich (St Louis, MO).
Fig. 3 Scanning electron micrograph of a polyimide microhole. (a)
Top view and (b) cross section of one polyimide microhole. The tiltcorrected
values for the cross section are: 3.3 μm thickness, 3.9 μm
diameter (bottom) and 4.1 μm diameter (top)
Biomed Microdevices
3 Cell culture
Freshwater snails (Lymnaea stagnalis) were maintained in
an aquarium containing aerated pond water at room
temperature (20–22°C), and fed lettuce. Snails, 1–2 months
old (10–15 mm in length), were used to isolate neurons for
experiments, whereas 2–4 month old animals (20–30 mm in
length), were used to prepare conditioned medium (CM:
isolated snail brains incubated in defined media for 3–
4 days). Detailed cell isolation and culture procedures have
been described previously (Syed et al. 1990). In brief, snails
were deshelled and anesthetized with 10% Listerine
solution (ethanol, 21.9%; and methanol, 0.042%) in normal
Lymnaea saline (in mM: 51.3 NaCl, 1.7 KCl, 4.0 CaCl2,
and 1.5 MgCl2) buffered to pH 7.9 with HEPES. The
central ring ganglia (CRG) were then removed and washed
three times (10 min each) with normal saline containing
gentamicin (50 μg.mL−1). CRG were then incubated for
20 min in defined medium (DM: Serum-free 50% L-15
medium with 20 μg.mL−1 gentamicin and inorganic salts at
mM: 40 NaCl, 1.7 KCl, 4.1 CaCl2, 1.5 MgCl2, and 10
HEPES, pH 7.9) containing trypsin (2 mg.mL−1; Sigma, St
Louis, MO), followed by a 15 min wash with DM
containing trypsin inhibitor (2 mg.mL−1; Sigma). After
enzymatic treatment, the CRG were pinned down in a
Sylgard dissection dish containing high osmolarity DM
(750 μl of 1 M glucose was added to 20 ml DM to raise
osmolarity from 130–145 to 180–195 mOsm). The outer
and inner sheath layers were removed using fine forceps
and the Left Pedal Dorsal 1 (LPeD1) interneuron was
extracted using gentle suction applied through a Hamilton
syringe (Hamilton Co., Reno, NV) and a fire-polished,
Sigmacoated (Sigma) glass pipette (60–70 μM tip diameter).
Under a dissecting microscope, isolated neurons were
manipulated over apertures on the microchip using a
suction pipette and then released by applying positive
4 Electrophysiological results and discussion
A total of 8 polymer microchips were tested. Of these, 4
microchips had neurons cultured over one aperture (Fig. 4
(a)) and 4 chips had neurons over both apertures (Fig. 4(b))
for a total of 12 neurons tested.
In response to the 5 mV test voltage-steps, 2 neurons
showed seal resistances of 5.5 and 8.2 GΩ, typical of cellattached
configurations (Fig. 5), and 4 neurons showed
large capacitive transients suggesting attainment of the
whole-cell patch-clamp configuration (Fig. 5). At this stage,
no negative pressure was applied to the microfluidic
channels to form a high resistance seal. The remaining 6
neurons tested showed a resistance of 4.03±0.6 MΩ and no
large capacitive transient. No changes in the resistance and
capacitance were observed after the application of suction.
We concluded that these neurons did not form adequate cell
membrane-to-chip seal resistance to permit recording of
biological activity.
In the 2 neurons in the cell-attached configuration no
single channel activity was present, probably due to the
absence of ion channels on the localized patch of
membrane (data not shown). This may have been due to a
low density of ion channels populating the membrane over
the microhole. This is consistent with reports of low ion
channel density on neuronal cell bodies isolated from the
snail Helix pomatia, where calcium channel density was 1
per 3 μm2 (Lux and Brown 1984), or in the soma and the
growth cones of Lymnaea neurons where the density of
stretch-activated channels selective for K+ was 1 per 1–
2 μm2 (Sigurdson and Morris 1989). These values are very
low when compared to the Na+ channel density in the squid
axon (166–533 per μm2; Hille 1984) or in squid stellate
ganglion cell bodies (37 per μm2; Brismar and Gilly 1987).
