ORIGINAL RESEARCH published: 10 February 2021 doi: 10.3389/fbioe.2021.615218 All-Polymer Printed Low-Cost Regenerative Nerve Cuff Electrodes Laura M. Ferrari 1,2,3† , Bruno Rodríguez-Meana 4† , Alberto Bonisoli 1,2† , Annarita Cutrone 2 , Silvestro Micera 2,5 , Xavier Navarro 4 , Francesco Greco 1,6,7* and Jaume del Valle 4* 1 Center for Micro-BioRobotics @SSSA, Istituto Italiano di Tecnologia, Pontedera, Italy, 2 The BioRobotics Institute and Department of Excellence in Robotics and AI, Scuola Superiore Sant’Anna, Pontedera, Italy, 3 Université Côte d’Azur, INRIA, Sophia Antipolis, France, 4 Department of Cell Biology, Physiology and Immunology, Institute of Neurosciences, Universitat Autònoma de Barcelona, and CIBERNED, Bellaterra, Spain, 5 Bertarelli Foundation Chair in Translational NeuroEngineering, Center for Neuroprosthetics and Institute of Bioengineering, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland, 6 Institute of Solid State Physics, NAWI Graz, Graz University of Technology, Graz, Austria, 7 Department of Life Science and Medical Bioscience, Graduate School of Advanced Science and Engineering, Waseda University, Tokyo, Japan Neural regeneration after lesions is still limited by several factors and new technologies are developed to address this issue. Here, we present and test in animal models Edited by: a new regenerative nerve cuff electrode (RnCE). It is based on a novel low-cost Diego Mantovani, fabrication strategy, called “Print and Shrink”, which combines the inkjet printing of Laval University, Canada a conducting polymer with a heat-shrinkable polymer substrate for the development Reviewed by: Tzu-Wei Wang, of a bioelectronic interface. This method allows to produce miniaturized regenerative National Tsing Hua University, Taiwan cuff electrodes without the use of cleanroom facilities and vacuum based deposition Lorenza Draghi, Politecnico di Milano, Italy methods, thus highly reducing the production costs. To fully proof the electrodes *Correspondence: performance in vivo we assessed functional recovery and adequacy to support axonal Francesco Greco regeneration after section of rat sciatic nerves and repair with RnCE. We investigated francesco.greco@tugraz.at the possibility to stimulate the nerve to activate different muscles, both in acute and Jaume del Valle jaume.delvalle@uab.cat chronic scenarios. Three months after implantation, RnCEs were able to stimulate † These regenerated motor axons and induce a muscular response. The capability to produce authors have contributed equally to this work fully-transparent nerve interfaces provided with polymeric microelectrodes through a cost-effective manufacturing process is an unexplored approach in neuroprosthesis field. Specialty section: Our findings pave the way to the development of new and more usable technologies for This article was submitted to Tissue Engineering and Regenerative nerve regeneration and neuromodulation. Medicine, Keywords: regenerative cuff electrodes, low-cost fabrication, inkjet printing, wrinkling, organic bioelectronics, a section of the journal PEDOT:PSS, peripheral nerve interfaces Frontiers in Bioengineering and Biotechnology Received: 08 October 2020 INTRODUCTION Accepted: 11 January 2021 Published: 10 February 2021 Injuries to the peripheral nervous system (PNS) result in the partial or total loss of the motor, Citation: sensory and autonomic functions of the body part innervated by the lesioned nerve. The quality Ferrari LM, Rodríguez-Meana B, of life of people who suffer from these conditions is significantly reduced and major social Bonisoli A, Cutrone A, Micera S, consequences are payed in terms of health-care (Rosberg et al., 2005). When the whole nerve Navarro X, Greco F and del Valle J (2021) All-Polymer Printed Low-Cost is transected and the two nerve ends cannot be surgically rejoined, an autologous nerve graft Regenerative Nerve Cuff Electrodes. is commonly interposed in clinical practice to support axons regeneration. Nerve conduits have Front. Bioeng. Biotechnol. 9:615218. been extensively reported as a clinical alternative to the autograft repair (Deumens et al., 2010). doi: 10.3389/fbioe.2021.615218 However, even with surgical repair, axonal regeneration may fail or be insufficient to allow Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 1 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes functional reinnervation of targets and recovery of all motor and electronics, sensors and cell cultures, among others (Chen et al., sensory functions. Prostheses and exoskeletons are therapeutic 2011; Greco et al., 2012). One of the possible approaches to obtain strategies used to improve patients’ quality of life after severe surface wrinkling relies on the use of heat-shrink thermoplastic nerve injuries or limb amputations. For the application of substrates for the 3D patterning of thin films deposited onto neuroprosthesis, peripheral nerve interfaces (PNIs) are a key them. By thermally inducing the in-plane shrinkage of the element as they provide a link between the nervous system and substrate, the top-deposited thin film buckles, resulting in a the mechanical device (del Valle and Navarro, 2013). conformal wrinkled skin (Greco et al., 2013; Bonisoli et al., 2017). Typically, PNIs are produced through state-of-the-art Concurrently, the heat-shrinking provides a suitable method microfabrication and micromachining processes. Standard for miniaturization (down to e.g., 10% of original size) of any lithographic techniques, characterized by expensive high pattern (i.e., circuit, electrode), while retaining the pristine large vacuum and high temperature processes, are commonly and surface area (Gabardo et al., 2017; Chan et al., 2018). In this efficiently used for the last 20 years (Stieglitz and Meyer, way, the charge injection capacity of the electrode per unit area 1999; Boretius et al., 2010). Together with the drawbacks of can be increased, which is an attractive feature especially for cleanroom processes, some of the traditional metallic materials stimulating electrodes (Green and Abidian, 2015). Moreover, adopted (e.g., Au, Pt) have limitations in interfacing the soft electrodes miniaturization has positive effects on improving biological tissues (Bellamkonda et al., 2012). In the last decade, signal resolution during neural recording (Lacour et al., 2016). organic electronic materials (mainly conducting polymers and Here we propose a novel non-conventional microfabrication semiconductors) have shown their potential in applications technique to produce low-cost all-polymer PNIs. The produced where electronics interfaces biology. Given their soft mechanical design, named regenerative nerve cuff electrode (RnCE), consists properties and mixed conductivity (ionic and electronic) (Rivnay in a flexible plastic tube embedding conducting polymer et al., 2013), they proved to be an optimal bioelectronic interface, microelectrodes facing the lumen. Miniaturized 3D textured as already demonstrated in a variety of applications (Berggren electrodes are produced by the combination of inkjet printing and Richter-Dahlfors, 2007; Khodagholy et al., 2013; Simon and shrink-induced surface wrinkling (Print and Shrink). To et al., 2016). Organic materials represent the new frontier this aim, a heat-shrinkable polyolefin (PO) wrap film is used for PNIs development, since