Streamlining the interface between electronics and neural systems for bidirectional electrochemical communication

Abstract

Seamless neural interfaces conjoining neurons and electrochemical devices hold great potential for highly efficient signal transmission across neural systems and the external world. Signal transmission through chemical sensing and stimulation via electrochemistry is remarkable because communication occurs through the same chemical language of neurons. Emerging strategies based on synaptic interfaces, iontronics-based neuromodulation, and improvements in selective neurosensing techniques have been explored to achieve seamless integration and efficient neuro-electronics communication. Synaptic interfaces can directly exchange signals to and from neurons, in a similar manner to that of chemical synapses. Hydrogel-based iontronic chemical delivery devices are operationally compatible with neural systems for improved neuromodulation. In this perspective, we explore developments to improve the interface between neurons and electrodes by targeting neurons or sub-neuronal regions including synapses. Furthermore, recent progress in electrochemical neurosensing and iontronics-based chemical delivery is examined.

Fig. 1. Effect of surface topographical features on neuronal cultures. (a) Colored SEM images of the guided culture of neurons by the geometric patterning of a nanopillar coated with poly-dl-ornithine (PDLO). Scale bar: 4 μm. Adapted from ref. with permission from the American Chemical Society, copyright 2018. (b) SEM images of planar gold and nanoporous gold surfaces. Scale bar: 200 nm. Bar graph showing the surface coverage of neurons and astrocytes on various substrates. Adapted from ref. with permission from the American Chemical Society, copyright 2015. (c) Surface topography affects the cleft distance between the cell membrane and underlying surface. FIB-SEM cross-sectional image showing the wrapping of the plasma membrane around nanopillars. Scale bar: 250 nm. (d) Cleft distance measurements show a drastically smaller cleft distance with nanopillars than nanopores or flat surfaces. Adapted from ref. with permission from the American Chemical Society, copyright 2017.

Fig. 2. Effect of surface modification on cellular growth and in vivo implantation. (a) Beta-amino acid subunits to compose synthetic polymers. All subunits are racemic and the synthesized polymers are heterochiral. General synthesis of polymers with a chain length of 20 amino acid residues. (b) Live/dead assay for preosteoblast cells seeded on a bare PEG hydrogel and DM₅₀CO₅₀-modified PEG hydrogel for 2 days. Adapted from ref. with permission from Nature Publishing Group, copyright 2021. (c) Bar graph quantifying normalized neurite outgrowth. Mean ± s.d. (n = 12 trials). Statistical significance was determined with a two-way ANOVA and Tukey's post hoc **p < 0.01 ***p < 0.001 ****p < 0.0001. Adapted from ref. with permission from John Wiley and Sons, copyright 2021.

Fig. 3. Observations of intracellular signal propagation by targeting neurites (axons or dendrites). (a) Recording of activity by two distal (green and blue circles) and two proximal (black and red circles) microelectrodes from the position of laser-induced dissection of axons. The activity recovery was observed on distal electrodes for several weeks. Adapted from ref. with permission from the Royal Society of Chemistry, copyright 2015. (b) Single neuron cultured on a nanowire (NW) transistor array. Axon and dendrite are aligned oppositely by micropatterned poly-l-lysine. Adapted from ref. with permission from the American Association for the Advancement of Science, copyright 2006.

