pubs.acs.org/Langmuir © 2009 American Chemical Society
Rectifying and Sorting of Regenerating Axons by Free-Standing Nanowire Patterns: A Highway for Nerve Fibers :: :: Waldemar Hallstrom,† Christelle N. Prinz,*,† Dmitry Suyatin,† Lars Samuelson,† Lars Montelius,† and Martin Kanje‡ †
Division of Solid State Physics and ‡Department of Cell and Organism Biology, Lund University, Lund, Sweden Received February 4, 2009. Revised Manuscript Received March 3, 2009
We present an EBL-defined nanowire pattern that can sort axons coming from different directions on a substrate. The pattern defines tracks for left-bound traffic and right-bound traffic, which opens up new possibilities for designing neural networks on a chip.
Introduction Cell patterning and guidance of nerve cell fibers-axons is of crucial importance when building neural networks, 2D organs, and neural interfaces. The ability to control the connections between neurons by guiding their axons on a chip surface offers several advantages. Among them is the possibility to address axons from different types of neurons (e.g., motor neurons from sensory neurons). This is a prerequisite condition for bidirectional neural implants such as brain machine interfaces. Cell and axonal guidance can be achieved by using a variety of chemical and/or topographical modifications.1-14 We have recently shown that neurons can grow on free-standing nanowire substrates and that parallel rows of nanowires provide an excellent way of controlling cell growth and the guidance of regenerating axons.15,16 The rows of wires act as fences, preventing the nerve fibers from growing in between two nanowires. In this manner, axonal outgrowth *Corresponding author: Christelle Prinz. Email address: christelle.
[email protected] (1) Wyart, C.; Ybert, C.; Bourdieu, L.; Herr, C.; Prinz, C.; Chatenay, D. J. Neurosci. Methods 2002, 117, 123. (2) Matsuzawa, M.; Liesi, P; Knoll, W. J. Neurosci. Methods 1996, 69, 189. (3) Schmalenberg, K. E.; Uhrich, K. E. Biomaterials 2005, 26, 1423. (4) Oliva, A. A.Jr.; James, C. D.; Kingman, C. E.; Craighead, H. G.; Banker, G. A. Neurochem. Res. 2003, 28, 1639. (5) Gustavsson, P.; Johansson, F.; Kanje, M.; Wallman, L.; Linsmeier, C. E. Biomaterials 2007, 28, 1141. (6) Johansson, F.; Carlberg, P.; Danielsen, N.; Montelius, L.; Kanje, M. Biomaterials 2006, 27, 1251. (7) Rajnicek, A. M.; Britland, S.; McCaig, C. D. J. Cell. Sci. 1997, 110, 2905. (8) Dalby, M. J.; Riehle, M. O.; Sutherland; Agheli, H.; Curtis, A. S. G. Eur. Cells Mater. 2005, 9, 1. (9) Dalby, M. J.; Gadegaard, N.; Riehle, M. O.; Wilkinson, C. D.; Curtis, A. S. Int. J. Biochem. Cell Biol. 2004, 36, 2005. (10) Dowell-Mesfin, N. M.; Abdul-Karim, M.-A.; Turner, A. M. P.; Schanz, S.; Craighead, H. G.; Roysam, B.; Turner, J. N.; Shain, W. J. Neural Eng. 2004, 1, 78. (11) Bayliss, S. C.; Buckberry, L. D.; Fletcher, I.; Tobin, M. J. Sens. Actuators, A 1999, 74, 139. (12) Johansson, F.; Kanje, M.; Eriksson, C.; Wallman, L. Phys. Status Solidi C 2005, 9, 3258. (13) Rajnicek, A. M.; Foubister, L. E.; McCaig, C. D. Biomaterials 2008, 29, 2082. (14) Rebollar, E.; Frischauf, I.; Olbrich, M.; Peterbauer, T.; Hering, S.; Preiner, J.; Hinterdorfer, P.; Romanin, C.; Heitz, J. Biomaterials 2008, 29, 1796. :: :: (15) Hallstrom, W.; Martensson, T.; Prinz, C.; Gustavsson, P.; Montelius, L.; Samuelson, L.; Kanje, M. Nano Lett. 2007, 7, 2960. :: :: (16) Prinz, C.; Hallstrom, W.; Martensson, T.; Samuelson, L.; Montelius, L.; Kanje, M. Nanotechnology 2008, 19, 345101.