they can enhance signal quality, as low-cost, optically transparent and flexible substrate. Heat- biocompatibility and chronic reliability (Bettinger, 2018). shrink PO films are commonly adopted in food packaging and Particularly, organic materials have shown an increased charge have been proposed as biocompatible substrates for cell culturing injection capacity compared to standard metallic materials (Chen et al., 2011; Lew et al., 2011; Sharma et al., 2011), (Green et al., 2013), thus enabling the miniaturization of as well as molds for PDMS substrates in similar applications electrode arrays (Bettinger, 2018). (Nguyen et al., 2009; Chen et al., 2014; Shum et al., 2017). The One of the benefits of adopting organic electronic materials is materials here adopted for the electrodes patterning are poly(3,4- that they can be suited for printing. Inkjet printing is a large-scale ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) and low-cost production process. It is a non-contact additive and SU-8. PEDOT:PSS, the most used conducting polymer manufacturing method that operates in ambient conditions, in bioelectronics, is commercially available as waterborne at low temperatures. This allows the adoption of flexible dispersion and it has shown excellent chemical-physical stability substrates, with limited stability at high temperature, such as and biocompatibility (Groenendaal et al., 2000; Bernards thermoplastic foils often exploited in flexible/printed electronics et al., 2008; Malhotra and Ali, 2017). Biocompatibility and (Wong and Salleo, 2009; Caironi and Noh, 2015). The technique stability of various printable PEDOT:PSS formulations have been is suited for the deposition of various functional materials, as investigated, revealing the crucial role of thermal annealing conductive polymers (CPs), organic semiconductors, polymers and additives (Stríteský et al., 2018). In implantable devices, used for dielectric/passivation, which are solution processable PEDOT:PSS has shown to induce lower foreign body reaction (Berggren et al., 2007; Elschner et al., 2010). However, because (FBR) over other polymers (Cellot et al., 2016). Moreover, the of intrinsic limitations in resolution, inkjet printing cannot reach use of CPs as coatings for neural electrodes was reported to lower the miniaturization level required in some applications and the impedance at the electrode-tissue interface as compared to addressable by cleanroom processing. Typically inkjet printing the bare metal electrode (Asplund et al., 2010), thus enhancing of features with lateral size smaller than ∼100 µm can be the charge transfer and downplaying the electrical insulation challenging, at least with commercially available facilities. On effect due to the encapsulation of the implant. SU-8, an epoxy the other hand, there is a growing interest in adopting self- resin-based dielectric, is used for electrical passivation, given its assembling methods for surface patterning, especially on a large biocompatibility (Nemani et al., 2013) and the widespread use in area. Surface wrinkling is a self-assembling phenomenon which neural probes applications (Cho et al., 2008; Altuna et al., 2013; has been proposed as a rapid and convenient method for Márton et al., 2020). surface patterning (Genzer and Groenewold, 2006; Rodríguez- One of the main challenges for regenerative PNIs is the Hernández, 2015). It allows for tunable texturing with quasi production of non-obtrusive interfaces, defined as transparent periodic topographic motifs (with features size ranging from interfaces, still retaining a good degree of stimulation selectivity. tens-hundreds of nm up to mm scale) over very large area (up Previous works remarked the effort in the assessment of to several m2 ). Wrinkled surfaces of various materials have been transparent electrodes, as the regenerative multielectrode successfully tested in various applications, such as stretchable interface (REMI) (Garde et al., 2009), the regenerative scaffold Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 2 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes electrode (RSE) (Clements et al., 2012) and more recently before each printing step with the following settings: 50 s, 10 W the double-aisle regenerative electrode (Delgado-Martínez before printing of PEDOT:PSS and Ag ink; 100 s, 10 W before et al., 2017), comprising a silicone tube with a polyimide foil SU-8 printing. interposed in between and gold microelectrodes located on The deposition started with the patterning of the electrodes, the two opposing sides of the foil. The latter was developed by the feed lines and the contact pads by printing of the exploiting the same manufacturing process used for intraneural PEDOT:PSS-based ink with 30 µm ds (Figure 1A). Ag ink was PNIs, such as the thin film longitudinal intrafascicular electrode then deposited with 40 µm ds onto the PEDOT:PSS terminal (tf-LIFE) (Lago et al., 2007) or the transversal intrafascicular pads (Figure 1B). After drying overnight at room temperature, multielectrode (TIME) (Boretius et al., 2010), and its design the dielectric layer was printed. Firstly, the SU-8 10% was printed allowed unobstructed nerve regeneration, further enabling an with 70 µm ds [SU-8 (1), Figure 1C], dried by heating (95◦ C, interfacing with the axons. Despite the plurality of advantages 60 s), cured for cross-linking by UV exposure (365 nm, 410 of the design, the techniques used for the manufacturing of mW cm-2, 30 s) with a spot light source (Lightningcure LC8; the above-mentioned devices are very expensive and require a Hamamatsu) and then heated again (95◦ C, 60 s). During the cleanroom environment. thermal processes, in order to avoid its shrinkage, the substrate In this article we describe the RnCEs low-cost fabrication was accurately fixed to the glass slide by using extra Kapton tape process and their performance as a regenerative nerve guide and and metal clips, three clips for long sides, one clip for short PNI in vivo. Firstly, we assessed the RnCEs design as permissive sides. A layer of SU-8 29% was then printed, with 50 µm ds, to for axonal regeneration across a nerve gap by using passive efficiently insulate the conductive traces [SU-8 (2), Figure 1D]. (i.e., without polymeric microelectrodes) PEDOT:PSS coated PO Shortly after the deposition, the tape was removed and the PO tubes (PPP tubes). Once the adequacy for axonal regeneration was fixed to the glass slide only by clamping the short sides has been confirmed, the active RnCEs, with patterned polymeric with metal clips (one clip for each). Then, by heating at 110◦ C microelectrodes, was implanted and tested in vivo for nerve for 4 min, the PO film uniaxially shrunk, along the direction stimulation. Functional analysis, comprising electrophysiological parallel to the clamped sides (Figures 1E–G). After shrinkage, tests and sensory recovery, was performed. After 90 days, the the PO was fixed to the glass slide by using Kapton tape and implanted RnCEs were tested for their ability to stimulate the clips on each side (to avoid further shrinkage), and the SU-8 regenerated nerve axons and their selectivity assessed in relation 29% layer was cured for cross-linking by UV exposure (30 s) to transparency. followed by heating (70◦ C, 4 min). In order to produce the proper final tube’s dimension (10 mm length, 2.5 