Engineering Highly Interconnected Neuronal Networks on Nanowire

Apr 24, 2017 - National Fabrication Facility, Research School of Physics and Engineering, Australian National University, Canberra, ACT 2601,. Austral...
0 downloads 3 Views 2MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Communication

ENGINEERING HIGHLY INTERCONNECTED NEURONAL NETWORKS ON NANOWIRE SCAFFOLDS Vini Gautam, Shagufta Naureen, Naeem Shahid, Qian Gao, Yi Wang, David R. Nisbet, Chennupati Jagadish, and Vincent R. Daria Nano Lett., Just Accepted Manuscript • Publication Date (Web): 24 Apr 2017 Downloaded from http://pubs.acs.org on April 25, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

ENGINEERING HIGHLY INTERCONNECTED NEURONAL NETWORKS ON NANOWIRE SCAFFOLDS

Vini Gautam†,§,‡*, Shagufta Naureen§, Naeem Shahid¶, Qian Gao§, Yi Wang‡, David Nisbet‡, Chennupati Jagadish§* and Vincent R. Daria†* †

Eccles Institute of Neuroscience, John Curtin School of Medical Research, Australian National University, Canberra, ACT – 2601, Australia.

§

Department of Electronic Materials Engineering, Research School of Physics and Engineering, Australian National University, Canberra, ACT – 2601, Australia.



Laboratory of Advanced Biomaterials, Research School of Engineering, Australian National University, Canberra, ACT – 2601, Australia. ¶

Australian National Fabrication Facility, Research School of Physics and Engineering, Australian National University, Canberra, ACT – 2601, Australia.

ABSTRACT Identifying the specific role of physical guidance cues in the growth of neurons is crucial for understanding the fundamental biology of brain development and for designing scaffolds for tissue engineering. Here, we investigate the structural significance of nanoscale topographies as physical cues for neurite outgrowth and circuit formation by growing neurons on semiconductor nanowires. We monitored neurite growth using optical and scanning electron microscopy, and evaluated the spontaneous neuronal network activity using functional calcium imaging. We show, for the first time, that an isotropic arrangement of indium phosphide (InP) nanowires can serve as physical cue for guiding neurite growth and aid in forming a network with neighbouring neurons. Most importantly, we confirm that multiple neurons, with neurites

ACS Paragon Plus Environment

1

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 22

guided by the topography of the InP nanowire scaffolds, exhibit synchronized calcium

activity,

implying

intercellular

communications

via

synaptic

connections. Our study imparts new fundamental insights on the role of nanotopographical cues in the formation of functional neuronal circuits in the brain and will therefore advance the development of neuroprosthetic scaffolds.

KEYWORDS: Nanowires, neurons, scaffolds, calcium imaging, guided growth, topography

One of the major challenges in neuroscience is to uncover the cellular and molecular processes underlying the formation of neural circuits in the brain. The ordered assembly of neurons in the cerebral cortex and other regions in the brain suggest that guidance cues are highly operational at the onset of brain development. Chemical and physical cues in the extracellular matrix (ECM) guide axons and dendrites to target other cells and form synaptic interconnections1, 2. Thus, understanding the interplay between these cues could elucidate theories on brain development and functional circuit formation, in turn providing insights into the design of biomaterials and neuroprosthetic scaffolds. To understand such interaction, neurons can be grown on artificial substrates using pre-determined external chemical cues and surface parameters or physical cues. Forming pre-defined patterned neural networks allows direct access to studying response of individual cells and small neuronal networks to external cues and stimuli3, 4. Novel biomaterials and advanced fabrication techniques in materials science enable patterning of chemical and physical cues on artificial

ACS Paragon Plus Environment

2

Page 3 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

substrates to guide neuronal cell adhesion5, 6, outgrowth7 and proliferation8, 9. For instance, chemical components in the extracellular matrix such as laminin, fibronectin or neuron-adhesive glycoproteins can be deposited in a desired layout on the cell culture substrates using techniques such as microcontact printing and laser ablation10-13. Contact-mediated mechanisms in the cytoskeleton then result in directional growth of the axons and engineered connectivity between cells13,

