On Chip Guidance and Recording of ... - ACS Publications

Letter pubs.acs.org/NanoLett

On Chip Guidance and Recording of Cardiomyocytes with 3D Mushroom-Shaped Electrodes Francesca Santoro, Jan Schnitker, Gregory Panaitov, and Andreas Offenhaü sser* Institute of Bioelectronics ICS-8/PGI-8, Forschungszentrum Jülich D-52425 Jülich, Germany ABSTRACT: The quality of the recording and stimulation capabilities of multielectrode arrays (MEAs) substantially depends on the interface properties and the coupling of the cell with the underlying electrode area. The purpose of this work was the investigation of a three-dimensional nanointerface, enabling simultaneous guidance and recording of electrogenic cells (HL-1) by utilizing nanostructures with a mushroom shape on MEAs.

KEYWORDS: Cell guidance, HL-1 cells, multielectrode array, 3D electrodes, cell−electrode interface

M

in influencing cell properties such as morphological and physiological changes due to the interaction of the cell with modified surfaces.20,21 It is biologically not completely understood how cells interact with different kinds of 3D structures, although there is a good indication that the actin is the major cellular component for the cell−chip interaction by forming a ring-shape morphology around the micro- and nanostructures.22 Furthermore, shaping the 3D microstructures using a biomimetic design has turned out to be a successful approach for applications with neuronal cells as shown in ref 23. Moreover, Hai et al. recently showed24 how the interspine space between 3D microspines in an array configuration influences the outgrowth of Aplysia californica neurons: matrices with interspine spaces of 4 and 8 μm exhibited a major influence on the neurites outgrowth in contrast to ones with a higher interspace. A similar approach was investigated, where primary cortical neurons showed preferential outgrowth when plated on gold mushroom-shaped spines functionalized with amino-terminated self-assembled monolayers (SAM).23 To create well-defined patterns of cells on MEAs, many methods have been proposed in the last years. Protein patterning was one of the first techniques used for guidance with defined geometries and patterns on a substrate using microcontact printing25,26 or with lithography methods.27 In these cases, the patterned protein is recognized by the receptors of the cells that are responsible for the cell−extracellular matrix boundary.28,29 The major limitation of these techniques is that the protein pattern is restricted to two dimensions, since the thickness of the protein layer is negligible compared with the

ultielectrode arrays (MEAs) have been largely used for characterizing cellular networks, especially for the investigation of electrical active cells.1−3 In the recent years, we continuously improved these devices regarding their interface properties and consequently allowing the extracellular recording of action potentials and electrical and optical stimulation.4−8 A crucial point for the recording of extracellular signals is the contact between the cell membrane and the active part of the device: a tight cell−electrode interface can provide transmission of the electrical signal without significant dissipation.9,10 Furthermore, it is of great interest to guide cells on MEAs to allocate cells on the electrodes and create defined cell patterns for understanding the signal propagation between cells. State of the art MEA and field-effect-transistor (FET) biosensors are adapted more toward three-dimensional (3D) microstructures,11 and it was shown that the interaction between the cell and an active device could be greatly improved by means of nanostructures such as metal nanopillars,12 carboncoated electrodes13,14 or microspines,11,15,16 and silicon nanowires,17 making these devices only suitable for physiological investigations by the reduction of the cleft between sensor and cell. It was shown that gold microspines can be engulfed by a cell membrane and not only increasing the effective contact area but also providing a high degree of coupling for the extracellular measurement with such electrodes.18 Yet the interface is not the only challenge for the cells being coupled with an electronic device. The topography and electrochemistry of the surface are just two examples of the properties of the materials involved at the interface; it still remains a challenge to improve the adhesion of very soft cells with a Young’s modulus in the kilopascal regime19 onto a rigid, inorganic material with a stiffness of MPa to GPa. The topographical shape and its chemical properties have a key role © 2013 American Chemical Society

Received: August 2, 2013 Revised: September 19, 2013 Published: October 2, 2013 5379

dx.doi.org/10.1021/nl402901y | Nano Lett. 2013, 13, 5379−5384

Nano Letters

Letter

Figure 1. (A) Scanning electron microscopy (SEM) image of a 3D 64-electrode array (scale bar 500 μm); (B) nanopillar electrodes with guide lines on the passivation layer (scale bar 20 μm); (C) silicon substrate with pattern of gold nanopillars for guidance tests (scale bar 200 μm); (D) single square design with nodes connected by 200 μm lines (scale bar 50 μm).