mm lumen) the PO film (≈30 µm final thickness, after shrinkage) was cut MATERIALS AND METHODS as shown in Figure 2A. The “tubing area”, 10 mm square, was rolled up by wrapping it around a cylindrical metal mold with Regenerative Nerve Cuff Electrode (RnCE) diameter 2.5 mm. The overlapping “binding areas” were bonded Fabrication by pressing them together with a flat scalpel heated with a hot The Print and Shrink fabrication started with the printing gun. A thin layer of poly(dimethyl siloxane) (PDMS; Sylgard 184) process, which entailed multiple layers deposition (Figure 1) was added on the closure of the tubes to ensure the sealing. The with a lab-scale inkjet printer (DMP-2800; Fujifilm). Heat-shrink PDMS was finely brushed onto the closure of the tubes, supported PO wrap films (Cryovac D-940; Shrink Packaging Sealed Air, by silanized glass pipettes. The silanization process, adopted to 18 µm thickness) were used as substrates. PO films adopted here prevent the tubes from sticking to the support, was performed by had a nominal free shrink of 70–75% upon heating at 120◦ C. exposing the glass pipette to 30 min chlorotrimethylsilane (99%, The conductive materials used for printing were a PEDOT:PSS- Sigma Aldrich) vapor created in an airtight plastic container by based ink (PJET700; Clevios) mixed with 10% (w/w) glycerol putting a small quantity of solution in a vial (Bernardeschi et al., (Farmalabor), and an Ag-precursor ink (Reactive silver ink; 2015). The sealed tubes were then left to dry in open air for 1 day. Sigma-Aldrich). As dielectric material, a solution of SU-8 (SU- The final RnCE is depicted in Figure 2B. A custom PCB 8 2002; MicroChem Corp.), an epoxy resin-based dielectric, was (CadLine) (Figure 2C) was connected to the RnCE pads used. The pristine SU-8 2002 (SU-8 29%) is characterized by area with conductive epoxy (Ag/epoxy adhesive; Chemtronics). 29% solid content in cyclopentanone solution. A less viscous SU- Connections were then insulated with UV-curable glue (Bondic; 8 formulation (SU-8 10%), characterized by 10% solid content, 365 nm, 410 mW cm-2, 60 s). A suture flap (Figure 2B) was left was obtained by adding cyclopentanone (Sigma-Aldrich) to the near the pads area to allow the anchoring of the RnCE at one of SU-8 2002. For each step, a single layer was printed, at room the nerve ends during the surgical procedure for implantation. temperature, using 1 jetting nozzle and a specific drop spacing (ds). The ds, which corresponds to the distance between centers of two contiguous jetted drops, was tuned in order to deposit the RnCE Surface and Electrical thinnest uniform layer possible. Before the deposition, polyimide Characterization (PI) adhesive tape (Kapton R ) and metal clips (Double clips The thickness of each printed layer was evaluated by printing 25 mm; Lebez) were used to fix the PO film (70 mm x 40 mm) to the inks on a silicon wafer and measuring with a stylus a glass support slide (75 × 50 mm). Air plasma (Colibrì plasma profiler (P-6 stylus profiler; KLA Tencor). The inner wall system; Gambetti) was performed to activate the PO surface surface topography was analyzed by means of atomic force Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 3 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes FIGURE 1 | The Print and Shrink fabrication process. (A) Inkjet printing of the conductive ink based on PEDOT:PSS. (B) Inkjet printing of the Ag ink. (C) Inkjet printing of the first SU-8 10% layer. (D) Inkjet printing of the SU-8 29% layer. (E) PO film clamped onto the glass slide. (F) Thermally-induced uniaxial shrinking. (G) Schematics of surface wrinkling induced by heat-shrinking. microscopy (AFM) in order to have an insight on the spatial operating in tapping mode with a silicon probe (NSG01; NT- periodicity and the height of wrinkles. AFM images were MDT), and scanning in the direction perpendicular to the acquired with a Veeco Innova Scanning Probe Microscope wrinkle axis. The images were then analyzed with a free and Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 4 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes FIGURE 2 | Regenerative nerve Cuff Electrodes (RnCEs). (A) Planar view of the final device, with the design of the tube lumen and geometrical references before the rolling up. (B) Picture of the final RnCE. (C) Rendering of the device with an inset of the custom PCB. open source software (Gwyddion1 ) by extracting the sequences female Sprague–Dawley rats (250–300 g) were anesthetized of peaks (x-z profiles) along the direction perpendicular to with ketamine/xylacine (90/10 mg/kg i.p.). Then, the sciatic the wrinkles (x axis). To describe the spatial periodicity, the nerve was exposed at the midthigh following a muscle splitting wrinkle wavelength (λ) was defined as the average distance incision. The sciatic nerve was transected 90 mm proximal to between first-neighboring peaks along the x axis. The wrinkle the third toe, a portion of 5 mm was resected, and the proximal height (h), i.e., the average of the peak heights, was found by and distal stumps were sutured to both ends of a silicone tube averaging the z axis distances between each peak and the first- (10 mm long, 2.5 mm i.d., 0.5 mm thick, n = 5) or a PPP tube neighboring left and right valleys along the profile. Scanning (10 mm long, 2.5 mm i.d., 30 µm thick, n = 6) leaving a 8 mm electron micrographs were acquired by using a EVO MA10 SEM gap between nerve ends (Figure 4). All implanted tubes were (Zeiss), operating at 5 kV accelerating voltage. The electrical filled with saline. After the nerve repair, the wound was closed characterization of the printed PEDOT:PSS traces was performed with silk sutures and anesthesia was reversed with atipamezole with a four point probe tester (Cascade PM5), averaged over hydrochloride (0.2 mg/kg s.c.). four samples. The active RnCEs (n = 6) were implanted following a similar procedure (Figure 4C). First, the sciatic nerve was dissected and In vivo Implantation transected, and the distal nerve was introduced in the tube to First, we tested the main features of the RnCEs, their test the functionality of the electrode for stimulation (section design and the active materials, producing the PPP tubes Electrophysiological Tests). After the tests, a 5 mm portion of (Supplementary Material, PEDOT:PSS coated PO tubes the nerve was resected, and the distal and proximal stumps fabrication and Supplementary Figure 1). In order to assess of the nerve were sutured to the tube of the electrode leaving whether PPP tubes are able to support nerve regeneration, an 8 mm gap. The tubes were filled with saline, the wounds they were implanted as nerve conduit to bridge a gap in the were closed and animals were left to recover. Although six transected sciatic nerve of rats (Valero-Cabré et al., 2001) animals received the RnCE, one RnCE was not functional once and compared with standard silicone tubes. For the implant, implanted so the electrode was left to assess regeneration but the animal was not included in the functional evaluation of 1 https://www.gwyddion.net RnCEs. Then, also in this group, one of the animals died due Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 5 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes to anesthesia-associated complications during nerve conduction 1% BSA in TBS) containing 0.1% Tween. Sections