View Article Online View Journal

Lab on a Chip Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: G. Petit-Pierre, A. Bertsch and P. Renaud, Lab Chip, 2016, DOI: 10.1039/C5LC01544H.

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication. Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available. You can find more information about Accepted Manuscripts in the Information for Authors. Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

www.rsc.org/loc

Page 1 of 8

Lab on a Chip View Article Online

DOI: 10.1039/C5LC01544H

Neural probe combining microelectrodes and a dropletbased microdialysis collection system for high temporal resolution sampling Guillaume Petit-Pierre,∗a Arnaud Bertsch,a and Philippe Renauda

We propose a novel neural probe which combines microfluidic channels with recording and stimulation electrodes. The developed microfabrication approach enables the concentration of every active element such as electrodes and the sampling inlet in close proximity on the same surface. As a first approach, full functional validation is presented in this work (in vivo testing will be presented in a next study). Electrical characterization by impedance spectroscopy is performed in order to assess electrodes properties. An advanced experimental setup enabling the validation of the fluidic functions of the neural probe is also presented. It allowed the achievement of high temporal resolution (170 ms) during sampling as a result of the integration of a T-junction droplet generator inside the probe. The droplets reached a volume of 0.84 nL and are separated by a non-aqueous phase (Perfluoromethyldecalin, PFD). This probe represents an innovative tool for neuroscientists as it can be implanted in precise brain structure while combining electrical stimulation with sampling at high temporal resolution.

1

Introduction

Neural interfaces and their abilities to closely communicate with the central and peripheral nervous system is the basis for the development of efficient Neuroprosthetics. The application of these technologies in neurology led to the development of effective medical procedures such as Deep Brain Stimulation (DBS) which treats Parkinson’s disease. The development of these therapies initially started with the use of neural probes, a neuroscientist tool first designed to record electrical signals from precise brain areas. Various types of neural probes have been developed in the past 60 years. The first devices which were used to perform an electrical recording of the brain consisted of a single micro wire and were developed in the 1950’s. 1 Michigan silicon arrays combined 4 recording sites along the probe shank and therefore provided higher spatial resolution compared to single micro wire. 2 The Utah array 3 provided for the first time a multichannel approach using methods borrowed from the integrated circuit manufacturing industry. This rigid 100 channels device is famous for being a reliable neural interface which performs recordings and stimulations at the level of the sensorimotor cortex. 4 More recently, biocompatible polyimide-based devices integrating

a

Laboratory of Microsystems LMIS4, Swiss Federal Institute of Technology Lausanne (EPFL), Lausanne, Switzerland. Tel: 0041 21 693 65 67; E-mail: [email protected]

platinum (Pt) electrodes were developed. 5–7 These systems demonstrated reduced tissue inflammation and improved long term biocompatibility. 8,9 In parallel to these developments, additional brain interaction strategies based on integrated microchannels in neural probes were developed. 10,11 The first applications were dedicated to drug delivery. 12 More recently, a spectacular demonstration of the combination of electrodes and microfluidic channels on the same device was presented by Altuna et al.. 13 They implanted SU-8 probes in anaesthetized rat brains and demonstrated a high correlation between drug delivery and consecutive recorded electrical activity. The device proposed by Minev et al. also demonstrated the potential of simultaneous electrical and fluidic stimulation, implemented on the same device to restore locomotion after paralyzing spinal cord injury. 14 As the extraction of brain fluids such as Cerebrospinal Fluid (CSF) started to represent an efficient way to monitor brain-related diseases, microdialysis devices were developed by neuroscientists. The working principle is based on a gradient of concentration existing between two compartments separated by a semi-permeable membrane to drive a molecular diffusion inside the probe. Their successful operation for quantifying molecules levels in neuronal tissue demonstrated by Ungerstedt et al. contributed significantly to the use of this method in neuroscience. 15,16 However, microdialysis probes are relatively large with a typical