length and the width of the pattern. As mentioned before, 2D guidance is driven by a receptor-mediated mechanism based on extracellular matrix recognition; it was proposed in the recent years to additionally create micro-nano patterning on substrates to induce cell proliferation, alignment, and stretching.30−32 In this Letter we report about the possibility to use 3D nanostructures for coupling cardiomyocyte-like cells with our recording device. In particular, we investigated how one can combine cell guidance and extracellular recordings of cardiac cells on a chip at the same time. Therefore, we studied the effect of cell guidance on mushroom-shaped 3D gold nanostructures which were fabricated onto a MEA chip for extracellular recordings. In this work, we used a MEA processed as shown in ref 33. The electrode layout was adapted from ref 34, while we increased the overall chip size to 24 × 24 mm2 to simplify the flip-chip encapsulation to a top contact chip. For the fabrication of the gold nanostructures we refer to ref 23. Briefly, a thin gold film was deposited on the substrate, which was then covered by a poly(methyl methacrylate) (PMMA) e-beam resist (Allresist GmbH, Berlin, Germany) with a thickness of about 1 μm. The apertures around 500 nm have been exposed by means of ebeam lithography. Consequently, the openings were filled by electroplated gold using the sputtered gold film as a background electrode. The gold nanostructures were electroplated on top of the planar MEA gold electrode with a diameter of 8 μm as shown in Figure 1A−B. Additionally, gold lines with a length of 180 μm have been fabricated between adjacent electrodes on the top of the passivation layer without shortcircuiting the electrodes (Figure 1B). The width of the lines was about 15 μm in the case of two parallel lines of pillar featuring a 10 μm pitch and was scaled up to six parallel lines when the smallest pitch was fabricated. In the next processing step all of the gold strip-lines were electrically short-circuited by a 50 nm aluminum layer. This circuit was electrically connected to the MEA electrodes to galvanize simultaneously the cell guiding mushroom lines and the electrode gold nanostructures

of the MEA itself. Finally, the resist was removed, and the aluminum shorts were etched with an aluminum wet etchant (ANPE80/5/5/10, Microresist Technology, Berlin, Germany). Adopting these parameters, we were able to fabricate mushrooms with a diameter of about 500 nm, a stalk height of about 1 μm, and a cap height of 200 nm. The overall effective surface area of an individual pillar was about 4.5−10 μm2, depending on the surface roughness of the cap, resulting in a total electrode surface of about 77−115 μm2 in case of the 8 μm (diameter) electrodes for electrophysiological measurements. We determined the equivalent capacitance values of about 30 pF per electrode with the help of impedance spectroscopy. In addition to the MEAs with cell-guiding nanostructures, we fabricated 64 gold pillar node containing samples on silicon oxide for the optimization of spine interspacing in cell guide lines (Figure 1C−D). The pitch between nanopillars was varied in the range from 2 to 10 μm. To cultivate cardiomyocyte-like cell line HL-1,35 the Si test samples and the MEAs were cleaned in flowing ultrapure water for 2 h and then sterilized for 30 min with UV light. After the sterilization, the surface of the substrates was coated with fibronectin at a concentration of 1 mL in 200 μL of 0.02% Bacto TM Gelatin (Fisher Scientific) for an incubation time of 1 h. Confluent HL-1 cells in a T-25 flask were treated with 0.025% trypsin/EDTA, suspended in 5 mL of Claycomb medium, and centrifuged for 5 min at 1700 rpm. The pellet was then resuspended in 3 mL of medium, and 30 μL was plated on the substrates. After 15 min 1 mL of medium was added, and the substrates were incubated for 3 days until the cells formed a confluent monolayer on the lines and started to contract. The living cells on the silicon substrates were stained with 1 mM calcein AM (Invitrogen) and 1 mM ethidium homodimer in phosphate-buffered saline (PBS) solution (137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4, 1.8 mM KH2PO4), and the image acquisition was performed using an Axio Imager Z.1 (Carl Zeiss AG, Oberkochen, Germany). 5380


Nano Letters

Letter

Figure 2. (A) HL-1 cell on Si test sample stained with 1 mM calcein AM and 1 mM ethidium homodimer (scale bar 200 μm); (B) 87% guided cells on Si test sample (i) and 96% guided cells (ii); (C) normalized number of cells plotted as a function of pillar pitch.