From these background recordings, we calculated the
standard deviation of the baseline current to be 0.7 pA:
the minimum signal distinguishable from the baseline
fluctuation at 95% confidence level is then 2.1 pA, within
Fig. 4 Micrograph of LPeD1 neurons on polymer microchips.
Micrograph for the (a) one neuron—one chip and (b) two neurons—
one chip experiments. For the single channel recordings of (a), the
lower microchannel was left dry. No process developed between the
two neurons in (b). Each neuron covered one aperture; see Fig. 1(b)
for microchip outline
Biomed Microdevices
the resolution of single ion channel current jumps.
Although we recorded no such activity in the cell-attached
configuration, rupturing the patch of cell membrane over
the aperture with a brief suction pulse to the microfluidic
channel resulted in the characteristic appearance of large
capacitive transients. These, as shown in Fig. 5, are
associated with the increased surface of the membrane,
indicative of attaining the whole-cell configuration.
We subsequently conducted current-clamp and voltageclamp
recordings on the 6 neurons in which the whole-cell
patch-clamp configuration was attained. Typical whole-cell
voltage and current responses to an incremental series of
intracellular current or voltage steps are shown in Fig. 6(a)
and (b), respectively. The observation of robust action
potentials in response to depolarizing current steps confirms
the acquisition of electrophysiological activity characteristic
of Lymnaea neurons (Fig. 6(a)), making this the first
recordings of action potential obtained from neurons
cultured directly on a planar microchip. We obtained
whole-cell recordings over prolonged periods of time: one
neuron was re-tested 1 h 40 min after the initial analysis
and found to behave identically. Other neurons were tested
only once, over a period of 20 to 40 minutes.
In conventional whole-cell patch-clamp experiments, the
current through the series resistance of the pipette and the
residual resistance of the ruptured patch is often sufficiently
large to introduce significant voltage errors. Ideally, the
resistance of the patch pipette in whole-cell experiments should
be zero and the time resolution for measuring membrane
currents and changing the membrane voltage would be limited
only by the speed of the electronics (typically a few microseconds).
Modern amplifiers use different techniques for
compensating these errors, i.e. “correction” and “prediction”
(Axon Instruments 1993). However, series resistance compensation
is never perfect. Even if 100% compensation is used
with stability, very fast currents cannot be fully corrected.
In the whole-cell recordings reported in Fig. 6, series
resistance was not compensated. The current-clamp recordings
in Fig. 6(a) show virtually no voltage offset in
response to current steps, confirming that series resistance
compensation was not required to record action potentials.
More precisely, in current-clamp mode the bridge balance is
used to separate the membrane potential (Vm) from the total
potential (Vp) recorded by the micropipette. The axon
guide: a guide to electrophysiology and biophysics laboratory
techniques (Axon Instruments 1993) states that “when
the current is stepped to a new value, there is a rapid
voltage step on Vp due to the ohmic voltage drop across the
micropipette (voltage offset)”. Since the “micropipette” is
intracellular, changes in Vm are included in Vp. Thus the Vp
trace shows an exponential rise to a new potential followed
by some membrane potential activity. The bridge amplifier
removes the instantaneous voltage step, leaving the Vm
trace shown.
To further assess the fidelity of our recordings, we
calculated the rise-time of currents evoked during the rising
phase of the action potential in voltage-clamp (Fig. 6(b) and
inset). Typically such currents are due to a Na+ conductance.