14

. Alternatively, specific surface topographies

can be fabricated on substrates to induce neuronal adhesion as well as neurite alignment and guidance14-17. For example, topographical features such as grooves and pillars are typically designed on cell culture substrates using techniques such as photolithography and reactive ion etching, and it has been shown that axonal growth aligns with these patterns18-20. Such guided growth of neurites (axons and dendrites) is hypothesised to be mediated by various transmembrane proteins and mechanosensitive ion channels in the cytoskeleton21. For instance, integrin molecules bind to the ECM protein fibrils like collagen and laminin, in turn causing adaptor proteins to bind to actin in the cytoskeleton and forming focal adhesion complexes. These events initiate a range of mechanotransduction signalling pathways, which affect migration, proliferation and differentiation of cells in response to external forces and topography22. Furthermore, a few electrophysiological studies have evaluated synaptic connections of neurons grown on chemically and physically patterned substrates. For instance, electrophysiological recordings of neurons grown on substrates with polymer (e.g. lysine) patterns show that predefined axonal growth in one direction results in unidirectional signal propagation between neurons23 and limited dendritic structures alter

ACS Paragon Plus Environment

3

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 22

the firing properties of the constrained neurons24. Paired recordings of neurons on micron-scale linear patterns, such as grooves, have also been conducted and it was shown that such neurons form uni-, bidirectional and autaptic connections25. These prior studies, however, involve micron-scale patterns, constrained network morphologies and limited synaptic connections between neurons. To provide a better understanding of the external cues in the ECM that affect neuronal growth, it is important to isolate the effect of physical and chemical cues, and also allow neurons to grow unconstrained, enabling mechanosensing properties to function naturally. Moreover, the size of the protein fibres composing the ECM, such as fibrillar collagens and elastins, range from ten to several hundreds of nanometres in diameter26. Therefore it is crucial to scale the size of the physical cues to nanoscale in order to investigate the spatial dimensions sensed by the cells and for designing scaffolds for guiding the growth of neuronal networks Nanoscale structures such as nanowires have been used for positioning neurons27, directing neurite growth28-30, as well as for extracellular and intracellular recording through neural cell membranes31-34. Semiconducting nanowires are particularly important for their potential integration with electronic circuits for recording and stimulating cellular activity. Chemically etched high-density vertical Silicon (Si) nanowires of diameter ~ 120 nm have been shown to affect the morphology of retinal neurons35. On the other hand, epitaxially grown high-density Gallium Phosphide (GaP) semiconductor nanowires of diameters ~ 50 nm and height ~ 2.5 µm have been shown to penetrate neuronal cell membrane of peripheral neurons36 as well as support the long-term growth of retinal neurons37. GaP nanowires with similar

ACS Paragon Plus Environment

4

Page 5 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

structural parameters, densely arranged in rows 10 µm apart, were also used as ‘fences’ to guide axons of sensory neurons29 as well as those from the central nervous system30. These prior studies achieved guided neurite growth using random arrangement of nanowires composed either of elemental semiconductors such as Si or compound semiconductors such as GaP. The nanowires were used as a topographical feature to either promote neuronal growth or constrain the axonal outgrowth to achieve directionality. An isotropic arrangement of Si nanopillars (with diameters, d ranging from 0.5 - 2 µm and pitch, p = 1.5 µm) was also used to direct neurite growth18. In general, the collective aim of these prior studies is to guide the growth of neurons and their neurites to build effective biomaterials for neuroprosthetics. While these prior studies demonstrate directional growth of neurites, the functional connectivity and circuit activity of the patterned neurons have yet to be explored. Unlike other cells in the body, neurons establish synaptic connections and form functional circuits with other neurons. Hence, aside from optimizing highly directed neurite growth, it is important to understand the underlying principles that link neuronal morphology and circuit activity of the neuronal network with the nanowire array topology as a way of understanding the significance of biophysical cues in the brain. None of these prior studies have assessed neuronal circuit function on semiconducting nanowires. Here, we provide a deeper understanding of mechanisms underlying neuronal circuit formation by associating the nanowire topology with the cell’s mechanosensing properties and how nanowires influence neuronal cell morphology, circuit formation and function. First, we demonstrate that Indium Phosphide (InP, a III-V compound semiconductor) nanowire scaffolds support