J our na l Na me, [ y ea r ] , [ vol . ] , 1–8 | 1

Lab on a Chip Accepted Manuscript

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

J our nal Name

Lab on a Chip

Page 2 of 8 View Article Online

diameter of 200 µm to 400 µm and a sampling length of 1 mm to 4 mm, thus drastically limiting spatial resolution. 17 Furthermore, the diffusion through membrane and Taylor dispersion occurring in the sampling tube limits the temporal resolution that can be achieved. 18 An answer to the diffusion issue could be found in the implementation of segmented flows and droplets transport within the probe microfluidic channels. Indeed, Taylor dispersion is an effect in which a finite distribution of solute in a liquid flowing through a tube tends to spread under the combined action of molecular diffusion and the variation of velocity over the cross-section. 19 The segmentation of the collected liquid by a non-aqueous phase limits Taylor dispersion because each droplet includes a sub-sample enclosed in a defined space. If the droplet is conserved in the same state and does not fuse with another one during the transportation to the analysis spot, the temporal resolution of the sampling, defined by the time between two generated droplets, is maintained until analysis of the sample. Tremendous work on droplets microfluidics has been carried out with the perspective to use this method for Lab-on-a-Chip applications. Droplets microfluidics is based on the separation of a continuous phase by droplets of an immiscible fluid or by bubbles. One of the motivations in developing this technology was to avoid axial dispersion occurring in non-segmented flows. 20 Several ways to generate droplets were demonstrated. With the T-junction configuration, the break-up of a continuous stream is caused by the meeting with a second perpendicular immiscible stream and the consequent shear stress generated. 21 Another method called flow-focusing uses two streams of continuous phase to narrow a perpendicular dispersed liquid in a cross junction until the pinch-off takes place. 22 These methods are of particular interest when combined with biological applications. The chemistrode presented by Chen et al. typically combines the field of droplet microfluidics with biology. 23 This device aims to manipulate chemical signals instead of electrical signals as usually done with electrodes. The chemistrode includes PDMS v-shaped microchannels. It uses droplets surrounded by an oily carrier phase in order to deliver or record molecular signals to/from a biological substrate such as a cell or tissues. Although the chemistrode opens new perspectives regarding chemical interaction, it is however not designed with implantable grade material and therefore excludes in-vivo neurological applications. In another application, Slaney et al. presented a push-pull probe combining microdialyisis with droplets segmentation. 24 In their approach, after the extraction, the liquid was separated by an oily phase which improved temporal resolution and facilitated the analysis of the samples. However, this method required the development of an external segmentation device connected to the output flow of the probe. Furthermore, it does not address the diffusion which still occurs between the sampling spot and the generation of the droplets. In the present article, we propose a novel concept of neural probe integrating Pt electrodes and a microfluidic T-junction in the distal tip. The positioning of the T-junction in close prox-

2|

J our na l Na me, [ y ea r ] , [ vol . ] , 1–8

imity with the sampling inlet (170 µm) enables to immediately segment the fluid in droplets (nanoliter range) separated by an immiscible carrier phase. The temporal resolution, defined by the time between the generations of two droplets, is therefore improved by preventing Taylor dispersion which usually occurs in continuous flow during transport to an analysis system. The T-junction allows to generate droplets at high frequency (under the second) demonstrating high temporal resolution. The use of biocompatible grade materials decreases the tissue reaction for in-vivo applications. On top of that, we propose the integration of recording and stimulation electrodes on the same device with the objective to trigger the release of neurotransmitters or record electrical activity of precise brain regions. The microfluidic inlet and the electrodes are integrated on the same surface, in close proximity to each other, in order to provide high spatial resolution.

2

Microfabrication

2.1

Material and Methods

The neural probe was fabricated using microfabrication methods. Instead of starting with the back of the probe and ending with the device top layer, the opposite procedure was followed. It allowed to concentrate every active elements on the same surface of the final device. The active elements include the Pt electrodes and the sampling inlet assembled on an initial polyimide (Pi) base. The main steps of the fabrication process are summarized in figure 1. Two silicon wafers were processed in parallel. On the first wafer (figure 1a), a 1 µm thick WTi(200 nm) - Al(800 nm) sacrificial layer was sputtered. Then, a 3 µm Pi layer was spincoated and patterned using photolithography. Before the etching step, a reflow of the photoresist mask (AZ-92XX) at 115 ◦C during 2 min was performed. This allowed to generate a conical profile rather than a vertical wall between the two levels of the metal layer (electrical tracks and electrodes) thus preventing rupture in case of bending. A 350 nm sandwich Ti-Pt-Ti metal layer was then sputtered and etched to form the electrodes and the electrical tracks. Electrical insulation of this layer was obtained with a second 3 µm spin-coated Pi layer patterned by photolithography. It allowed to open electrodes included into the microfluidic channel as well as the sampling inlet. Finally, a 40 µm thick SU-8 photoresist layer was deposited and patterned by photolithography. It served as a structural material forming the microchannels walls. A 50 µm Mylar film was laminated on the second wafer (figure 1b). A droplet of water was applied between the wafer and the Mylar foil to enhance the bonding. 25 Then, a 40 µm thick SU8 layer was spin-coated over the film and soft-baked. After this step, the Mylar film with the SU-8 layer on the top was peeled off the second wafer. Then, the Mylar and SU-8 structure was flipped bottom-up and assembled on the top of the first wafer. The bonding between SU-8 layers was obtained by lamination applying heat (60 ◦C) and a compression force (36 kg, between the rolls) (figure 1c). A final photolithographic step (exposition through the Mylar film) allowed to reinforce the bonding (figure 1d).