Langmuir 2009, 25(8), 4343–4346
can be confined in between two rows of nanowires. The ultrasmall radius of the wires prohibits axon climbing because the growth cone always encounters the wires at sharp angles (90°), in contrast to microstructured walls where fibers can reach the top of the wall by climbing at an intermediate angle. We have also shown that focal adhesions of the nerve fibers form specifically at the nanowires. Compared to walls, grooves, and channels, the nanowire fences offer a higher degree of guidance. They also have a very open architecture, allowing a free flow of nutrients and oxygen and avoiding deficiencies that might occur in solid channels. So far, this technique has allowed us to guide axons but not to differentiate between axons growing in different locations. For example, if axons coming from two different places on a chip are mixed, then they cannot be sorted once they have entered the rows of nanowires. Here we show that patterns of nanowires can be used to rectify axonal outgrowth and demonstrate that axons from two different populations can be fully separated, thus creating the possibility to address two populations of axons on a chip surface. These results, together with our earlier findings, provide a basis for the advanced control of neuronal growth on a chip, where a large range of functionalities can be implemented, including chemical sensors and electrodes to investigate neuronal function at high temporal and spatial resolution.17
Results and Discussion The nanowire pattern is shown in Figure 1. The patterned surface was first incubated with laminin, a basal lamina protein that promotes axonal outgrowth.18 Laminin adsorbs on the nanowires, making it possible to visualize the nanowires by immunostaining. Figure 1a is a fluorescence microscopy image of the nanowire pattern, showing arrays of short nanowire rows (10 μm long) oriented at 30° compared to the expected axonal growth direction. The periodicity of the rows is 8.7 μm in the direction parallel to the growth and 15 μm in the direction perpendicular to the growth. The pattern is designed so that the axons should grow in alternating directions with a separation of 5 μm between each growth :: :: (17) Johansson, F.; Hallstrom, W.; Gustavsson, P.; Wallman, L.; Prinz, C.; Montelius, L.; Kanje, M.2008, J. Vac. Sci. Technol., B, 26, 2558. (18) Powell, S. K.; Kleinman, H. K. Int. J. Biochem. Cell Biol. 1997, 29, 401.
Published on Web 3/18/2009
DOI: 10.1021/la900436e
4343
Letter
Figure 1. Laser scanning microscopy of fluorescently labeled nanowire rectifiers. (a) Wide-field image of the pattern. White arrows indicates the presumed growth directions in two consecutive tracks. Scale bar 25 μm. (b) Three-dimensional reconstruction from a z stack showing the short rows of nanowires imaged at 45°. Scale bar 2 μm. track/highway (white arrows). Figure 1b is a 3D confocal close-up image of the short rows of 4-μm-high, 75-nmdiameter nanowires spaced by 400 nm. We tested the nanowire pattern on axonal outgrowth by mounting a ganglion (a collection of nerve cells) next to the pattern. We used the superior cervical ganglion (SCG) from a green fluorescent protein (GFP) mouse. This ganglion contains postganglionic sympathetic neurons that innervate autonomic structures. It is involved in stress responses such as fight-or-flight reactions. We investigated the outgrowth properties after 7 days of culturing. During this period, axons grow out of the ganglion onto the nanowire-patterned surface. The higher the nanowire length, the larger the number of axons that can be guided and rectified. A length of 4 μm was chosen as a safe value to ensure that all of the axons from an SCG ganglion could be guided. Figure 2 shows axonal outgrowth from one GFP-SCG on the rectifier pattern. The axons are green, and the nanowires have been stained blue for laminin. The ganglion was mounted to the right of the image. The axons are growing toward the left in every other track, leaving the highways for the right-bound traffic empty. The inset shows the preparation at lower magnification. Here the axons have been immunostained red for anti-β-tubulin III. Figure 2 c shows the ratio of the fluorescence in pairs of adjacent tracks at different distances from the ECM-gel border. Within the ECM gel (at -100 μm), the fluorescence ratio is around 50%, and no guidance/rectification has taken place because the axons are growing inside the gel and have not encountered the nanowire pattern. At 100 μm from the gel border, 93% of the fluorescence is in the correct track. The ratio is close to 100% after 300 μm. In the next experiment, we mounted ganglia from a different mice breed on the opposite side of the nanowire pattern: an SCG from a GFP mouse at