Figure 3. (A) SEM images of a guided HL-1 cell on nanopillars with an interspace of 2 μm (scale bar 5 μm); (B) SEM images of a guided HL-1 cell on nanopillars with an interspace of 10 μm (scale bar 8 μm).

For the analysis we first determined the absolute number of cells on the samples and differentiated between cells on the gold nanostructures and the silicon oxide areas (cells not being fully on the pillars were counted as on the silicon oxide) and accounted the underlying total area of both counts for the normalization. The analyzed samples (n = 10) showed a high cell vitality with nearly 100% living cells (calcein staining marked in green in Figure 2A). To investigate the highest guidance effect, the normalized number of cells was plotted as a function of the pitch between adjacent pillars (Figure 2C). As shown in the plot (Figure 2C), the normalized number of notguided cells was on average 10%, in contrast with the 90% of the cells guided by the 3D gold nanostructures. We assume that this is due to the fact that the effective adhesion area of the 3D nanostructures is higher than the flat Si area and in principle cardiomyocytes would rather anchor via focal adhesion proteins and spread on the rough nanopillar surface than on the flat surface. The nanopillars with a pitch from 3 to 7 μm were shown to have the best guidance effect on the HL-1 cells: on average 93−96% of cells were guided in this range of pitches in contrast to only about 87% guided cells by a 2 μm pitch (Figure 2B,i) and an even lower degree of guided cells for pitches between 8 and 10 μm (84−86%). Thus, we were able to guide

up to 96% of the cardiomyocytes in the presence of mushroomshaped structures with a 4 μm pitch (Figure 2B,ii). Similar results were shown for LRM55 cells on silicon pillars36 where 70% of the cells have grown preferably on the 3D structures than on smooth surface. To investigate the performance of the nanopillars that showed the lowest guidance effect, we performed on the pillars with 2 and 10 μm additional scanning electron microscopy (SEM). After the calcein stain, the cells were washed with prewarmed PBS and fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. Dehydration was carried out with ethanol in different concentrations ranging from 10% up to 100% (v/v). Afterward, a critical point drying was performed with CO2 as an intermediate medium for drying the cells. For the SEM images, a thin layer of platinum was sputtered on the sample, and a LEO 1550 (Carl Zeiss AG, Oberkochen, Germany) scanning electron microscope was used for the acquisition. For the acquisition a voltage of 20 kV was applied using an “in lens” detector for the secondary electrons. The images were then acquired in scanning electron mode. From the SEM investigation, we noticed that the cardiomyocytes tended to spread more and flattened in the presence of the bigger pitch as shown in Figure 3B, still engulfing the 3D 5381


Nano Letters

Letter

Figure 4. (A) FIB cross section of HL-1 on 2 μm pitch nanopillars (scale bar 4 μm, tilt 52°); (B) details of membrane not attached on the substrate (scale bar 0.5 μm, tilt 52°); (C) FIB cross section of HL-1 on 10 μm pitch nanopillar (scale bar 5 μm, tilt 52°); (D) details of cell attaching the nanopillar (scale bar 0.5 μm, tilt 52°).

Figure 5. (A) Spontaneous action potentials recorded by 3D electrodes in a time frame of 12 s; (B) single action potentials recording in a time frame of 1.5 s.