were then test after 45 days of implantation. incubated overnight at 4◦ C in a mixture in BS of NF200 The experimental procedures were approved by the ethical (AB5539, 1:500 Millipore) and S100 (22520, 1:100 DiaSorin) committee of the Universitat Autònoma de Barcelona in primary antibodies for staining axons and Schwann cells, accordance with the European Communities Council Directive respectively. Samples were then washed and incubated again 2010/63/EU. Animals were kept on standard laboratory for 2 h at room temperature for biotin amplification of the conditions with a light-dark cycle of 12:12 h and ad libitum S100 antibody (BA-1000, 1:200 Vector). Samples were washed access to food and tap water. All efforts were made to minimize again and incubated with secondary antibodies Alexa Fluor 594 pain and animal discomfort during surgery and treatments. Goat anti-Chicken (A11042 1:200 Thermofisher) and Alexa Fluor 488 Streptavidin (S11223, 1:200 Thermofisher) diluted in TBS. Evaluation of Axonal Regeneration Finally, sections were washed and cover-slipped with Mowiol Assessment of Motor Reinnervation containing DAPI (1:10,000, Sigma) for nuclear counterstain. Functional reinnervation of target muscles was assessed at Sections were visualized with an epifluorescence microscope 30, 45, 60, and 90 days post-injury (dpi) by motor nerve (Eclipse Ni, Nikon) attached to a digital camera (DS-Ri2, Nikon). conduction tests. Briefly, animals were anesthetized and sub- To evaluate the microstructure of the nerve, distal segments dermal needle electrodes were placed transcutaneously at the were post-fixed with osmium tetroxide (2%, 2 h) and dehydrated sciatic notch for electrical stimulation using single electrical through ethanol series prior to embedding in epon resin. pulses (Synergy Medelec, Viasys HealthCare). The compound Semithin 0.5 µm thick sections were stained with toluidine muscle action potentials (CMAPs) of tibialis anterior (TA) blue. Microscope (Olympus BX51) images were taken and and plantar interosseus (PL) muscles were recorded using thin measurement of cross-section area of the distal segment of the needles placed in the muscle belly in monopolar configuration. sciatic nerve and quantification of the number of myelinated The reference electrode was placed at the fourth toe and a ground nerve fibers was carried out using Image J software (Del Valle electrode was placed at the knee. The amplitude and latency et al., 2018). of the CMAP was measured, and contralateral intact limbs (n = 10) were used as control. The rat body temperature was maintained throughout the test with a thermostatic warming Electrophysiological Tests flat coil. Functional evaluation of RnCEs was performed just after implantation (0 dpi) and after 90 dpi. As indicated above, at 0 dpi, Assessment of Skin Sensory Reinnervation the sciatic nerve was sectioned and introduced in the tube portion The progression of nociceptive reinnervation of the hind paw of the RnCE, and stimulated as with common cuff electrode was assessed by means of pinprick test at 30, 45, 60, and 90 implants. Afterwards, proximal and distal nerve stumps were dpi. Animals were gently kept in a cloth with the sole of the sutured to the ends of the RnCE for allowing nerve regeneration. injured paw facing upward, and the skin was stimulated with a At 90 dpi, electrical stimulation was applied with the RnCE to the needle. Five sites of the lateral side of the paw were tested, being regenerated sciatic nerve to assess the stimulation performance each site stimulated three times, and responses were recorded of the implanted electrodes. At both timepoints, biphasic current as positive only when clear reaction (as fast withdrawal and pulses were delivered through each one of the electrodes against vocalization) was triggered by the stimulation (Navarro et al., a small needle reference electrode placed proximally near the 1994). The percentage of positive areas per animal was calculated, nerve. Increasing current pulses with a width of 10–100 µs and and contralateral paws were tested as controls. an intensity up to 8 mA (max charge 800 nC) were delivered by a Digitimer DS4 stimulator. The CMAP was recorded from Histology of Regenerated Sciatic Nerve gastrocnemius medialis (GM), TA and PL muscles using small At the end of functional evaluation (90 dpi), animals were needle electrodes placed in each muscle (Badia et al., 2011). euthanized by an intraperitoneal injection of pentobarbital, The CMAPs were amplified (P511AC, Grass), band-pass filtered perfused with 4% paraformaldehyde in PBS for 30 min and (3 Hz to 3 kHz) and digitized with a Powerlab recording system the sciatic nerves were collected. The nerve guides were (PowerLab16SP, ADInstruments) at 20 kHz. The amplitude of clipped and sciatic nerves divided in two parts. Proximal each CMAP was measured baseline to peak and normalized to halves were stored in cryoprotected solution of PBS-sucrose the maximum CMAP amplitude obtained in each experiment by 30% with azide 0.1% at 4◦ C before cryosectioning and stimulation of the sciatic nerve with a needle electrode. For each immunofluorecent processing. Distal halves were post-fixed in electrode, the threshold current of stimulation that elicited 5, 30, 3% glutaraldehyde−3% paraformaldehyde in cacodylate-buffer and 95% of the maximum CMAP was determined. The electrode solution (0.1 m, pH 7.4) at 4◦ C for epon embedding. with the lowest threshold value in each RnCE (best AS) was used Proximal segments were embedded in OCT gel (Tissue for data analysis. Finally, the selectivity index (SI) was calculated Tek) before cryosectioning. A cryostat (Leica Microsystems, to quantify the specific activation of a single muscle among Germany) was used to collect 15 µm thick cross sections. For the set of three muscles (GM, PL, TA) when stimulating from immunostaining, sections were first hydrated with TBS, blocked each electrode, as previously described (Veraart et al., 1993) and for endogenous biotins (E21390, ThermoFisher), and incubated the maximum SI (SImax) from each electrode for each muscle for 90 min at room temperature in a blocking solution (BS, was used. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 6 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes Data Analysis PEDOT:PSS and SU-8 10%. Notably, both materials are Data are presented as mean ± standard error of the mean (SEM). well adherent to the PO wrap film, since they interlocked with Results have been statistically analyzed by using GraphPad the substrate during its softening. The process ensures the Prism 8 (GraphPad Software, USA). Statistical significance for development of compliant interfaces (Supplementary Material, comparison between groups has been analyzed using one-way Supplementary Figures 2, 3) without signs of delamination, ANOVA, or mixed effect models followed by Tukey’s post-hoc even after prolonged incubation periods in cell culture medium, tests when required. Statistical significance is considered when as demonstrated in previous studies with a similar heat-shrink P-value was < 0.05. substrate (Bonisoli et al., 2017). The PEDOT:PSS film is characterized by an anisotropic (i.e., uniaxial) wrinkled texture, RESULTS uniform over the whole surface of the electrodes (Figures 3A,B). An average spacing between wrinkles (wrinkle wavelength, λ) RnCEs Fabrication Strategy