Lab on a Chip Accepted Manuscript

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

DOI: 10.1039/C5LC01544H

Page 3 of 8

Lab on a Chip View Article Online

(a) 1st wafer: Al WTi

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

wafer 1 Pi

Pt+Ti

Pi

SU-8

(b) 2nd wafer: SU-8 Mylar

wafer 2

(c) Bonding:

Two microfluidic channels of rectangular section (40 µm × 80 µm) cross the structure from the legs to the tip of the probe. They join at the distal tip where they form a T-junction with a third microchannel connected to the sampling inlet (D = 80 µm). The white dashed-line in figure 2 highlights the pathway of the microfluidic channels. Close to the sampling inlet, two stimulation electrodes with a diameter of 150 µm (distal) and 100 µm (proximal) are distributed longitudinally. The proximal stimulation electrode is smaller in order to provide sufficient room for the electrical tracks (10 µm) passing nearby. Four recording electrodes with a diameter of 20 µm are distributed in a tetrode-like configuration. This permits to discriminate signals from individual neurons when recording. Additionally, square electrodes (80 µm × 80 µm) are integrated inside the channel, facing down. On the left of figure 3 is presented a cross-section picture of the probe tip performed with a scanning electron microscope (SEM, LEO 1550). The openings of the two microchannels as well as the 8 electrical tracks assembled on the top of the structure are well distinguishable. One can observe that a slight misalignment occurred between the SU-8 cover and the rest of the structure during the bonding. This is particularly noticeable on the sides of the probe cross-section, along the structure thickness (see figure 3, left). On the right of figure 3, a SEM picture shows the two stimulating electrodes as well as the 4 recording electrodes. The sampling inlet at the middle of the tetrode electrodes can be observed as well (white arrow).

3

Experimental Validation

3.1

Electrical Characterization

wafer 1

(d) Release:

microchannel

Fig. 1 Main fabrication steps of the neural probe.(a) The first wafer includes Pi (polyimide), Pt (platinum) electrodes and SU-8 (microchannels wall).(b) The second wafer includes a Mylar film and a SU-8 layer (microchannels cover).(c) The bonding is performed by lamination before the final release of the probe (d).

The release of the probes was performed by anodic dissolution of the sacrificial layer according to the method proposed by Metz et al.. 26 After the release, the devices were dipped in a 10:1 HF-DI water solution for 10 s to remove the Ti layer above the Pt layer. 2.2

Neural Probe

A neural probe including microfluidic channels, recording and stimulation electrodes was successfully fabricated as shown in figure 2. A zoom on the distal tip shows the electrodes. A T-junction droplets generator has been included in the tip of the probe. This enables the collection and the direct segmentation of the extracted fluid. The proximal area of the probe includes two legs which constitute the input and output of the microfluidic channels. On the same region, the electrical contacts (2 mm × 2 mm) allow to make the connection with the electrodes of the distal tip.

In a first approach, the electrochemical behaviour of the electrode/electrolyte interface was examined using impedance spectroscopy compared with a silver/silver-chloride reference electrode (Ag/AgCl). The measurements were conducted using a commercially available impedance analysis system (Agilent 4294A, Agilent technologies). Impedance modulus and phase were recorded at discrete frequencies in the 100 Hz to 5 MHz range. The probe and the counter electrode were immersed in a saline solution (0.9 % NaCl pH 7.0, ρ= 16 mS cm−1 ) and the impedance spectroscopy performed at an oscillation amplitude of 50 mV. The electrode-electrolyte behaviour can be approximated by an equivalent circuit including a resistive and a capacitive element mounted in parallel. 9,27 The interfacial capacitance is approximated by a constant phase element (CPE) which best describes the double layer behaviour at the interface with the solution as well as the contribution of the electrode surface roughness. The charge transfer resistance (RCT ) constitutes the resistive element and model the electron-ion exchange at the electrode surface. The CPE and the RCT (in parallel) are combined in series with a spreading resistance (Rs) and a bulk resistance (Rbulk) as shown in figure 4 (a). Figure 4 (b) and (c) show a typical impedance/phase spectrum for the recording electrode (D = 20 µm) and the stimulation electrode (D = 150 µm), respectively. The spectra obtained are in accordance with similar devices previously characterized with the same method. 8,28 Using the Peak J our na l Na me, [ y ea r ] , [ vol . ] , 1–8 | 3