1 cells grew on the 3D nanostructures (Figure 4C), while the bottom membrane in addition spread when attaching the planar gold surface below the pillars (Figure 4D). For the voltage recording of extracellular signals, we developed a 64 channel MEA amplifier system which consists of a headstage and a main amplifier connected to a highresolution A/D converter (USB-6255, National Instruments, Austin, Texas, USA) and to a controlling PC. A self-developed LabView software (National Instruments, Austin, USA) controls the recording of the data stream and allows to set amplifier parameters such as gain and filter settings. The headstage connects the MEA chip and amplifies the signal with a gain of 10. The signal was further amplified with a gain of 100 in the main amplifier, resulting in a total nominal gain of 1000. The amplifier system features a parallel readout of 64 channels at a sampling rate of 10 kHz and can record voltages of less

nanostructures without being confined by them. On the other hand, the cells have a more stretched configuration in the presence of nanostructures with a 2 um pitch (Figure 3A): the stretching is due to the cell phenotype where the cytoskeleton rearrangement is driven mostly by myosin and actin filaments. These phenomena were similarly discussed for groove structures37−39 and for human fibroblasts.40 As an additional proof we accomplished a focused ion beam (FIB) cross-section obtained with a Helios Nanolab Dual-beam (FEI, Hillsboro, USA): the polishing and the milling of the cross section have been performed using an ion voltage of 30 kV and current of 80 pA. We performed FIB cross sections for the 2 and 10 μm pitches: in the case of the 2 μm pitch the cells tended to grow completely on the top of the 3D nanostructures (Figure 4A) without approaching the bottom part of the substrate (Figure 4B). On the other hand, the 10 μm pitch HL5382


Nano Letters

Letter

than 1 μV rms with a dynamic range of ±10 V (after amplification). For cellular measurements we limited the effective bandwidth with a high pass filter (AC coupling) and a low pass filter (high frequency cutoff) from 1 Hz to 3 kHz. The analysis of the data was performed with a script written in MATLAB (Mathworks, Natick, USA) and ORIGIN (OriginLab, Northampton, USA). The spontaneous activity of the HL-1 was recorded by the 3D nanostructured MEAs with a nominal electrode footprint of 8 μm diameter with mushroom nanopillars, and on the guiding lines we fabricated nanopillars of a 4 μm pitch according to the guidance results. The traces of spontaneous action potentials are shown in Figure 5A, showing the voltage recordings of four electrodes as a function of time. The frequency of the action potentials amounted 0.62 ± 0.15 Hz for every trace of the example. Due to the peculiar signal propagation of the cardiomyocytes, it is possible to determine a signal propagation speed of 28.6 mm/s. Considering an interval of 1−2 s (Figure 5B), one can calculate the delay between an action potential on one electrode and another electrode serving as a reference in time. To chemically stimulate the cells, norepinephrine was added to cell media with a concentration of 1:1000. Norepinephrine effects an increase of the contraction rate of the cardiomyocytes (Figure 6), resulting in a frequency of about 0.89 ± 0.17 Hz in

have more than a 90% effective guidance effect on HL-1 cells. Moreover, we were able to investigate the cell response to the pillar interspace with focused ion beam cross sectioning and scanning electron microscopy. These guidance results represent an improvement compared to results obtained with other techniques previously mentioned and in addition the 3D nanostructure-based guidance represents a novel technique for cell patterning. The mushroom-like structures are furthermore suitable for extracellular recordings of action potentials and combine effective cell guidance with electrophysiological investigations for cardiomyocyte-like cells. To the knowledge of the authors, this is one of the very few examples, potentially the first, shown in the literature where a defined nanostructure was used effectively and simultaneously for cell patterning and extracellular measurements.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: a.off[email protected]. Address: Leo Brandt Strasse 1, 52428, Jülich, Germany. Tel. Office: +492461612330. Fax: +49246161-8733. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We would like to acknowledge Marco Banzet for the help in the fabrication process. Unless otherwise noted, all the chemicals were supplied by Sigma Aldrich GmbH, Seelze, Germany.