Fig. 4. Biohybrid strategies for the formation of the synaptic interface. (a) Schematic illustration of the implanted μTENN which is able to transmit signals with neurons bidirectionally into the brain. Optogenetically active μTENNs as transplantable input/output channels. Inputs: an LED array (1) optically stimulates a unidirectional, channelrhodopsin-positive μTENN (2) to activate layer IV neurons (3). Outputs: layer V neurons (4) are connected via synapses to bidirectional μTENN (5) and relayed neuronal activity is recorded by a photodiode array on the brain surface (6). (b) Phase image of a part of bidirectional GCaMP+ μTENN before implantation in the rodent cortex. Scale bar: 50 μm (left). Multiphoton image of the same μTENN obtained immediately after implantation. Scale bar: 20 μm (right). (c) in vivo recording of calcium concentration changes with (b). Each red and grey trace corresponds to the time course fluorescence intensity of each region of interest (ROI) and mean trace is marked black. Adapted from ref. with permission from the American Association for the Advancement of Science, copyright 2021. (d) Conceptual schematic of the induced presynapse interface. Presynaptic differentiation occurs at the contact between the neuronal membrane and the presynapse inducing protein functionalized substrate. (e) Fluorescence analysis of pre- and post-synaptic proteins under each bead condition. Intensities are compared between each bead condition. (37 < n < 68 for each condition. Statistical significance is indicated by n.s. for p > 0.05 and *** for p < 0.001). Adapted from ref. with permission from the American Chemical Society, copyright 2019. (f) Neuroglin1 (NL1-R), neuroligin2 (NL2-R) and Slitrks3 (SL3-R) coated beads induce the formation of glutamatergic and inhibitory presynaptic boutons, respectively. Representative confocal images of induced hemisynapses immunostained with synapsin1 (Syn1, red), vesicular glutamate transporter 1 (VGluT1, green), and vesicular γ-aminobutyric acid (GABA) transporter (VGAT, cyan). (g) Structures of genetically engineered synapse-inducing proteins modified on beads (Red: NL1-R and blue: SL3-R), which can induce excitatory or inhibitory synapses. Adapted from ref. with permission from Springer Nature, copyright 2016. (h) Poly-d-lysine (PDL)-beads induce the formation of synaptic vesicle complexes on axons. SEM (left) and transmission electron microscopy (TEM) (right) images of neurons cocultured for 24 h with PDL-beads (top) or uncoated beads (bottom). Scale bars: 1 μm (SEM) and 250 nm (TEM). Adapted from ref. with permission from Society for Neuroscience, copyright 2009. (i) Molecular interactions at a single (1) NLG1/Nrx1β contact, leading to the recruitment of postsynaptic proteins, including PSD-95 (yellow), and (2) SynCAM1/SynCAM contact, leading to the recruitment of presynaptic vesicles (red). Immunostained images of synaptic protein recruitment at coated micropatterns. Left image shows the Nrx1β-Fc coated dots (blue) and enrichment of NLG1 (red) and PSD-95 (green). Right image shows SynCAM1-Fc boated dots (blue) and accumulation of SynCAM1 (red) and synapsin1 (green). Adapted from ref. with permission from Nature Publishing Group, copyright 2013. (j) Scheme of eNL1 immobilized on an AuNP electrode. Structure of genetically engineered neuroligin-1, eNLG1 reconstructed after NLG1 (PDB ID, 3BIW) and TagRFP (PDB ID, 3NED). (k) Presynaptic differentiation of primary hippocampal neurons induced by the eNLG1 electrode. FE-SEM and confocal image of neurites immunostained with Syn1 (green) and PSD95 (cyan) on the eNLG1 electrode. Scale bar: 500 nm (top). Confocal images of neurites immunostained with Syn1 (green) and VGluT1 (red) on the eNLG1 electrode. Scale bar: 10 μm (bottom). (l) Displacement of somas and induced presynapses on the eNLG1 electrode observed over 7 days. Adapted from ref. with permission from the American Chemical Society, copyright 2021.

Fig. 5. Representative designing considerations for seamless neural probes. (a) In vivo monitoring of dissolved oxygen in the hippocampus of the rat brain with an electrografted silica nanoporous membrane (SNM) modified column-shaped carbon fiber microelectrode (CFME) every 30 minutes for 2 h. Adapted from ref. with permission from the American Chemical Society, copyright 2019. (b) Chemical structure of polytannic acid doped nanoporous conductive polyaniline (PTA-PANI). (c) The ratios of sensitivities by postcalibration of the PTA-PANI-coated CFME (black column) and bare CFME (red column) to that by precalibration in artificial cerebral spinal fluid before and after implant in the rat brain for 2 h. Adapted from ref. with permission from the American Chemical Society, copyright 2019. (d) Microglial surface coverage of poly(sulfobetaine methacrylate) (PSBMA) coated probe (blue bars) and non-coated probe (orange bars) surfaces (Two-way ANOVA and Tukey's post-hoc tests; ****p < 0.0001 for both 2 and 6 h). Adapted from ref. with permission from Wiley, copyright 2020. (e) Cyclic voltammogram obtained in 10 μM dopamine (DA) solution at the electrodeposited graphene oxide microband-CFME before (solid curve) and after (dotted curve) immersion into 20 mg mL⁻¹ BSA for 2 h. The inset shows energy dispersive X-ray analysis (EDX) of the electrodeposited graphene oxide microband-CFME. Au and graphene oxide are respectively colored yellow and blue. Adapted from ref. with permission from John Wiley and Sons, copyright 2021. (f) Three-dimensional reconstructed images of neuron-like electronics (NeuE)-neural tissue interface 3 months after implantation. (neuron (green) and NeuE (red)) Scale bar: 100 μm. Inset shows the structural similarity between neurons and NeuE. Adapted from ref. with permission from Springer Nature, copyright 2019. (g) Experimental scheme for a Pavlovian reward learning task in freely moving mice using NeuroString (top). Exemplar time-aligned DA signals from a mouse on day 1 (middle) and day 6 (bottom). Adapted from ref. with permission from Springer Nature, copyright 2022. (h) False-colored SEM image of cultured neuron on vertical nanoelectrode array. Scale bar: 4 μm. Adapted from ref. with permission from the American Chemical Society, copyright 2017.