Lab on a Chip Accepted Manuscript

DOI: 10.1039/C5LC01544H

Lab on a Chip

Page 4 of 8 View Article Online

DOI: 10.1039/C5LC01544H

10 mm

Contacts (8x)

Distal tip

240 µm

Microfluidic channels (2x) Proximal side Droplets detection electrodes

Stimulation electrodes (Ø=100, 150 µm)

Recording electrodes (Ø=20 µm, 4x)

Fig. 2 Polyimide and SU-8 based neural probe after microfabrication. The device includes stimulation and recording Pt (platinum) electrodes as well as 2 microfluidics channels joining at the distal tip to form a T-junction (droplets generator). The white-dashed line represents the microchannels pathway.

20 µm

100 µm

Fig. 3 (Left) SEM picture of the probe tip after a cross-section cut. The microchannels are open, and, on the top of them, we observe the 8 electrical tracks.(Right) The stimulation and recording electrodes as well as the sampling inlet (white arrow) are integrated on the same surface, in close proximity to each other.

Resistance Frequency method (PRF), we observe a difference of one order of magnitude for the spreading resistance (Rsel.20 µm = 2.0 × 104 Ω, Rsel.150 µm = 3.1 × 103 Ω ) when the phase is closest to 0. 9 This gap is explained by the difference in size of the electrode (20 µm vs. 150 µm), the spreading resistance being inversely proportional to the diameter of the electrode. 29 Additionally, the electrical resistance of each track was measured with a 4 points probe station (SüssMicrotec, PM8). The measurement was performed track by track, between the proximal contact and its corresponding electrode (see figure 2). Contacting the electrodes integrated in the microfluidic channels required the delamination of the bottom SU-8 layer in order to access the conductive metal layer. Each electrodes was successfully connected and the mean track electrical resistance was 709 ± 46 Ω. Compared to the impedance magnitude found at the biologically relevant frequency of 1 kHz (20 µm electrode, R = 4.7 × 104 Ω), the contribution of the mean track electrical resistance will have a negligible influence on the measurement (709 Ω vs. 4.7 × 104 Ω).

3.2 3.2.1

Fluidic Characterization Droplet Collection

The droplet collection system is controlled by modulating the flow of a Perfluoromethyldecalin (PFD) phase which circulates in the microchannels of the probe. The PFD is injected in one of the fluidic entry and sucked up by the other one. For this purpose, fused silica capillaries (ID = 250 µm, OD = 360 µm) have been connected to the probe legs (proximal area, see figure 2). In order to ensure a pressure tight connection, the legs have been introduced into the capillary and assembled with epoxy glue. Each capillary is connected to a pressure-driven reservoir (Fluigent, module MFCS-MZ) filled with PFD . This system enables to apply a positive or negative pressure on the probe legs, with respect to the ambient pressure. The fluidic behaviour of the probe can therefore be modelled by a 3 entries system connected by the Tjunction (included in the probe tip). Qin and Qout are the flow resulting from the pressure applied on each microchannel of the probe. Qsampled is the extracted flow from the sampling inlet. By the conservation law, we can write: Qin + Qout = Qsampled From this equation, we deduce:

4|

J our na l Na me, [ y ea r ] , [ vol . ] , 1–8

(1)

Lab on a Chip Accepted Manuscript

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

Fluidic inputs and outputs

Page 5 of 8

Lab on a Chip View Article Online

DOI: 10.1039/C5LC01544H

Qout Qin

RCT Rs

water

Rbulk

PFD

CPE Electrode-electrolyte interface

(b) -20

10 5

-40

10 4

-60 -80

10 3 10 3

10 4

10 5

10 6

Frequency (Hz)

t=0

t = 0.126 s

t = 0.140 s

t = 0.154 s

Fig. 5 (Left) T-junction model with the fluidic input Qin , the fluidic output Qout and the collected fluid Qsampled . (Right) Sequence of a water droplet generation in a PFD carrier phase occurring near the T-junction. At t=0, the droplet (Volume = 0.62 nL) is in preparation. The droplet detaches (t=0.126 s) and propagates along the microchannel (from t=0.140 s to t=0.154 s).

(c) 10 7

0 = impedance = phase

-20 -40

10 5

-60

10 4 10 3 10 2

Phase (deg)

10

6

-80 10

3

10

4

10

5

10

6

Frequency (Hz) Fig. 4 Equivalent electrical model for an electrode-electrolyte interface (a). Impedance magnitude and phase for a 20 µm recording electrode (b) and a 150 µm stimulation electrode (c).