ACS Paragon Plus Environment

5

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 22

neuronal growth and formation of highly connected neuronal circuits. Our selection of InP nanowires arises from their extensive use as building blocks for nanoscale electronic and optoelectronic applications38. Unlike Si and GaP, InP is a direct band-gap semiconductor, which could be utilized for a superior optoelectronic interface for neuronal stimulation. However, the biocompatibility of InP with neuronal cells is not known, and our work is the first demonstration that neural cells can grow on InP-based optoelectronic substrates. We use an isotropic arrangement of high aspect ratio InP nanowires to achieve directional neurite growth and to systematically analyse their influence on neuronal morphology and function. As the first characteristic of neuronal growth, we show that an InP nanowire (diameter, d ~ 200 nm) is an effective physical cue for guided growth of neurites. We then show that vertically aligned nanowires act as scaffolds, which guide neurite extension to grow along a periodic nanowire array. As the neurons extend their neurites and form circuit connections on these scaffolds, we then evaluate the neuronal activity of the cellular network by monitoring intracellular calcium (Ca2+) dynamics. Most importantly, we show that the Ca2+ activity between neurons on nanowire scaffolds is highly correlated, which implies an increased probability of forming synaptic connections between the neurons on the nanowires. We first optimized a top-down fabrication method to obtain high aspect ratio InP nanowires in a 200x200 µm lattice with a height, h = 2 µm and diameters varying from 100 nm < d < 800 nm (Supporting information, Figure S1). We also fabricated nanowire arrays in a square lattice with varying pitch from 2 µm < p < 10 µm while keeping d fixed (Supporting information, Figure

ACS Paragon Plus Environment

6

Page 7 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

S2). Epitaxially grown InP nanopillars were also fabricated for cell culture and the details of the fabrication and neuronal growth on these nanowires growth are discussed in the supporting information. The InP nanowires were mechanically stable during handling for cell culture preparation and in physiological conditions. We then isolated cells from brains of either postnatal rodent pups or from pre-natal embryos. Post-natal primary cultures last up to 21 days in vitro (21 DIV) and can form network connections with support cells (e.g. glia). On the other hand, embryonic primary cultures last less than 10 DIV and grow only neurons with no support cells. We plated the cultures on to the nanowire scaffolds and control substrates (plain InP substrates and glass coverslips), and characterized the cell growth using scanning electron microscopy (SEM) and immunocytochemistry (for details of the experimental procedures, see supporting information). Figure 1 shows SEM images of neurons on the nanowire arrays. Fig. 1a shows both the cell body and neurites of hippocampal neurons on the nanowires. On regions away from the soma, the neurites align with the nanowires (Fig. 1b). In embryonic cortical neuron cultures, the alignment of neurites can be observed as early as 2 DIV. A detailed view of the neurites shows the filopodia navigating their way through the nanowires (Fig. 1c, inset). In early stages of the culture, the neurites extend to a length crossing from one side of the nanowire region to the other (up to 200 µm), which indicates nanowire-induced polarization of the neurites. We further observed that the interconnection between neurites and ordering of the network occur preferentially at the region of the nanowires and predominately terminate at the edge of the nanowire region at 7 DIV (Fig. 1d). Figure 1e shows a SEM

ACS Paragon Plus Environment

7

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 22

image of post-natal primary cultures. At 21 DIV, we observe an increased density of the neurites on the nanowires and a termination of the neuronal network at the boundary between nanowires and smooth InP surface. These results indicate that the neural cells and their neurites exhibit an affinity towards the nanoscale surface topography provided by the nanowires. Figure 1f shows the ordering and alignment of the neurites along the square lattice arrangement of the nanowires after 5 DIV. This alignment is indicative of increased secondary branching of the neurites in response to isotropic nanotopographical cues provided by the nanowires. We also demonstrate the growth and support of neural cells on scaffolds of epitaxially grown InP nanowires (Supporting information, Figure S4). Our results showing growth of hippocampal and cortical neurons on InP nanowire arrays confirm that InP nanowires are non-cytotoxic, and InP nanowire patterns can be an effective physical cue to engineer the growth of neurites on 2D substrates.

ACS Paragon Plus Environment

8

Page 9 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 1. SEM images of the growth of neurons and neurites on nanowire arrays (a) A cell-body from hippocampal culture on the nanowire array, top view, after 20 DIV. (b) Another area on the same substrate as in (a), where neurites can be seen to grow along the nanowires, top view. (c) A neuron from cortical culture, after 2 DIV, inset 1 and inset 2 depict a closer view around the soma and the axon respectively. (d), (e) Cellular network on the edge of the nanowire array after 7 and 21 DIV, respectively. (f) Neurite growth on an area of nanowires after 5 DIV showing anchoring and secondary branching of neurites at the nanowires.