We found that these currents had a rise time (10–
90%) of 0.84±0.02 ms. Lymnaea neurons have previously
been reported to have Na+ conductance channels (Kostyuk
and Krishtal 1977) and such invertebrate Na+ channels have
been shown, using conventional patch-clamp, to have
activation kinetics ranging from 0.99 ms to 2.72 ms (Gilly
et al. 1997). This suggests that we are capturing the highest
frequencies of electrophysiological activity in these neurons,
showing the ability of our polymer microchip to
record patch-clamp signals with high fidelity. Results of
Fig. 6(a) show textbook-quality action potentials obtained
under ideal conditions of high seal resistance and good R-C
chip characteristics.
Gigaseal formation between the cell membrane and
aperture’s outer rim is a prerequisite to high signal-tonoise
patch-clamp recordings, therefore gigaseal formation
is a parameter used in evaluating the efficacy of planar
patch-clamp methods. Gigaseal formation is poorly understood,
but published studies on planar patch-clamp systems
have suggested the importance of several parameters. The
surface making contact with the cell membrane should be
smooth, clean and hydrophilic; it should have a high area of
Fig. 5 Typical responses to 5 mV voltage-steps obtained in cell-attached
and whole-cell configuration. Note the capacitive transients due to the
increased cell membrane area in the whole-cell configuration
Biomed Microdevices
contact with the cell membrane, thus favoring an hourglass
shape for the aperture (Sordel et al. 2006; Chen et al. 2009;
Lehnert et al. 2007). The type of cells, their passage
number and minimal debris in the cell suspension also
play a role. These parameters are relevant to seal
formation based upon drawing the cell membrane into
close proximity with the chip surface surrounding the
aperture using suction. The approach we used in this
study was different: we placed cells over the aperture
and allowed them to establish a gigaseal without further
manipulation. The small number of microchips tested in
the present study precluded in-depth investigation of
this complex issue.
Of the 12 neurons cultured on chips, 6 initially formed
high resistance seals between the cell membrane and the
chip surface. This rate of success compares well with other
polymer planar patch-clamp systems (Stett et al. 2003b;
Ionescu-Zanetti et al. 2005). Plasma treatment of the
polyimide surface followed by poly-L-lysine functionalization
rendered the surface hydrophilic, thus favoring gigaseal
formation. Also, the polymide surface around the
aperture is smooth: a 1 nm root-mean-square surface
roughness was recorded by atomic force microscopy
(AFM, not shown), in-line with published roughness data
for efficient cell-to-surface seals (Lehnert et al. 2007). From
the SEM cross-section shown in Fig. 3(b), the aperture’s
edges are rounded and have a lip. Although not as
pronounced as in the micromachined hourglass shape
(Chen et al. 2009), this lip may enhance contact area
between the cell membrane and polyimide surface, thus
improving seal formation and allowing the recording of
high-quality signals.
5 Conclusion
We describe a polymer microchip for high quality planar
patch-clamp recording of primary neurons cultured on-chip.
The combination of high impedance seal formation between
the cell membrane and the chip, and the low R-C
characteristics of the hybrid PDMS-PI microchip permitted
high-fidelity recording of neuronal electrophysiological
activity. We report the first recording of action potentials
from neurons cultured on a polymer planar patch-clamp
device. Having established the functionality of this microchip
for patch-clamp recordings from neurons cultured onchip,
future work will focus on simultaneous recordings
from neurons that are synaptically connected (Smit et al.
2001; Woodin et al. 2002; Munno et al. 2003). This will
allow us to examine neuronal communication and patterns
of synaptic plasticity with a resolution not yet attained on a
planar patch-clamp device.
Acknowledgments The authors wish to gratefully acknowledge the
technical support of Allan Horner and Robert Trepanier for Plexiglas
packages fabrication, Hue Tran for metallization, Jeff Fraser for
scanning electron microscopy imaging and Raluca Movileanu for
AFM data. Naweed Syed was supported by the Canadian Institute
for Health Research (CIHR) grant. Collin Luk is the recipient of
NSERC and Alberta Heritage Foundation for Medical Research
(AHFMR) studentships.