of 1.1 ± 0.4 µm, and an average wrinkle height (h) of 0.4 ± The Print and Shrink process consists of four steps: (1) the direct 0.2 µm were found (Figure 3C). The SU-8 10% film defined a patterning of materials by means of inkjet printing on PO foil, topography with wrinkled micro-grooves in a periodic pattern (2) the heat-shrink induced self-assembly miniaturization, (3) the with a pitch of around 40 µm (Figure 3D), while the wrinkled rolling-up of the foil into a tube and (4) the PCB packaging. areas (Figures 3E,F) showed characteristics very similar to the The printing process, depicted in Figures 1A–D, defines the PEDOT:PSS wrinkled electrodes. Average SU-8 10% wrinkle microelectrode array (Figure 2A), which consists of 5 electrodes, wavelength was found to be λ = 1.0 ± 0.4 µm and average with pre-shrink size of 200 × 130 µm2 , and the feed lines. The wrinkle height h = 0.3 ± 0.2 µm. The 40 µm pitch grooves are PEDOT:PSS-based ink, adopted as main conductive material, was the result of the self-assembly process in the SU-8 10% layer, formulated with 10% glycerol, enabling an increased conductivity where printed lines did not coalesce in a uniform layer. The and facilitating the printing process (Garma et al., 2019). A wrinkling is here confined in the central region of each line, printed Ag pattern is added on top of the PEDOT:PSS terminal whereas a smooth surface is observed at the edges. The smooth pads (i.e., at the external connection interface) to improve the areas did not undergo wrinkling probably because they were electrical junction with the PCB. SU-8 is used as dielectrics, in two too thick due to the so-called coffee ring effect (Hu and Larson, different formulations. SU-8 10% is used to define the electrodes 2006). Nevertheless, nor delamination or cracks were observed and SU-8 29% to provide the full insulation of the feed lines (SU- in the whole surface. A different texture is observed in the 8 29%). SU-8 29% was not used for the whole coverage since SU-8 29% layer, deposited only on top of the conductive traces. it tends to delaminate and crack upon PO shrinking, due to its Such dielectric has not been dried neither UV-cured before stiffness. The less viscous SU-8 10% ink was therefore adopted to shrinking. As a consequence, upon PO shrinking, it resulted in obtain a thin and flexible coverage. Insights on the SU-8 printing a uniform flat coating insulating the underlying bilayer. The and shrinking are provided in Supplementary Material (SU-8 coating thickness was 2.7 µm, two orders of magnitude larger printing process, Supplementary Figure 2). than the thickness of the PEDOT:PSS patterns (40 nm). The After printing, the heat-shrinking is performed. By heating SU-8 29% cover ensured an adherent interface with PEDOT:PSS above PO glass transition temperature the shrink-wrap film wrinkled electrodes (Supplementary Figure 2F). softened and irreversibly shrunk in-plane, as a result of relaxation The electrical resistance of the PEDOT:PSS traces ranged of pre-stretched polymer chains. As a consequence, the skin from 144 ± 9 to 249 ± 23 k (average ± standard deviation, layers (PEDOT:PSS and SU-8) were induced to be buckled, averaged over four samples), depending on the length of each upon the compressive planar forces exerted by the substrate trace. Detailed electrical resistance data with respect to traces (Figures 1E,F). In the actual setup, the printed PO film uniaxially length are reported in Table 1 (Supplementary Figure 4 for shrunk to 65% of its original lateral size (≈26 mm final width, traces reference). Figures 1E,F). The final geometrical area of the electrodes is therefore 130 × 130 µm2 . Notably, after shrinking the PO wrap film remains flexible, optically transparent, puncture resistant Regeneration After Nerve Section and and it can be wrapped and rolled-up in tubular conduits resulting Repair in a flexible PNI (Figure 2B). A custom PCB is finally connected Silicone tubes, PPP tubes and RnCEs were compared as to the RnCE pads through a miniaturized package (Figure 2C). conduits to support sciatic nerve regeneration along a 8 mm gap in rats. In all cases, we found a regenerated nerve cable RnCEs Surface and Electrical inside the tubes (Figures 4D–F) at 90 dpi. Nerve histological Characterization evaluation showed a centered nerve containing a dense core Surface characterization allowed to analyze the 3D topography of regenerative units with myelinated and unmyelinated axons of the RnCE lumen. We found two different textures, a wrinkled and small blood vessels. The immunolabeling of regenerated one and a smooth one, which are essentially related to the nerves showed in the three groups a wide distribution of composition of printed materials, the printing parameters and axons surrounded by Schwann cells (Figures 5A–C). Histological the curing processes. sections showed high amount of regenerated nerve fibers without The first texture, which is found on the electrodes signs of axonal damage, such as no or very thin myelin and the major dielectric, is related to the wrinkling of sheath, or signs of Wallerian degeneration (Figures 5D,E). Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 7 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes FIGURE 3 | PEDOT:PSS and SU-8 wrinkles. (A,B) SEM images of PEDOT:PSS wrinkled active site. Scale bar: 10 and 2 µm. (C) AFM image of wrinkled PEDOT:PSS (scale bar: 5 µm), and extracted x-z profile (gray line scan). (D,E) SEM images of SU-8 10% wrinkled micro-grooves. Scale bar: 20 and 2 µm. (F) AFM image of wrinkled SU-8 10% (scale bar: 5 µm), and extracted x-z profile (gray line scan). TABLE 1 | Electrical characterization of PEDOT:PSS traces, averaged over 4 had slightly higher CMAP amplitudes of both TA and PL muscles samples. than the other two groups but differences were not significant PEDOT:PSS traces Length (mm) Resistance (k) (p > 0.05) at any time point (Figure 5F). The pinprick test on the plantar skin showed that pain sensation returned to 1 6.4 144 ± 9 the paw from 45 dpi, without differences (p > 0.05) between 2 7.6 164 ± 26 the three groups (Figure 5G). In summary, functional results 3 12.6 249 ± 23 indicated motor and sensory reinnervation in animals implanted 4 12.2 214 ± 22 with RnCEs, comparable to levels obtained with a standard 5 9.6 176 ± 28 silicone tube. See Supplementary Figure 4 for traces reference. In vivo Assessment of RnCE Functionality The RnCEs were used to stimulate the sectioned but still functional sciatic nerve at the implant time (0 dpi), and Quantification of myelinated axons showed no differences (p then following the same stimulation protocol at 90 dpi on > 0.05, one-way ANOVA) between the three tested tubes the regenerated nerve (Figure 6). All the RnCEs were able to (Figure 5H). efficiently stimulate regenerated axons and induce a CMAP We also carried out functional evaluation tests to assess in the three tested muscles at 90 dpi. Recruitment curves of muscle and skin reinnervation following nerve regeneration. muscle activity were plotted for each muscle and compared Motor nerve conduction tests showed complete denervation of between 0 and 90 dpi (Figures 6A,B). The thresholds of the hindlimb muscles after injury. At 45 dpi, CMAPs of very charge needed to reach a 5, 30, and 95% of the maximal small amplitude and long latency were recorded from the TA CMAP amplitude significantly increased up to six times after muscle, while responses from the PL muscle, more distal in the 90 dpi (p < 0.05, Mixed-effects) in comparison with values paw, started to appear at 60 dpi. All animals showed evidence needed at day 0 (Figure 6C). Finally, the capability of the of reinnervation of TA and PL muscles at 90 dpi. The CMAP RnCE to selectively stimulate different groups of axons leading amplitude of both TA and PL muscles was significantly lower to selective recruitment of different muscles was maintained in the injured side than in the contralateral paw (p < 0.05) in over time, and the SImax showed no significant differences all the groups during follow-up, indicating partial reinnervation. between the acute and chronic implant time-points (p > 0.05) Regarding the implanted nerves, the group with a silicone tube (Figure 6D). Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 8 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes FIGURE 4 | The implanted regenerative cuff electrode. Representative images of: (A) silicone tube, (B) PEDOT:PSS coated PO tube (PPP tube), and (C) Regenerative nerve Cuff Electrode (RnCE) at the time of implant (0 dpi) bridging a 8 mm gap in the rat sciatic nerve, and of (D) silicone tube, (E) PPP tube and (F) RnCE at 90 dpi, showing the regenerated nerve cable inside. Yellow arrows point proximal and distal sciatic stumps sutured to each device. White arrows point the PCB. Scale bar: 5 mm. DISCUSSION texturing of electrodes is meant to increase their surface area while minimizing the electrode size. This strategy has been In this study we present a novel regenerative PNI, the RnCE, indeed already reported to enhance charge injection capacity fabricated through a non-conventional low-cost approach, the in PNIs (Cogan, 2008). Regarding the biocompatibility of the Print and Shrink strategy. RnCEs connect the microfabrication active material, in vitro studies already demonstrated the non- of CPs, the new frontier in PNIs material, with the development cytotoxicity of PEDOT:PSS coatings in neuronal cells (Charkhkar of non-obtrusive regenerative interfaces, the challenge of the next et al., 2014) and its in vivo stability in acute to early-chronic generation PNIs. To the best of our knowledge RnCE is the first studies (Kozai et al., 2015). Although an exhaustive long-term in fully transparent all-polymer regenerative PNIs. vivo analysis has not been performed yet, recent studies showed The Print and Shrink entails low-cost substrate material and stability of PEDOT:PSS in in vitro long-term experiments [one or cost-effective fabrication techniques. We have already reported 3 weeks (Cellot et al., 2016) up to 4 months (Dijk et al., 2020)]. inkjet printing as an affordable technique for fabricating all- Hence, the facts that our results show a similar amount of axons polymer PEDOT:PSS based multi-electrode arrays on flexible between standard silicone tubes and the PEDOT:PSS tubes and substrates, with application in in vitro electrophysiology (Garma also that the degree of functional recovery is similar, suggest that et al., 2019). Since inkjet printing does not require any kind this material should be suitable for in vivo neural applications. of masking, the process enables to save materials and quickly Remarkably, the here reported low-cost fabrication strategy customize the design of the electrodes. Moreover, the optimized enables a facile tailorability of the electrodes design. The ease printing method here reported allowed the deposition of thin of customization, in terms of electrodes sizing and overall layers, with all the materials adopted, which is crucial to ensure device dimensions, opens for multiple applications with diverse a conformal wrinkling. Indeed, neither cracks nor delamination animal models up to personalized peripheral interfaces in of all the printed layers were observed after shrinking and case of human lesions. Moreover, the developed technique the inkjet printed films were stable upon bending of the allows reliable adhesion of the patterned materials, enabling flexible substrate and remained stable after in vivo tests. The to overcome mechanical instability, as cracks or delamination wrinkling manufacturing process here reported is an expansion reported in multiple implants and limiting the neuroprostheses of a previous work on smart multifunctional biointerfaces for outcome (Barrese et al., 2016; Ganji et al., 2018). muscle and neuronal cells (Greco et al., 2013; Bonisoli et al., Given their flexibility, RnCEs are suitable for soft 2017). In those cases the self-assembled anisotropic topography interfacing with biological tissues. Nevertheless, the polymeric and the conductive properties of PEDOT:PSS were found to microelectrode array arranged on the PNI lumen and the whole promote and orient cell growth and differentiation. Here, the design do not impair the natural process of nerve regrowth. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 9 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes FIGURE 5 | Histological and electrophysiological evaluation of nerve regeneration. (A–E) Representative images of cross sectioned nerves repaired with silicone, PEDOT:PSS coated PO tube (PPP) and Regenerative nerve Cuff Electrode (RnCE) at 90 dpi. (A) NF200 (red) for axons; (B) S100 (green) for Schwann cells; (C) Merge of A and B with nuclear staining (DAPI, blue); (D,E) Brightfield images of regenerated nerves stained with toluidine blue labeling myelin sheaths at high (D) and low (E) magnification. Scale bar: 50 µm in (A–D), 200 µm in (E). (F) Amplitude of the CMAP of tibialis anterior (TA) and plantar interossei (PL) muscles along the 90 days follow-up after sciatic nerve section and repair. (G) Skin paw reinnervation tests by pinprick test. (H) Quantification of myelinated fibers in the sciatic nerve distal to the tube in the three tested groups. *p < 0.05 vs. other groups. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 10 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes FIGURE 6 | Muscle recruitment curves. (A,B) Representative muscle recruitment curves from selected active sites at (A) 0 dpi and (B) 90 dpi. The increase of the electrical charge delivered through the device elicited recruitment of more muscular fibers thus increasing the CMAP amplitude. (C) Values of the electrical charge needed to elicit 5, 30, and 95% of the maximum CMAP amplitude of gastrocnemius (GM), tibialis anterior (TA) and plantar interossei muscles (PL). (D) Maximal selectivity index (SImax) for GM, TA, and PL muscles over time. *p < 0.05 vs. time. In this sense, the RnCEs allowed axons to grow through a gap such as the use of porous materials (Meyer et al., 2016), or fillers of 8 mm between nerve stumps (Figure 4), a longer distance with neurotrophic factors (Del Valle et al., 2018) or cells (Allodi than used in previous studies with other regenerative interfaces, et al., 2014), nor better also than the current gold standard as such as sieve (Lago et al., 2005), microchannel (Srinivasan the autograft (Dietzmeyer et al., 2020). Finally, tube or cuff et al., 2015) or double-aisle (Delgado-Martínez et al., 2017) electrodes, such as the RnCE, should be flexible, self-sizing, electrodes. The 8 mm gap chosen is a relatively long distance, have a larger diameter than the nerve to be implanted and with but