To investigate the degree of neurite guidance along the nanowires in relation with the number of days in culture, we bulk-labelled the cells with Cal520AM (Fig. 2a-c). From the acquired fluorescence images, we performed two-dimensional Fourier transform (2DFT) to quantify the degree of alignment of neurite growth. We first performed 2DFT on sections of the image where neurons are grown on nanowires (dashed squares). From the frequency (Fourier) space, we then took the values of the horizontal, fhoriz, and diagonal,

fdiag, frequency components. The fhoriz represents the spectrum describing the density of horizontal neurites while fdiag represents that of orthogonal diagonal neurites in the image. The ratio, fhoriz/fdiag hence quantifies the amount of horizontally aligned (guided) neurites with respect to diagonally aligned (unguided) neurites and Fig. 2d plots this ratio as a function of DIV. The plot shows that the density of aligned neurites increases as the neuronal circuits develop, consequently increasing the probability of neurites from different neurons to connect with each other. Further details of 2DFT analysis are discussed in the supporting information (Figure S3). Our results confirm the selective and preferential orientation of the neurites to grow along the isotropic arrangement of vertically aligned InP nanowires.

ACS Paragon Plus Environment

9

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 22

Figure 2. Fluorescence images of the neurons on nanowires arrays after (a) 4 DIV (b) 5 DIV and (c) 6 DIV. The data is shown for cortical cells grown on nanowires with pitch, p = 2 µm and diameter, d = 400 nm. The dashed square represents the area of the substrate with the nanowires. (d) Quantification of neurite ordering calculated from the ratio of the density of horizontally over diagonally aligned neurites plotted as a function of DIV.

We further observed a dependence of neurite guidance on the distance,

δ=p-d, between the edges of the nanowire pillars. We hypothesize that there is a finite range of mechanosensing at the filipodia, which influences neurite guiding between the pillars. To test this hypothesis, we cultured embryonic cortical neurons and imaged the cells after 5 DIV. The cells were grown on substrates with a fixed d = 200 nm and varying δ. Figure 3a and 3b shows the network of neurons grown on δ = 4.8 µm (p = 5 µm) and δ = 1.3 µm (p = 1.5 µm), respectively. We found that the neurite growth were more random and did not follow the nanowire scaffolds when δ = 4.8 µm. This effect could be observed during early stages of growth at 2 DIV (Supporting information, Figure S5). From Fig. 3b, we further hypothesize that the range of mechanosensing could be within δ~1.3µm and guidance is influenced with slight changes in δ. To test this, we cultured neurons on one substrate with multiple nanowire arrays with varying δ and a constant p = 2 µm. Figure 3c

ACS Paragon Plus Environment

10

Page 11 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

shows our results where neurites exhibit alignment along nanowires starting with δ = 1.8 µm. We observed that neurite guidance increased as the distance between nanowires is decreased to δ = 1.3 µm. We then quantified the guidance of the neurites by performing the 2DFT as described before. Figure 3d plots the ratio of fhoriz/fdiag as a function of a normalised parameter, δ/p. Included in the plot are ratios for Fig. 3a and Fig. 3b (triangles). This analysis shows that the neurites exhibit an increased ordering as the nanowire pillars are closely packed (higher degree of alignment for a low δ/p value). Our results demonstrate that the neuronal growth cones have a finite range of interaction with the nanowires, and this interaction decreases in strength as the nanowires are loosely packed.

Figure 3. Dependence of neurite growth on nanowire packing. Cortical neurons grown on nanowire array with pitch, (a) p = 5 µm and (b) p = 1.5µm (insets show SEM images of the corresponding nanowire arrays). (c) Cortical neurons (labelled with β-III-tubulin) on nanowire arrays with p = 2 µm and varying distance, δ, between nanowires. The dashed squares identify the areas with the nanowires. (d) Quantification of neurite ordering taken as a ratio between the number of horizontally aligned over diagonal neurites as a function of δ/p.