■

REFERENCES

(1) Thomas, C. A., Jr.; Springer, P. A.; Loeb, G. E.; Berwald-Netter, Y.; Okun, L. M. Exp. Cell Res. 1972, 74, 61−66. (2) Gross, G. W.; Rieske, E.; Kreutzberg, G. W.; Meyer, A. Neurosci. Lett. 1977, 6, 101−105. (3) Pine, J. J. Neurosci. Methods 1980, 2, 19−31. (4) Choi, D. S.; Fung, A. O.; Moon, H.; Villareal, G.; Chen, Y.; Ho, D.; Presser, N.; Stupian, G.; Leung, M. J. Nanosci. Nanotechnol. 2009, 9, 6483−6486. (5) Kim, J.-H.; Kang, G.; Nam, Y.; Choi, Y.-K. Nanotechnology 2010, 21, 085303. (6) Wesche, M.; Hüske, M.; Yakushenko, A.; Brüggemann, D.; Mayer, D.; Offenhäusser, A. Nanotechnology 2012, 23, 495303. (7) Wang, K.; Fishman, H. A.; Dai, H.; Harris, J. S. Nano Lett. 2006, 6, 2043−2048. (8) Yakushenko, A.; Gong, Z.; Maybeck, V.; Hofmann, B.; Gu, E.; Dawson, M.; Offenhäusser, A.; Wolfrum, B. J. Biomed. Opt. 2013, 18, 111402−111402. (9) Verma, P.; Melosh, N. A. Appl. Phys. Lett. 2010, 97, 033704− 033704-3. (10) Weis, R.; Fromherz, P. Phys. Rev. E 1997, 55, 877−889. (11) Spira, M. E.; Hai, A. Nat. Nanotechnol. 2013, 8, 83−94. (12) Brüggemann, D.; Wolfrum, B.; Maybeck, V.; Mourzina, Y.; Jansen, M.; Offenhäusser, A. Nanotechnology 2011, 22, 265104. (13) Keefer, E. W.; Botterman, B. R.; Romero, M. I.; Rossi, A. F.; Gross, G. W. Nat. Nanotechnol. 2008, 3, 434−439. (14) Gabay, T.; Ben-David, M.; Kalifa, I.; Sorkin, R.; Abrams, Z. R.; Ben-Jacob, E.; Hanein, Y. Nanotechnology 2007, 18, 035201. (15) Hai, A.; Shappir, J.; Spira, M. E. Nat. Methods 2010, 7, 200−202. (16) Fendyur, A.; Spira, M. E. Front. Neuroeng. 2012, 5, 10.3389/ fneng.2012.00021. (17) Duan, X.; Gao, R.; Xie, P.; Cohen-Karni, T.; Qing, Q.; Choe, H. S.; Tian, B.; Jiang, X.; Lieber, C. M. Nat. Nanotechnol. 2012, 7, 174− 179. (18) Hai, A.; Shappir, J.; Spira, M. E. J. Neurophysiol. 2010, 104, 559− 568.

Figure 6. Action potentials recorded by 3D electrodes after chemical stimulation with noradrenaline during a time frame of 12 s.

our experiment. Earlier works demonstrated the application of nanostructured MEAs to be suitable for extracellular recordings from HL-1 cells in the millivolt regime.6,12 On the other hand, nanostructured MEA electrodes were successfully employed for the recording of action potentials in the 10 mV range from neuronal cell cultures.15 We show here the coupling of cardiomyocytes on nanostructured electrodes. The recorded signals from our gold nanopillars reached about 180 μV (peakto-peak) on some electrodes and were clearly distinguishable from the background electrode noise. Xie et al.41 also measured with their array of nanopillars (height 1.5 μm, diameter 150 nm) spontaneous activity from HL-1 cells in the range of about 100 μV (peak-to-peak) which is comparable with our results. In conclusion, our 3D gold mushroom-shaped nanostructures were shown to be an appropriate tool for chip-based simultaneous guidance and recording experiments. We were able to determine that nanopillars with a pitch from 3 to 7 μm 5383