still below the critical length that can be repaired with the thinnest wall without compromising their stability (Rodri standard polymer conduits (Lundborg et al., 1982; Valero- et al., 2000). While potential damage to the regenerated nerve Cabré et al., 2001; Yannas and Hill, 2004). We chose the 8 mm induced by the chronically implanted RnCE is a possibility, the gap to prove that both the PPP tube and the RnCE do not small thickness of the device and the enhanced flexibility of hinder regeneration and this material behaves similarly as the materials give the tube of the RnCE a soft interface with other plastics such as Teflon (Navarro and Kennedy, 1991) or the nerve. silicone (Deumens et al., 2010). Moreover, motor and sensory The RnCE exploits the PNS regeneration capabilities to reinnervation outcomes were similar between the silicone, the directly interface the nervous system. In this way the RnCEs PPP tube and the RnCE; and the number of myelinated axons overcome the issues related to nerve compression and impaired did not differ between the three groups. Therefore, our results nerve regeneration, known to affect the interface viability indicate that the RnCE device (which includes also the PCB reported for sieve and microchannel devices (Navarro et al., that can induce traction forces which could affect regeneration) 1996; Wallman et al., 2001; FitzGerald, 2016). RnCEs are able and the PPP tube alone are as good as conventional silicone to stimulate regenerated axons, although a higher electrical tubes in terms of supporting peripheral nerve regeneration. charge is needed in comparison to axons that had been recently However, the three groups did not reach complete functional cut. This is attributable to the physiopathological changes of recovery after 90 days of implantations, as expected. As with regenerated axons, which are considerable thinner in caliber and other regenerative electrodes, the RnCE cannot be considered as with thinner myelin sheaths than normal, for long time after a better regenerative strategy than other tubulization techniques injury (Gómez et al., 1996), and thus have a higher stimulation Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 11 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes threshold (Blair and Erlanger, 1933; McNeal, 1976). Moreover, By maintaining the maximal transparency of the PNI and resting membrane hyperpolarization is increased in regenerated increasing the microelectrodes array resolution, next generation motor axons (Moldovan and Krarup, 2004), contributing to the RnCEs will combine the benefit of good support for axonal observed lower excitability. regeneration without obstacle or compression with improved We have characterized the RnCEs in terms of stimulation selective stimulation and recording capabilities. Finally, the selectivity, that in PNIs is usually inversely related to the cost-effective Print and Shrink strategy would be useful for invasiveness of the device implantation (del Valle and Navarro, the manufacturing of neural interfaces wherein the optical 2013). Compared with previous studies in the literature, the transparency of the device plays a pivotal role, as in the case RnCEs showed a selectivity similar than non-regenerative cuff of optogenetics. electrodes (Badia et al., 2011), but lower compared to the double- aisle regenerative electrodes which separated the regenerated DATA AVAILABILITY STATEMENT nerve fascicles (Delgado-Martínez et al., 2017), or the TIME intrafascicular electrodes applied to the intact rat sciatic nerve The datasets presented in this study can be provided by the (Badia et al., 2011). Hence, RnCEs may not be the best approach authors under reasonable request. for implants in a large nerve trunk, such as the sciatic nerve, but it may be more efficient for implants in smaller nerves, that ETHICS STATEMENT innervate a limited target. The combination of the technology developed for the RnCE with other regenerative interface designs, The animal study was reviewed and approved by Comissió such as the double-aisle regenerative (Delgado-Martínez et al., d’Ètica en l’Experimentació Animal i Humana de la Universitat 2017) or the REMI (Garde et al., 2009) electrodes, in which the Autònoma de Barcelona. tube lumen is divided in separate sections by a septum where electrodes can be also located, may be a favorable option to reach higher interface selectivity. In addition, innovative multiple AUTHOR CONTRIBUTIONS aisle regenerative devices, may allow guidance by molecular cues LMF and AB developed the fabrication strategy (Print and for sorting the regeneration of specific axonal populations (Del Shrink) under the supervision of FG. AC helped in the Valle et al., 2018) directing motor or different types of sensory design of the electrodes array, developed the PCB, and the fibers to specific areas of the regenerative interface. RnCEs may electrical interconnections. SM, FG, and XN conceived the also be suitable to be used in cases where nerve transections study. BR-M and JV performed the in vivo experiments are surgically repaired with nerve conduits. Interestingly, nerve under the guide of XN. LMF, AC, AB, BR-M, JV, FG, electrical stimulation with RnCEs might be further exploited to and XN contributed to the analyses and discussion of the promote axonal regeneration and treat neuropathic pain within results. All authors contributed to writing and correcting the rehabilitation program after nerve repair (Asensio-Pinilla the manuscript. et al., 2009; Cobianchi et al., 2013). In conclusion, we report here a new fabrication method that can produce a fully transparent regenerative nerve electrode, FUNDING the RnCE. This electrode can be adjusted in size to house This work was partially supported by FLAG-ERA JTC 2017 different nerve calibers with a fast and cheap manufacturing project GRAFIN, grant PCI2018-093029 from Ministerio process. Future studies will determine whether it has the potential de Ciencia, Innovación y Universidades of Spain, co- advantages of regenerative interfaces, such as the possibility funded by European Union (ERDF/ESF, Investing in to record the signals of early regenerating fibers, to generate your future), TERCEL (RD16/0011/0014) and CIBERNED electrical field to improve regeneration (Gordon, 2016) or to (CB06/05/1105) funds from Instituto de Salud Carlos III of interface axons that have regenerated to bidirectionally connect Spain. FG acknowledges partial financial support from Top the subject with an external device (del Valle and Navarro, Global University Project at Waseda University, Tokyo from 2013). Future and more prolonged studies will be aimed at MEXT Japan. determining whether the RnCE could cause discomfort in the regenerated nerves after more than 3 months, or the materials cause any specific damage in the surrounding tissue. Moreover, ACKNOWLEDGMENTS it will be investigated the capability of the Print and Shrink strategy to produce microelectrode arrays at higher resolution The authors thank Monica Espejo and Jessica Jaramillo for their with increased electrodes density. A significant reduction in technical help. electrode size can be obtained by fully exploiting the heat shrinking of the PO substrate. Nominally a bi-dimensional free SUPPLEMENTARY MATERIAL shrinkage of 70–75% at 120◦ C can be reached, much higher than the 35% uniaxial shrinkage here obtained. Such improvement The Supplementary Material for this article can be found will require the setting of a dedicated fabrication methodology online at: https://www.frontiersin.org/articles/10.3389/fbioe. to ensure accuracy and repeatability in pattern miniaturization. 