Qout > Qin ⇒ Qsampled > 0

(2)

Such a system is represented on the left of figure 5. On the right of figure 5 a sequence in which a droplet is generated is shown. In order to obtain this result, the neural probe has been mounted on a dedicated setup allowing to collect clear water in the PFD carrier phase while filming with a camera (IDS, uEye UI1240SE). The red crosses highlight the droplet outlines. At t = 0 s, the droplet is in preparation and still attached to the sampling inlet, at t = 0.126 s the droplet detaches and then propagates along the channel. At the end of the sequence, we can notice that the next droplet is already in preparation. For this demonstration, a constant pressure difference between the microchannels of ∆P = 50 mbar was applied. This enabled to generate droplets of 0.62 nL at a frequency of 0.13 Hz, meaning a droplet each 7.8 seconds. This is rather low but higher frequencies can be obtained by increasing ∆P as presented in the next chapter. 3.2.2

High Frequency Droplet Sampling

Experimental Setup. In order to validate the concept of direct droplets sampling, a dedicated setup has been assembled. A representation of this experimental setup can be seen on the top of figure 6. Two syringes filled with blue-colored water and clear water are connected to a single tube (Saint-Gobain, Tygon, ID=0.51 mm, OD=1.52 mm). They are operated independently

by 2 syringe-pumps (Cetoni, Nemesys) and are used in order to generate a color transition in the common tube. A neural probe is inserted in the same tube. It is operated by the two pressurepumps in order to perform a continuous droplet collection of the water circulating in the tube. A first measurement of the mean light intensity through the tube, before the sampling spot, is performed with a photodiode (Vishay, model BPW34) and an amplifier (Stanford Research System, SR570). A second measurement of mean light intensity, focused on the passing droplet on the microchannel of the probe, is performed with a microscope (Zeiss, Axiowert S100 TV). The signal is recorded using a camera (IDS, uEye UI-3600CP) and the mean light intensity of images computed using the software ImageJ. Experimental Conditions. A fast transition from clear water to blue water was generated in the tube. In order to obtain this effect, the tube was first filled with clear water and flushed during 5 s at a flow rate of 10 µL s−1 using the syringe pumps. Then, the clear water syringe pump was switched off. The droplet sampling process was started by modulating the pressure applied to the PFD circuit connected to the probe. A pressure of Pin = 50 mbar and Pout = −400 mbar was applied (relatively to the local pressure). Then, the flow of blue dyed water from the second syringe pump was turned on at a flow rate of 10 µL s−1 . This resulted in creating a fast color change in the tube recorded both before the sampling spot, via the photodiode, and after sampling, via the droplets in the probe, as illustrated on the top of figure 6. When the mean light intensity measurement is stabilized in both spots (A and B) the recording is stopped. Result. Under these conditions, fast color changes were generated and captured within the sampling system of the probe. In the bottom of figure 6, the concentration change measured at the location A (dashed line) and B (continuous line) is showed. On measurement A (located before the sampling spot) we can note a fast decrease of intensity starting at time t = 1 s that stabilizes after 2.5 s. The major part of this transition occurs in 1.5 s. On measurement B, light intensity through the probe microchannel is presented. Each peak corresponds to a passing droplet thus demonJ our na l Na me, [ y ea r ] , [ vol . ] , 1–8 | 5

Lab on a Chip Accepted Manuscript

= impedance = phase

10 6

10 2

Qsampled

0

Phase (deg)

| Z | (Ω)

10 7

| Z | (Ω)

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

(a) Equivalent model:

Lab on a Chip

Page 6 of 8 View Article Online

DOI: 10.1039/C5LC01544H

Qin

flow direction

Qout

tube

Qin

7

125

120

5 4

115 3 2

110

1 continuous = B

dashed = A

0 0

0.5

1

1.5

2

2.5

3

Mean Intensity B (a.u.)

6

105

3.5

Time (s) Fig. 6 (Top) Representation of the setup designed in order to validate the concept of droplet sampling. (Bottom) Mean intensity measured at position A and B, related to the position before and after the sampling spot. Strong temporal correlation between both measurements is demonstrated.

strating a relatively fast collection at a frequency of 6 Hz. Under these conditions, the mean volume of a droplet was 0.84 nL. The level at which the signal returns after every droplet passes corresponds to the signal intensity of the PFD phase. A strong correlation regarding the time at which the change occurs and the shape of the curves is demonstrated on measurement A and B. This indicates that the droplet sampling concept of the neural probe is able to thoroughly capture a rapid change of concentration occurring in the external fluid.