ACS Paragon Plus Environment

11

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 22

Previous studies on hippocampal cell growth on silicon substrates have also shown that neurite guidance on micropillar arrays occurs at p = 3 µm, and start to become random for p > 6 µm30. While it is important to note that range of mechanosensing falls within ~ 1.3 µm, our results further show that the neurites were found to freely extend either on the top or between the nanowires (Fig. 1c, inset 2). Prior works on guiding neurites using semiconducting nano- or micro-pillars have relied on constrained growth of the neurites. For example, the work by Hällström et. al.36 and Piret et. al.37 used randomly arranged nanowires of d = 80 nm at a high density (~ 4 – 10 nanowires/µm2) with δ < 500 nm thereby constraining cell bodies and neurites on top of the nanowires. On the other hand, Dowell-Mesfin et. al.18 used isotropically arranged micron-sized pillars (2 µm) with δ = 1.5 µm, resulting in a channelling effect where neurites grew along the base of the nanowires, i.e. at the surface of the substrate. In our work, the isotropic arrangement of vertically aligned high aspect-ratio nanowires of d = 200 nm to d = 800 nm and δ > 1.5 µm provides a non-constricting physical cue to the neurites resulting in a naturally occurring branching and growth. Most importantly, the growth of neurites is not constrained to either top or bottom of the nanowires but nonetheless maintains a pattern corresponding to the design of the nanowire scaffolds. Our results confirm that the size of the physical cue sensed by an extending neurite can be in the order of nanoscale (d ~ 200 nm) and guided growth of the neurites to form specific patterns can be achieved using nanowires. Our results provide initial insights into mechanotransduction pathways in the cell membrane that are activated by the isotropically arranged

ACS Paragon Plus Environment

12

Page 13 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

nanowires at an inter-pillar distance of δ ~ 1.5 µm, resulting in an increased directional outgrowth of the neurites. We have designed future experiments to study the influence of conductivity and mechanical properties of the nanowires on the mechanosensing properties of neurons and identify specific mechanotransduction pathways via the application of ion-channel blockers. We further evaluated the spontaneous activity of the neuronal network growing on the nanowire scaffolds using functional Ca2+ imaging (for details of the experimental procedure, see supporting information). Functional Ca2+ imaging is widely used for monitoring cellular activity in the brain both in vitro and in vivo. Ca2+ influx in neurons is part of the signalling mechanism and can be monitored using Ca2+-indicator dyes. Here, we used Cal520-AM, which has been shown to have a high signal-to-noise ratio and intracellular retention39. We monitored spontaneous Ca2+ activity in hippocampal cells on both glass coverslips (control substrates) and nanowire arrays by recording the fluorescence using a confocal microscope at 4 frames per second (∆t = 0.25 s). From the images, we calculated the change in fluorescence as,

∆F F(t) − Fbaseline = F Fbaseline

(1)

The glial cells play a crucial role in calcium signalling in neuronal cultures; hence we specifically selected the postnatal hippocampal cultures over the embryonic cortical cultures for these studies, as the embryonic cortical cultures lacked glial cells. We observed two types of Ca2+ signals: the slow waves (rise time > 2 s) from the glia and the fast transients from the neurons (rise time < 0.25 s) and glia (rise time < 0.5 s) during synaptic transmission. Cells grown on glass coverslips exhibit unguided neurites with random

ACS Paragon Plus Environment

13

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 22

network and Ca2+ activity appeared in distinct regions and at sparse intervals (Supporting Movie 1). For cells growing on plain InP substate without the nanowires, we observed similar Ca2+ activity in distinct regions but with more frequent activity from the glia (Supporting Movie 2). For the cells grown on nanowires, we observed bursts of transient activity synchronized throughout the cellular network (Supporting Movie 3). Such bursts of transients were not observed on control substrates. The Ca2+ signals in the form of ∆F/F traces can be clearly seen in a heat-map representation (Supporting Information, Figure S6). In order to quantify our observations in Ca2+ activity, we analyzed the spatial distribution of ∆F/F on both nanowire and control samples (Fig. 4) by evaluating the spatial correlation of ∆F/F on all three samples. We first divide each 512×512 pixel frame spatially into M×M blocks (for M = 64, each block = 8×8 pixels). Then, for each block, the average ∆F/F vs time trace is calculated and denoted as fij(t), where i,j (0,1,2,X, M) are indices associated to the spatial coordinates of the block. The correlation coefficient for each block, rij, is then calculated using: 2  M M K f ∆t ⋅ k ⋅ f ∆t ⋅ k − Kf ∆t ⋅ k ⋅ f ∆t ⋅ k ( ) ( ) ( ) ( ) ( ) 1 ij lm ij lm   rij2 = ∑ ∑∑ M × M  l m k  fij2 ( ∆t ⋅ k ) − Kfij2 ( ∆t ⋅ k ) ⋅  flm2 ( ∆t ⋅ k ) − Kflm2 ( ∆t ⋅ k )   