one end and an SCG from a wild-type mouse (not naturally fluorescent) at the other end of the pattern. After 7 days of regeneration, the culture was stained red for anti-β-tubulin, which stains all nerve fibers indiscriminately. The results are shown in Figure 3. Figure 3a shows a fiber bundle of GFP-expressing neurons coming from the ganglion mounted on the right side of the pattern. The nanowires are stained blue. The axons occupy only every other track. Figure 3b shows a composite image of green, blue, and red emissions, where all axons are stained red. The wild-type ganglion (red only) was mounted to the left of the pattern. The GFP type is still visible and appears yellow because of double staining. The blue stains visible outside the nanowire pattern correspond most likely to aggregates of laminin or drops of the ECM gel that is used to 4344
DOI: 10.1021/la900436e
Figure 2. Axonal outgrowth on rectifying nanowire patterns (ganglion mounted to the right). (a) GFP-expressing nerve fibers exclusively following the intended tracks. Scale bar 25 μm. (b) Overview of immuno-labeled (β-tubulin) nerve fibers near the ganglion. The rectification starts immediately as the fibers leave the ECM gel and enter the nanowire tracks. Scale bar 200 μm. (c) Statistical analysis of the outgrowth in b. The graph shows the ratio of relative fluorescence intensity from pairs of adjacent tracks (together with the standard error) at different distances from the ECM gel border. The ECM border (distance 0 μm) is outlined by a thin line in panel b. mount the ganglia on the substrate (which is mostly composed of laminin). Figure 4 shows 3D confocal images of axonal growth at low and high densities on the nanowire pattern. Figure 4a represents a cross section of tracks with excessive fiber growth. Note that the height of the nerve fiber bundles is higher than the nanowires; even so, track crossing does not occur. Figure 4b shows nerve fibers at low density. The axons adhere to the innermost wires, resulting in two thin bundles of fibers growing on the edges of the track. There are two points worth mentioning in this context. The first one is that the nanowire pattern influences axonal Langmuir 2009, 25(8), 4343–4346
Letter
Figure 3. Confocal microscopy of bidirectionally rectified axonal outgrowth. (a) Nerve fibers from the GFP-expressing neurons from a ganglion mounted to the right. (b) The same image now showing the result of anti β-tubulin-labeled nerve fibers. The new (red) fibers originate from the wild-type ganglion mounted to the left. Because all fibers stain red, the GFP fibers now appear yellow. Scale bars 25 μm.
Figure 4. Three-dimensional reconstructions of β-tubulin-labeled axons in the nanowire tracks. (a) Nerve fibers forming dense bundles, exceeding the nanowire height. Note that the nerve fibers do not cross the nanowires. (b) Three-dimensional view showing axonal guidance at lower nerve fiber density. The nerve fibers grow in contact with the nanowires and form two thin bundles but stay strictly within the growth tracks. Scale bars 10 μm. outgrowth well above the height of the nanowires. This is most likely due to axon-axon interactions and the tendency of axons to fasciculate (bundle), a phenomenon inherent to axons.19 The second one is that the guidance properties of the pattern are very powerful because one can observe both high and low nerve fiber densities on the same pattern without track crossing. It is also interesting how such very thin structures as nanowires can exert an influence on the much larger fiber bundles. This is demonstrated in Figure 5, which shows a scanning electron micrograph of axon bundles on the nanowire pattern, imaged at a 45° tilt. Note how thin and flexible the wires are, yet they can control the growth of the much thicker axon bundles.
Materials and Methods Nanowire Substrates. The nanowires were grown by metalorganic vapor-phase epitaxy (MOVPE) from Au growth seeds defined with electron beam lithography (EBL) on (111)B gallium phosphide (GaP) substrates.20,21 The EBL pattern was made on resist PMMA950A5 using single pixel mode and was followed by development in 1:3 MIBK/IPA. Au was then deposited by thermal evaporation (200 A˚). The nanowire positions were predefined in short rows, each consisting of 25 wires (19) Cervello, M.; Lemmon, L.; Landreth, G.; Rutishauser, U. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 10548. :: (20) Martensson, T.; Borgstrom, M.; Seifert, W.; Ohlsson, B J.; Samuelson, L. Nanotechnology 2003, 14, 1255. :: (21) Seifert, W.; Borgstrom, M.; Deppert, K.; Dick, K. A.; Johansson, J.; :: Larsson, M. W.; Martensson, T.; Skold, N.; Svensson, C. P. T.; Wacaser, B. A.; Wallenberg, L. R.; Samuelson, L. J. Cryst. Growth 2004, 272, 211.