4

Discussion

Microdialysis probes are typically 200 µm to 400 µm in diameter with a sampling length of 1 mm to 4 mm. 17 Such a volume limits the analysis to rather large brain regions. In term of spatial resolution, the neural probe presents the advantage to be much more specific; the sampling zone is a single circular area of a diameter of 80 µm. In term of collection area, this is 2 order of magnitude smaller than the membrane exchange zone 6|

J our na l Na me, [ y ea r ] , [ vol . ] , 1–8

Achieving high temporal resolution during in-vivo sampling in brain tissues is a challenge since neurotransmitters release typically occurs in the range of the second. 30,31 The maximum achieved sampling rate of our current device was 6 Hz (meaning a droplet every 170 ms). It indicates we should be able to capture in real time a concentration change of a neurotransmitter occurring in the brain with this tool. One of the main limitation to this approach relates to the maximum volume of interstitial liquid we can remove from the tissue without causing any damages. Since the volume of each droplet is 0.84 nL, it represents, at a frequency of 6 Hz, a removal of an equivalent constant flow of 0.30 µL min−1 . The impact of direct sampling in the neural tissue was demonstrated to be minimized when the flow rate was kept under 50 nL min−1 and the total volume removed under 2.5 µL. 32 Our approach could deplete quickly the medium since the equivalent flow rate is 6 times higher than 50 nL min−1 . However, the effect of depletion can be strongly limited by performing the sampling during short periods of time typically during 1 minute or less which would be fully compatible with our approach. In this situation and with the parameters cited above (0.30 µL min−1 during 1 min), the total volume removed is 300 nL which seems acceptable. We can even imagine to generate single droplets "on-demand" with fast burst of pressure. This aspect will have to be further investigated in the future. Another approach for limiting the depletion would be to replace the liquid removed by an equivalent volume of artificial cerebrospinal fluid injected at the same spot. It could easily be implemented by adding a microfluidic channel on the probe dedicated to this function. Intracranial pressure (ICP) variation over time and its impact on device operation must also be evaluated. ICP variation in rat brain is reported to be relatively stable over the time, unless the animal is affected by strong neurological injury. 33,34 Standard variation of ICP were not reported to be an issue for the devices operating in a direct sampling mode or in a push-pull mode. 24,32,35 However, immediately after insertion of the probe in brain tissues, the respect of a delay before starting the sampling should allow the local ICP to stabilize. Indeed, the insertion procedure and possible exposure to the room local pressure could affect the ICP value and disturb the sampling process in a short term range.

Lab on a Chip Accepted Manuscript

Qout

B

A

Mean Intensity A (a.u.)

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

probe

of microdialysis probes (diameter = 200 µm, length = 1 mm considered). This should enable to record chemical signals at more specific targets of the brain. Furthermore, the microfabrication process developed allowed the electrodes to be in close proximity with the sampling inlet (see figure 2). Regarding the volume, the probe shank is 240 µm wide and 86 µm thick, which is sensibly smaller and less invasive than microdialysis probes. Devices combining microchannels and microelectrodes previously presented in scientific literature are of similar size of ours. For example, the one presented by Altuna et al. comprises a 150 µm wide and 55 µm thick shank which is slightly smaller than our probe. 13

Page 7 of 8

Lab on a Chip View Article Online

Regarding the time resolution, there are good chances to further increase the sampling rate by driving the system at a superior difference of pressure between the entries. This will result in the reduction of the space between droplets (PFD segment volume) and the increase of the droplets volume as supported by Gupta et al.. 36 Increasing the volume of the droplets is a benefit regarding the analysis of the samples. A limitation can however be related to the actuation system which must be fast and compatible with high pressure controls. Admitting that such equipment is available, special care will be required to reinforce the fluidic connections with the neural probe. In a preliminary work, the neural probe was connected to syringe pumps that could operate the system at high flow rate. In this configuration, the fluidic connection ruptured when the flow rate reached a value above 1 µL/ sec, meaning a pressure of 1.5 bar with our geometry.

strate the in vivo application of this technology.