Fig. 6. Representative electrochemical techniques for neural interfaces in recent studies. (a) Microscopy images of the insertion and withdrawal process of a conical carbon fiber nanoelectrode (CFNE) in the synaptic cleft in chronological order. Adapted from ref. with permission from the Royal Society of Chemistry, copyright 2020. (b) in vivo FSCV detection of dopamine in the nucleus accumbens by a three-channel NeuroString sensor during optogenetic stimulation (20 Hz with 15 pulses) of dopaminergic neurons in the ventral tegmental area. Adapted from ref. with permission from Springer Nature, copyright 2022. (c) Amperometric trace for a nanotip conical CFME placed inside a PC12 cell. Adapted from ref. with permission from John Wiley and Sons, copyright 2015. (d) Schematic illustration of the potentiometric sensor consisting of a voltmeter, an Ag/AgCl reference electrode, and a closed bipolar CFME which can oxidize ascorbate spontaneously. (e) In vivo potentiometric sensing of the concentration dynamics of ascorbate in the striatum upon locally injecting 100 μM glutamate (red curve) or a mixture of 100 μM glutamate and ascorbate oxidase (40 units mL⁻¹) at a rate of 2 μL min⁻¹ for 60 s (black curve). Adapted from ref. with permission from John Wiley and Sons, copyright 2020.

Fig. 7. Design of enzymatic and apta-sensors for the detection of neurochemicals. (a) Schematic diagram of an ultra-fast enzyme sensor for glutamate detection. Carbon fiber microelectrode was modified with gold nanoparticles (red) and a thin layer of glutamate oxidase (yellow). Glutamate is electrochemically detected by the sensor via reduction of the reporter molecule hydrogen peroxide. Averaged amperometric current–time trace for individual exocytosis events detected from acute brain slices of the mouse is shown. Adapted from ref. with permission from the American Chemical Society, copyright 2019. (b) Ribbon diagram of glucose oxidase monomer (from A. niger) with the FAD cofactor (blue) and immobilized gold nanoparticle (AuNP, yellow). Mutated amino acid residues are shown as space-filling models: cysteine (yellow), histidine (red), serine (purple), alanine (orange), tyrosine (pink), and glutamate (light blue). Binding of the AuNP to cysteine via covalent click chemistry is shown in the scheme. Adapted from ref. with permission from the American Chemical Society, copyright 2011. (c) Schematic illustration of the implantable aptamer-graphene microtransistor probe design and conformation change of the aptamer upon dopamine binding. (d) Representative response trace for the neural probe implanted in striatum after local perfusion with 60 mM KCl. VDS = 100 mV and VG = 50 mV. Adapted from ref. with permission from the American Chemical Society, copyright 2022.

Fig. 8. Chemical stimulation of neural systems via iontronic devices. (a) Potential profile across the palladium (Pd) based OEIP device. Proton and neurotransmitter GABA movement from the source to the target is shown. At certain potentials, the Pd electrode selectively absorbs protons to enable the passage of the drug to the target electrolyte. Adapted from ref. with permission from the American Association for the Advancement of Science, copyright 2021. (b) Schematic illustration of the working principle for a PGF micropipette and comparison with a conventional surface-charged nanopore. Inverted ion current rectification enhances the outward flow in the PGF micropipette for chemical delivery. Adapted from ref. with permission from the American Chemical Society, copyright 2021.

Wonkyung Cho

Sun-heui Yoon

Taek Dong Chung