Langmuir 2009, 25(8), 4343–4346
Figure 5. Scanning electron micrographs of sympathetic nerve fibers on rectifying nanowire patterns. (a) Overview showing fiber bundles of different thicknesses guided and rectified by the nanowires. Scale bar 50 μm. (b) Close-up of one fiber bundle. Note how extremely thin the nanowires are compared to the fiber bundle, yet they are able to control the outgrowth with high precision. Scale bar 10 μm. spaced 400 nm apart. The entire patterns were 2 mm long in the intended axonal growth direction and 2-4 mm wide. After liftoff in hot acetone, the samples were transferred to a growth chamber (Aixtron). The temperature for nanowire growth was 470 °C, and wire growth was initiated by supplying Ga(CH3)3 in addition to PH3. The growth time, temperature, and gas pressure all affect the wire length, which in this case was controlled to around 4 μm. The nanowire diameter (determined by the gold nanoparticle size) was 75 nm. Precursor molar fractions were 4.3 10-6 and 8.5 10-2 for Ga(CH3)3 and PH3, respectively, in a hydrogen carrier gas flow of 6 L min-1. The growth was conducted under low pressure (10 kPa), and the nanowire growth time was on the order of a few minutes. The resulting wires, grown in the [111] B direction (vertical on the surface), are hexagonal in cross section and very slightly tapered (the base is broadened). The nanowire substrates were incubated in a 10 μg 3 mL-1 laminin (Sigma) solution in phosphate-buffered saline (PBS) for 30 min at 37 °C, followed by rigorous rinsing in PBS. Cell Culture. All animal-related procedures were conducted with the permission of the Local Animal Ethics Committee at Lund University. We used SCGs from a wild-type mouse :: (NMRI, Mollegard, Denmark) and from a transgenic mouse expressing GFP (Okabe). The ganglia were removed by dissection and rinsed in RPMI 1640 medium. An extracellular matrix (ECM) gel (Sigma) was used to glue the ganglia onto the substrate. To this end, the ganglia were dipped in the ECM gel and then mounted on opposite sides of the pattern. Dipping the ganglia in ECM gel instead of adding a drop of gel to the ganglion reduces the amount of ECM gel spreading on the surface. Excessive gel on nanowire patterns creates a 3D matrix for the nerve fiber growth, preventing the axons from sensing the surface. The ECM gel was allowed to polymerize for 5 min at 37 °C. Thereafter, 2 mL of RPMI 1640 medium (with penicillin and streptomycin) containing 10% fetal calf serum, 1% ECM gel, and 42 ng 3 mL-1 nerve growth factor was added to each culture. The preparations were cultured for 7 days at 37 °C in a carbogen atmosphere (95% O2, 5% CO2). After this period, the cell culture were rinsed three times with PBS (37 °C) and then fixed in Stefanini’s fixative (4% paraformaldehyde, 0.03% saturated picric acid in 0.1 M phosphate buffer, pH 7.2) for 30 min. The sample was then washed seven times in PBS before further processing. Immunostaining. The preparations were incubated at 37 °C for 30 min in 1:200 rabbit/antilaminin IgG (Sigma) and 1:200 mouse/anti β-tubulin III IgG (Sigma) in PBS containing 0.25% Triton X100 and 0.25% BSA. After being rinsed seven times in PBS, the samples were incubated with 1:200 goat anti mouse Alexa Fluor 594 IgG (Invitrogen) and 1:100 goat anti rabbit cascade blue IgG (Invitrogen) in PBS containing 0.25% triton DOI: 10.1021/la900436e
4345
Letter X100 and 0.25% BSA. The samples were rinsed seven times in PBS, and the ganglia were gently removed with a pair of forceps before mounting for microscopy. Laser Scanning Confocal Microscopy. The fluorescently labeled samples were placed upside down in a PBS-filled glassbottomed cell-culture dish (Mattek) and analyzed with laser scanning confocal microscopy (Zeiss LSM510) with an oil-immersion objective (63, NA1.4). All data were obtained using the multitrack mode to eliminate channel cross talk and bleeding. Scanning Electron Microcopy. The immunostained preparations were dehydrated in a series of ethanol solutions and subjected to critical-point drying (Bal-Tech CPD 030). Thereafter, they were sputter coated with 10 nm of palladium/gold (VG microtech SC7640). The metal was sputtered on both sides of the sample to reduce charging during scanning electron microscopy. The scanning electron microscopy study was performed with a FEI Nova NanoLab 600. Statistics. The relative degree of rectification of the axons was analyzed in the case where only one ganglia was cultured on the substrate, in other words, where every other track is expected to be left empty. The total fluorescence value of each track was compared to the fluorescence of its adjacent track along a 500-μm-long (16 double tracks) singel-pixel line. The percentage
4346
DOI: 10.1021/la900436e
of fluorescence belonging to the correct track was calculated along with the standard error. The same calculation was repeated every 100 μm along the growth direction, from 100 μm before to 900 μm after the ECM gel border.
Conclusions Patterns of nanowires can be used to guide and rectify axonal outgrowth in a hitherto unprecedented manner. By creating a rectifier pattern, it is possible to separate nerve fibers coming from different places on a substrate. This opens up new possibilities to address nerve cell processes from different sensory modalities on the same chip or to differentiate between motor nerve fibers and sensory nerve fibers. Acknowledgment. This work was performed within the Nanometer Structure Consortium and the Department of Cell and Organism Biology at Lund University and was financed by the Knut and Alice Wallenberg Foundation, the Swedish Research Council (VR), and the Swedish Foundation for Strategic Research (SSF). The authors thank Peter :: Ekstrom, Rita Wallen, and Inger Antonsson for their help.
Langmuir 2009, 25(8), 4343–4346