The choice of the analysis method must require attention since each approach is limited both by a minimal volume of detection and a given sensitivity. 17 In addition to these constraints, the segmented flow combining hydrophobic (PFD) and hydrophilic phases further limit the choice. This segmentation guarantees the temporal resolution of the measurement meaning that any cross-contamination must be avoided. HPLC (highpressure liquid chromatography) coupled to electrochemical or fluorescence detection would have been good candidates but are incompatible with the current volume of the droplets (V = 0.84 nL). It is too small compared to the typical 10 µL to 20 µL injected samples. 37 Electrospray Ionisation Mass Spectroscopy has also been proposed as a compatible method for analysing droplets segmented samples. 24 However this method required the addition of a reagent of identical volume in each droplet before analysis which can be challenging to control. Another potential approach achieving high sensitivity with low sample volume (pL to nL) is capillary electrophoresis combined with LIF. This would require further developments in order to sort and manipulate the droplets of interest while still controlling their sequence. Offline measurement is probably more appropriate than online measurements for the analysis of the segmented samples collected by the probe. Indeed, droplets and the carrier PFD phase could be accumulated and stored in the capillaries interfaced with the probe fluidic output and later analysed. This would further enable to manipulate the droplets, analyse them one by one, without the constraint of maintaining a high flow rate. Potential applications of the technology include collecting samples (accumulated droplets) and precise separation correlated with different brain activity states such as sleep, awake or seizures periods. Specific neurochemical release patterns after electrical stimulation of precise brain regions could also be monitored with this device. Closed-loop monitoring with collection of samples triggered by specific recorded neuronal activity could be of great interest as well. For instance, it could lead to a better characterization of brain chemical release mechanisms in neurodegenerative diseases such as in Parkinson’s or Alzheimer’s disease. Such a work will be carried out by our group in a next step as an increment to demon-

The authors are grateful to the staff of the EPFL Center for Micronanotechnology (CMI) where clean room microfabrication was performed. More particularly, the assistance of C. Hibert, K. Suter, J. Dorsaz , J. Pernollet and G.-A. Racine during processing was really helpful and they are acknowledged for their support.

5

Conclusion

We present a novel neural probe which combines, in one single tip, recording and stimulation electrodes with a droplet generator. The probe’s electrical and fluidic functions were assessed. In particular, the droplet generation system, based on a T-junction, allowed to efficiently sample external fluid at a temporal resolution of 170 ms. The use of microfabrication methods provided a reliable mean to concentrate all the active elements of the probe in close proximity, thus proving high spatial resolution. This system is proposed as an advanced tool which can electrically trigger the release of molecules of interest (such as neurotransmitters) in targeted brain area, combining both electrical stimulation and droplet sampling methods.

Acknowledgements

References 1 R. Gesteland, B. Howland, J. Lettvin and W. Pitts, Proceedings of the Ire, 1959, 47, 1856–62. 2 K. L. Drake, K. D. Wise, J. Farraye, D. J. Anderson and S. L. BeMent, IEEE transactions on bio-medical engineering, 1988, 35, 719–32. 3 P. K. Campbell, K. E. Jones, R. J. Huber, K. W. Horch and R. A. Normann, IEEE Transactions on Biomedical Engineering, 1991, 38, 758–768. 4 E. M. Maynard, C. T. Nordhausen and R. A. Normann, Electroencephalography and Clinical Neurophysiology, 1997, 102, 228–239. 5 S. Metz, F. Oppliger, R. Holzer, B. Buisson, D. Bertrand and P. Renaud, 1st Annual International IEEE-EMBS Special Topic Conference on Microtechnologies in Medicine and Biology., 2000, pp. 10–14. 6 T. Stieglitz, H. Beutel, M. Schuettler and J.-U. Meyer, Biomedical Microdevices, 2000, 2, 283–294. 7 P. J. Rousche, D. S. Pellinen, D. P. Pivin, J. C. Williams, R. J. Vetter and D. R. Kipke, IEEE Transactions on Biomedical Engineering, 2001, 48, 361–370. 8 A. Mercanzini, K. Cheung, D. Buhl, M. Boers, A. Maillard, P. Colin, J.-C. Bensadoun, A. Bertsch, A. Carleton and P. Renaud, 2007 IEEE 20th International Conference on Micro Electro Mechanical Systems (MEMS), 2007, pp. 573–576. 9 A. Mercanzini, P. Colin, J.-C. Bensadoun, A. Bertsch and P. Renaud, IEEE transactions on bio-medical engineering, 2009, 56, 1909–1918. 10 J. Chen, K. D. Wise, J. F. Hetke and S. Bledsoe, IEEE Transactions on Biomedical Engineering, 1997, 44, 760–769. 11 S. T. Retterer, K. L. Smith, C. S. Bjornsson, K. B. Neeves, A. J. H. Spence, J. Turner, W. Shain and M. S. Isaacson, J our na l Na me, [ y ea r ] , [ vol . ] , 1–8 | 7

Lab on a Chip Accepted Manuscript

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

DOI: 10.1039/C5LC01544H

Lab on a Chip

Page 8 of 8 View Article Online

Published on 29 January 2016. Downloaded by Middle East Technical University (Orta Dogu Teknik U) on 02/02/2016 22:53:43.