(2)

where K is the total number of frames. Figures 4d, 4e and 4f summarize the spatial distribution of rij calculated for the nanowires, plain InP substrate and glass, respectively. In order to associate neuronal structures (soma, dendrites, glia) with rij, we performed an image multiplication between Fig. 4a and 4d for nanowires, Fig. 4b and 4e for InP substrate and between Fig. 4c and 4f for

ACS Paragon Plus Environment

14

Page 15 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

glass, as shown in Figure 4g, 4h, and 4i respectively. It is evident that the Ca2+ activity of neurons grown on the InP nanowires was correlated over the whole neuronal network, with the rij values for the Ca2+ activity on nanowires (maximum rij = 0.085) being five-fold higher than that on InP substrate (maximum rij = 0.015) and glass (maximum rij = 0.004). From the identified neuronal structures giving high values of rij, we then identified the glia, dendrites and soma and plotted their ∆F/F traces (Fig. 4j, 4k, 4l). In the case of nanowires, the maximum magnitude of ∆F/F was observed on the somas and dendrites of the neurons. We further analyse the signalling between neurons and glia on the nanowire samples (Fig. 5). Our results show that the glial Ca2+ activity is synchronized when the burst of activity occurs in the neurons (Fig. 5b). This is consistent with the theory of reciprocal signalling between neurons and glia where neuronal activity can trigger Ca2+ increase in glia40. As seen in Fig. 5, we are able to pick up this bidirectional flow of information between neurons and glia during synaptic transmission. Our results further show that the rise time of the Ca2+ activity for cells on nanowires is less than 0.25 s or within two consecutive frames (Fig. 5b). Hence, we have reason to conclude that the bursts of activity between the neurons identified in Figure 4 are due to synaptic transmission.

ACS Paragon Plus Environment

15

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 22

Figure 4. Recording of Ca2+activity. Hippocampal neurons (labelled at 15 DIV with Cal-520AM) on (a) InP nanowires (p = 2 µm & δ = 1.6 µm), (b) plain InP substrate, and (c) plain glass. Spatial distribution of the correlation coefficient of the Ca2+ activity of neurons in (d) nanowires, (e) InP substrate, and (f) plain glass, plotted in log10 scale. Color scale for the correlation coefficient is shown below panel (f). Image multiplication (g)=aൈd, (h)=bൈe, and (i)=cൈf showing correlated neuronal structures. ∆F/F vs time from numbered regions marked in (g), (h) and (i) for neurons grown on (j) nanowires, (k) InP substrate and (l) glass, respectively. Scale bar = 50 µm in b and applies to a-f.

ACS Paragon Plus Environment

16

Page 17 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 5. Neuron-glia signalling on nanowire substrates. (a) A region on nanowires with selected neurons (1, 2) and glia (3). (b) A zoom-in view of the Ca2+ activity as ∆F/F traces from the selected areas (red circles) in neuron 1, 2 and glia 3. The solid lines are the average of the respective traces.

Designing nanoscale topographical features on cell culture substrates provides new insights into the significance of physical cues in the development and formation of neural circuits in the brain. This will lead to significant advancements in designing machine-tissue interfaces and ultimately the development of neuroprosthetics in regenerative medicine. In this study, we demonstrated that the growth of cells from hippocampal and cortical areas of rodent brains can be engineered in vitro using InP nanowires. We have shown that InP nanowires are non-cytotoxic, and InP nanowire patterns can be an effective physical cue to engineer the growth of neurites on 2D substrates. Neurites from different neurons align along the isotropic arrangement of vertically aligned InP nanowires thereby increasing the probability of synaptic connections between the neurons.

We have also

shown that there is a finite range of parameters (δ/p