Nano Letters

Letter

(19) Lu, Y.-B.; Franze, K.; Seifert, G.; Steinhäuser, C.; Kirchoff, F.; Wolburg, H.; Guck, J.; Janmey, P.; Wei, E.-Q.; Käs, J.; Reichenbacj, A. Proc. Natl. Acad. Sci. 2006, 103, 17759−17764. (20) Sniadecki, N. J.; Desai, R. A.; Ruiz, S. A.; Chen, C. S. Ann. Biomed. Eng. 2006, 34, 59−74. (21) Andersson, A.-S.; Bäkhed, F.; von Euler, A.; Richter-Dahlfors, A.; Sutherland, D.; Kasemo, B. Biomaterials 2003, 24, 3427−3436. (22) Braeken, D.; Huys, R.; Jans, D.; Loo, J.; Rand, D. R.; Borghs, G.; Callewaert, G.; Bartic, C. In World Congress Medical Physics and Biomedical Engineering, Munich, Germany, Sept. 7−12, 2009; Dössel, O., Schlegel, W. C., Eds.; Springer: Berlin, 2010; pp 212−215 and at http://link.springer.com/chapter/10.1007/978-3-642-03887-7_59. (23) Panaitov, G.; Thiery, S.; Hofmann, B.; Offenhäusser, A. Microelectron. Eng. 2011, 88, 1840−1844. (24) Hai, A.; Kamber, D.; Malkinson, G.; Erez, H.; Mazurski, N.; Shappir, J.; Spira, M. E. J. Neur. Eng. 2009, 6, 066009. (25) James, C. D.; Davis, R.; Meyer, M.; Turner, A.; Turner, S.; Withers, G.; Kam, L.; Banker, G.; Craighead, H.; Issacson, M.; Turner, J.; Shain, W. IEEE Trans. Biomed. Eng. 2000, 47, 17−21. (26) Von Philipsborn, A. C.; Lang, S.; Bernhard, A.; Loeschinger, J.; David, C.; Lehnert, D.; Bastmeyer, M.; Bonhoeffer, F. Nat. Protocols 2006, 1, 1322−1328. (27) Cheng, J.; Zhu, G.; Wu, L.; Du, X.; Zang, H.; Wolfrum, B.; Jin, Q.; Zhai, J.; Offenhäusser, A.; Xu, Y. J. Neurosci. Methods 2013, 213, 196−203. (28) Offenhäusser, A.; Böcker-Meffert, S.; Decker, T.; Helpenstein, R.; Gasteier, P.; Groll, J.; Möller, M.; Reska, A.; Schäfer, S.; Schulte, P.; Vogt-Eisele, A. Soft Matter 2007, 3, 290−298. (29) Yu, T. W.; Bargmann, C. I. Nat. Neurosci. 2001, 4, 1169−1176. (30) Au, T. H.; Cui, B.; Chu, Z. E.; Veres, T.; Radisic, M. Lab Chip 2009, 9, 564−575. (31) Wang, L.; Liu, L.; Magome, N.; Agladze, K.; Chen, Y. Biofabrication 2013, 5, 035013. (32) Park, J.; Kim, H.-N.; Kim, D.-H.; Levchenko, A.; Suh, K.-Y. IEEE Trans. NanoBiosci. 2012, 11, 28−36. (33) Hofmann, B.; Kätelhön, E.; Schottdorf, M.; Offenhäusser, A.; Wolfrum, B. Lab Chip 2011, 11, 1054−1058. (34) Schnitker, J.; Afanasenkau, D.; Wolfrum, B.; Offenhäusser, A. Phys. Status Solidi 2013, 210, 892−897. (35) Claycomb, W. C.; Lanson, N. A.; Stallworth, B. S.; Egeland, D. B.; Delcaprio, J. B.; Bahinski, A.; Izzo, N. J. Proc. Natl. Acad. Sci. 1998, 95, 2979−2984. (36) Turner, A. M. P.; Dowell, N.; Turner, S. W. P.; Kam, L.; Isaacson, M.; Turner, J. N.; Craighead, H. G.; Shain, W. J. Biomed. Mater. Res. 2000, 51, 430−441. (37) Kim, D.-H.; Lipke, E. A.; Kim, P.; Cheong, R.; Thompson, S.; Delannoy, M.; Suh, K.-Y.; Tung, L.; Levchenko, A. Proc. Natl. Acad. Sci. 2010, 107, 565−570. (38) Nikkhah, M.; Edalat, F.; Manoucheri, S.; Khademhosseini, A. Biomaterials 2012, 33, 5230−5246. (39) Ross, A. M.; Jiang, Z.; Bastmeyer, M.; Lahann, J. Small 2012, 8, 336−355. (40) Kolind, K.; Dolatshahi-Pirouz, A.; Lovmand, J.; Pedersen, F. S.; Foss, M.; Besenbacher, F. Biomaterials 2010, 31, 9182−9191. (41) Xie, C.; Lin, Z.; Hanson, L.; Cui, Y.; Cui, B. Nat. Nanotechnol. 2012, 7, 185−190.

5384


On Chip Guidance and Recording of ... - ACS Publications

Recommend Documents