12 13

14

15 16 17 18 19 20 21 22 23

24

8|

IEEE Transactions on Biomedical Engineering, 2004, 51, 2063– 2073. S. Metz, A. Bertsch, D. Bertrand and P. Renaud, Biosensors and Bioelectronics, 2004, 19, 1309–1318. A. Altuna, E. Bellistri, E. Cid, P. Aivar, B. Gal, J. Berganzo, G. Gabriel, A. Guimerà, R. Villa, L. J. Fernández and L. Menendez de la Prida, Lab on a chip, 2013, 13, 1422–30. I. R. Minev, P. Musienko, A. Hirsch, Q. Barraud, T. Milekovic, L. Asboth, R. F. Torres and N. Vachicouras, Science, 2015, 347, 23–28. T. Ljungberg and U. Ungerstedt, Psychopharmacology, 1978, 56, 239–247. T. Zetterstrom and U. Ungerstedt, European Journal of Pharmacology, 1984, 97, 29–36. R. T. Kennedy, Current Opinion in Chemical Biology, 2013, 17, 860–867. R. Aris, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 1956, pp. 67–77. G. Taylor, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 1953, 219, 186–203. R. Seemann, M. Brinkmann, T. Pfohl and S. Herminghaus, Reports on Progress in Physics, 2011, 75, 016601. H. Song, J. D. Tice and R. F. Ismagilov, Angewandte Chemie International Edition, 2003, 42, 768–72. S. L. Anna, N. Bontoux and H. A. Stone, Applied Physics Letters, 2003, 82, 364–366. D. Chen, W. Du, Y. Liu, W. Liu, A. Kuznetsov, F. E. Mendez, L. H. Philipson and R. F. Ismagilov, Proceedings of the National Academy of Sciences of the United States of America, 2008, 105, 16843–48. T. R. Slaney, J. Nie, N. D. Hershey, P. K. Thwar, J. Linderman,

J our na l Na me, [ y ea r ] , [ vol . ] , 1–8

25

26 27

28 29 30 31 32 33 34 35 36 37

M. A. Burns and R. T. Kennedy, Analytical Chemistry, 2011, 83, 5207–13. J. Steigert, O. Brett, C. Müller, M. Strasser, N. Wangler, H. Reinecke, M. Daub and R. Zengerle, Journal of Micromechanics and Microengineering, 2008, 18, 095013. S. Metz, S. Jiguet, A. Bertsch, P. Renaud and ., Lab on a chip, 2004, 4, 114–20. A. Altuna, L. Menendez de la Prida, E. Bellistri, G. Gabriel, A. Guimerá, J. Berganzo, R. Villa and L. Fernández, Biosensors and Bioelectronics, 2012, 37, 1–5. D. R. Merrill and P. A. Tresco, IEEE transactions on bio-medical engineering, 2005, 52, 1960–5. J. Newman, Journal of The Electrochemical Society, 1966, 113, 501–02. B. J. Venton, T. E. Robinson and R. T. Kennedy, Journal of neurochemistry, 2006, 96, 236–46. S. Rossel, L. Gonzales and L. Hernandez, Journal of Chromatography B, 2003, 784, 385–93. R. T. Kennedy, J. E. Thompson and T. W. Vickroy, Journal of Neuroscience Methods, 2002, 114, 39–49. G. Wei, X.-C. M. Lu, X. Yang and F. C. Tortella, Journal of neurotrauma, 2010, 27, 1635–41. C. Hiploylee and F. Colbourne, Experimental neurology, 2014, 255, 49–55. S. Kottegoda, I. Shaik and S. A. Shippy, Journal of Neuroscience Methods, 2002, 121, 93–101. A. Gupta and R. Kumar, Microfluidics and Nanofluidics, 2009, 8, 799–812. A. Zapata, V. l. Chefer, T. S. Shippenberg and L. Denoroy, Current protocols in neuroscience, 2009, Chapter 7, Unit 7.4.1–30.

Lab on a Chip Accepted Manuscript

DOI: 10.1039/C5LC01544H

Neural probe combining microelectrodes and a droplet-based microdialysis collection system for high temporal resolution sampling.

We propose a novel neural probe which combines microfluidic channels with recording and stimulation electrodes. The developed microfabrication approac...
3MB Sizes 4 Downloads 9 Views