Fig. 6 Whole-cell patch clamp recordings from an LPeD1 neuron
cultured on-chip. (a) Voltage responses (top) of an LPeD1 neuron to
graded series of intracellular current pulses (bottom), after the rupture
of the patch membrane and the attainment of the whole-cell
configuration. The two upper curves were shifted up for clarity. The
current pulses were applied at a voltage membrane potential of
−60 mV (Ic=−45 pA). (b) Current responses (top) of the same LPeD1
neuron shown in A to a series of intracellular voltage pulses (10 mV.
pulse−1) from a voltage membrane potential of −60 mV. The inset
shows a region with an expanded time scale
Biomed Microdevices
Axon Instruments, in The axon guide: a guide to electrophysiology
and biophysics laboratory techniques (1993), p. 262.
A. Bruggemann, M. George, M. Klau, M. Beckler, J. Steindl, J.C.
Behrends, N. Fertig, Assay Drug Dev. Technol. 1, 665 (2003)
T. Brismar, W.F. Gilly, Proc. Natl Acad. Sci. 84, 1459 (1987)
C.-H. Chen, A. Folch, Lab Chip 6, 1338 (2006)
C.-Y. Chen, T.-Y. Tu, C.-H. Chen, D.-S. Jong, A.M. Wo, Lab Chip 9,
2370 (2009)
E. Dahan, V. Bize, T. Lehnert, J.D. Horisberger, M.A.M. Gijs, Lab
Chip 8, 1809 (2008)
M.W. Denhoff, J. Phys. D Appl. Phys. 39, 1761 (2006)
A.E. Dubin, N. Nasser, J. Rohrbacher, A.N. Hermans, R. Marrannes,
C. Grantham, K. Van Rossem, M. Cik, S.R. Chaplan, D.
Gallacher, J. Xu, A. Guia, N.G. Byrne, C. Mathes, J. Biomol.
Screen. 10, 168 (2005)
J. Dunlop, M. Bowlby, R. Peri, D. Vasilyev, R. Arias, Nat. Rev. Drug
Discov. 7, 358 (2008)
N. Fertig, R.H. Blick, J.C. Behrends, Biophys. J. 82, 3056 (2002)
N. Fertig, A. Tilke, R.H. Blick, J.P. Kotthaus, J.C. Behrends, G. ten
Bruggencate, Appl. Phys. Lett. 77, 1218 (2000)
W.F. Gilly, R. Gillette, M. McFarlane, J. Neurophysiol. 77, 2373
O.P. Hamill, A. Marty, E. Neher, B. Sakmann, F.J. Sigworth, Pflügers
Arch. 391, 85 (1981)
S. Hediger, A. Sayah, M.A.M. Gijs, Sens. Actuators B 56, 175 (1999)
B. Hille, Ionic Channels of Excitable Membranes (Sinauer Associates,
Sunderland, 1984)
C. Ionescu-Zanetti, R.M. Shaw, J. Seo, Y.-N. Jan, L.Y. Jan, L.P. Lee,
Proc. Natl Acad. Sci. 102, 9112 (2005)
L. Kiss, P.B. Bennett, V.N. Uebele, K.S. Koblan, S.A. Kane, B.
Neagle, K. Schroeder, Assay Drug Dev. Technol. 1, 127 (2003)
K.G. Klemic, J.F. Klemic, F.J. Sigworth, Pflügers Arch. 449, 564 (2005)
P.G. Kostyuk, O.A. Krishtal, J. Physiol. 270, 569 (1977)
J. Kutchinsky, S. Friis, M. Asmild, R. Taboryski, S. Pedersen, R.
K. Vestergaard, R.B. Jacobsen, K. Krzywkowski, R.L.
Schroder, T. Ljungstrom, N. Helix, C.B. Sorensen, M. Bech,
N.J. Willumsen, Assay Drug Dev. Technol. 1, 685 (2003)
A.Y. Lau, P.J. Hung, A.R. Wu, L.P. Lee, Lab Chip 6, 1510 (2006)
S.S. Lee, S.W. Lee, Microsyst. Technol. 14, 205 (2008)
T. Lehnert, M.A.M. Gijs, R. Netzer, U. Bischoff, Appl. Phys. Lett. 81,
5063 (2002)
T. Lehnert, D.M.T. Nguyen, L. Baldi, M.A.M. Gijs, Microfluid.
Nanofluid. 3, 109 (2007)
H.D. Lux, A.M. Brown, J. Gen. Physiol. 83, 727 (1984)
P.K. McCamphill, T.W. Dunn, N.I. Syed, Eur. J. Neurosci. 27, 2033
G. Mealing, M. Bani-Yaghoub, C. Py, R. Voicu, R. Barjovanu, R.
Tremblay, R. Monette, J. Mielke, K. Faid, in Proceedings of the
International Joint Conference on Neural Networks (2005), pp.
D.W. Munno, D.J. Prince, N.I. Syed, J. Neurosci. 23, 4146 (2003)
W.-L. Ong, K.-C. Tang, A. Agarwal, R. Nagarajan, L.-W. Luo, L.
Yobas, Lab Chip 7, 1357 (2007)
R. Pantoja, J. Nagarah, D. Starace, N.A. Melosh, R. Blunck, F.
Bezanilla, J.R. Heath, Biosens. Bioelectron. 20, 509 (2004)
A. Priel, Z. Gil, V.T. Moy, K.L. Magleby, S.D. Silberberg, Biophys. J.
92, 3893 (2007)
C. Py, G. Mealing, M. Denhoff, A. Charrier, R. Monette, T. Comas, T.
Ahuja, D. Martinez, A. Krantis, S.Wingar, in Proceedings of the 12th
Micro-Total Analysis Systems Conference (2008), pp. 507–509
C. Schmidt, M. Mayer, H. Vogel, Angew. Chem. Int. Ed. 39, 3137
K. Schroeder, B. Neagle, D.J. Trezise, J. Worley, J. Biomol. Screen. 8,
50 (2003)
W.J. Sigurdson, C.E. Morris, J. Neurosci. 8, 2801 (1989)
F.J. Sigworth, K.G. Klemic, IEEE Trans. Nanobiosci. 4, 121 (2005)
A.B. Smit, N.I. Syed, D. Schaap, J. van Minnen, J. Klumperman, K.S.
Kits, H. Lodder, R.C. van der Schors, R. van Elk, B.
Sorgedrager, K. Brejc, T.K. Sixma, W.P.M. Geraerts, Nature
411, 261 (2001)
T. Sordel, S. Garnier-Raveaud, F. Sauter, C. Pudda, F. Marcel, M.
De-Waard, C. Arnoult, M. Vivaudou, F. Chatelin, N.
Picollet-D’hahan, J. Biotechnol. 125, 145 (2006)
A. Stett, C. Burkhardt, U. Weber, P. van Stiphout, T. Knott, Recept.
Channels 9, 59 (2003a)
A. Stett, V. Bucher, C. Burkhardt, U. Weber, W. Nisch, Med. Biol.
Eng. Comput. 41, 233 (2003b)
N.I. Syed, A.G.M. Bulloch, K. Lukowiak, Science 250, 282 (1990)
N.J. Willumsen, M. Bech, S.P. Olesen, B.S. Jensen, M.P.G.
Korsgaard, P. Christophersen, Recept. Channels 9, 3 (2003)
M.A. Woodin, D.W. Munno, N.I. Syed, J. Neurosci. 22, 505 (2002)
Biomed Microdevices

Related Posts Plugin for WordPress, Blogger...
Be Sociable, Share!

About the Author

has written 1822 posts on this blog.

Copyright © 2017 Medical Technology & Gadgets Blog All rights reserved.
Proudly powered by WordPress. Developed by Deluxe Themes