2021.615218/full#supplementary-material Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 12 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes REFERENCES of human embryonic stem cells and their cardiac derivatives. Adv. Mater. 23, 5785–5791. doi: 10.1002/adma.201103463 Allodi, I., Mecollari, V., González-Pérez, F., Eggers, R., Hoyng, S., Verhaagen, J., Cho, S.-H., Lu, H. M., Cauller, L., Romero-Ortega, M. I., J.-,Lee, B., and Hughes, G. et al. (2014). Schwann cells transduced with a lentiviral vector encoding Fgf- A. (2008). Biocompatible SU-8-based microprobes for recording neural spike 2 promote motor neuron regeneration following sciatic nerve injury. Glia 62, signals from regenerated peripheral nerve fibers. IEEE Sens. J. 8, 1830–1836. 1736–1746. doi: 10.1002/glia.22712 doi: 10.1109/JSEN.2008.2006261 Altuna, A., Bellistri, E., Cid, E., Aivar, P., Gal, B., Berganzo, J., et al. (2013). SU-8 Clements, I. P., Mukhatyar, V. J., Srinivasan, A., Bentley, J. T., Andreasen, based microprobes for simultaneous neural depth recording and drug delivery D. S., and Bellamkonda, R. V. (2012). Regenerative scaffold electrodes for in the brain. Lab Chip. 13, 1422–1430. doi: 10.1039/c3lc41364k peripheral nerve interfacing. IEEE Trans. Neural Syst. Rehabil. Eng. 21, Asensio-Pinilla, E., Udina, E., Jaramillo, J., and Navarro, X. (2009). 554–566. doi: 10.1109/TNSRE.2012.2217352 Electrical stimulation combined with exercise increase axonal Cobianchi, S., Casals-Diaz, L., Jaramillo, J., and Navarro, X. (2013). Differential regeneration after peripheral nerve injury. Exp. Neurol. 219, 258–265. effects of activity dependent treatments on axonal regeneration and doi: 10.1016/j.expneurol.2009.05.034 neuropathic pain after peripheral nerve injury. Exp. Neurol. 240, 157–167. Asplund, M., Nyberg, T., and Inganäs, O. (2010). Electroactive polymers for neural doi: 10.1016/j.expneurol.2012.11.023 interfaces. Polym. Chem. 1, 1374–1391. doi: 10.1039/c0py00077a Cogan, S. F. (2008). Neural stimulation and recording electrodes. Annu. Rev. Badia, J., Boretius, T., Andreu, D., Azevedo-Coste, C., Stieglitz, T., and Biomed. Eng. 10, 275–309. doi: 10.1146/annurev.bioeng.10.061807.160518 Navarro, X. (2011). Comparative analysis of transverse intrafascicular del Valle, J., and Navarro, X. (2013). Interfaces with the peripheral nerve multichannel, longitudinal intrafascicular and multipolar cuff electrodes for the control of neuroprostheses. Int. Rev. Neurobiol. 109: 63–83. for the selective stimulation of nerve fascicles. J. Neural Eng. 8:036023. doi: 10.1016/B978-0-12-420045-6.00002-X doi: 10.1088/1741-2560/8/3/036023 Del Valle, J., Santos, D., Delgado-Martínez, I., de la Oliva, N., Giudetti, Barrese, J. C., Aceros, J., and Donoghue, J. P. (2016). Scanning electron G., Micera, S., et al. (2018). Segregation of motor and sensory axons microscopy of chronically implanted intracortical microelectrode arrays in regenerating through bicompartmental tubes by combining extracellular non-human primates. J. Neural Eng. 13:026003. doi: 10.1088/1741-2560/13/2/ matrix components with neurotrophic factors. J. Tissue Eng. Regen. Med. 12, 026003 e1991–e2000. doi: 10.1002/term.2629 Bellamkonda, R. V., Pai, S. B., and Renaud, P. (2012). Materials for neural Delgado-Martínez, I., Righi, M., Santos, D., Cutrone, A., Bossi, S., D’Amico, interfaces. MRS Bull. 37, 557–561. doi: 10.1557/mrs.2012.122 S., et al. (2017). Fascicular nerve stimulation and recording using a Berggren, M., Nilsson, D., and Robinson, N. D. (2007). Organic materials for novel double-aisle regenerative electrode. J. Neural Eng. 14:046003. printed electronics. Nat. Mater. 6:3. doi: 10.1038/nmat1817 doi: 10.1088/1741-2552/aa6bac Berggren, M., and Richter-Dahlfors, A. (2007). Organic bioelectronics. Adv. Mater. Deumens, R., Bozkurt, A., Meek, M. F., Marcus, M. A. E., Joosten, E. A. J., Weis, 19, 3201–3213. doi: 10.1002/adma.200700419 J., et al. (2010). Repairing injured peripheral nerves: bridging the gap. Prog. Bernardeschi, I., Greco, F., Ciofani, G., Marino, A., Mattoli, V., Mazzolai, B., et al. Neurobiol. 92, 245–276. doi: 10.1016/j.pneurobio.2010.10.002 (2015). A soft, stretchable and conductive biointerface for cell mechanobiology. Dietzmeyer, N., Förthmann, M., Grothe, C., and Haastert-Talini, K. (2020). Biomed. Microdevices. 17:46. doi: 10.1007/s10544-015-9950-0 Modification of tubular chitosan-based peripheral nerve implants: applications Bernards, D. A., Owens, R. M., and Malliaras, G. G. (2008). Organic for simple or more complex approaches. Neural Regen. Res. 15, 1421. Semiconductors in Sensor Applications. Berlin Heidelberg: Springer. doi: 10.4103/1673-5374.271668 doi: 10.1007/978-3-540-76314-7 Dijk, G., Rutz, A. L., and Malliaras, G. G. (2020). Stability of PEDOT: PSS-coated Bettinger, C. J. (2018). Recent advances in materials and flexible gold electrodes in cell culture conditions. Adv. Mater. Technol. 5:1900662. electronics for peripheral nerve interfaces. Bioelectron. Med. 4:6. doi: 10.1002/admt.201900662 doi: 10.1186/s42234-018-0007-6 Elschner, A., Kirchmeyer, S., Lovenich, W., Merker, U., and Reuter, K. (2010). Blair, E. A., and Erlanger, J. (1933). Comparison of properties of individual axons in PEDOT: Principles and Applications of an Intrinsically Conductive Polymer. the frog. Proc. Soc. Exp. Biol. Med. 30, 728–729. doi: 10.3181/00379727-30-6647 Boca Raton, FL: CRC Press. doi: 10.1201/b10318 Bonisoli, A., Marino, A., Ciofani, G., and Greco, F. (2017). Topographical FitzGerald, J. J. (2016). Suppression of scarring in peripheral nerve implants by and electrical stimulation of neuronal cells through microwrinkled drug elution. J. Neural Eng. 13:026006. doi: 10.1088/1741-2560/13/2/026006 conducting polymer biointerfaces. Macromol. Biosci. 17:1700128. Gabardo, C. M., Adams-McGavin, R. C., Fung, B. C., Mahoney, E. J., Fang, Q., doi: 10.1002/mabi.201700128 and Soleymani, L. (2017). Rapid prototyping of all-solution-processed multi- Boretius, T., Badia, J., Pascual-Font, A., Schuettler, M., Navarro, X., Yoshida, lengthscale electrodes using polymer-induced thin film wrinkling. Sci. Rep. 7, K., et al. (2010). A transverse intrafascicular multichannel electrode (TIME) 1–9. doi: 10.1038/srep42543 to interface with the peripheral nerve. Biosens. Bioelectron. 26, 62–69. Ganji, M., Hossain, L., Tanaka, A., Thunemann, M., Halgren, E., Gilja, V., et al. doi: 10.1016/j.bios.2010.05.010 (2018). Monolithic and scalable Au nanorod substrates improve PEDOT–metal Caironi, M., and Noh, Y. Y. (2015). Large Area and Flexible Electronics. Weinheim: adhesion and stability in neural electrodes. Adv. Healthc. Mater. 7:1800923. Wiley-VCH Verlag GmbH & Co. doi: 10.1002/9783527679973 doi: 10.1002/adhm.201800923 Cellot, G., Lagonegro, P., Tarabella, G., Scaini, D., Fabbri, F., Iannotta, S., et al. Garde, K., Keefer, E., Botterman, B., Galvan, P., and Romero-Ortega, M. I. (2016). PEDOT: PSS interfaces support the development of neuronal synaptic (2009). Early interfaced neural activity from chronic amputated nerves. Front. networks with reduced neuroglia response in vitro. Front. Neurosci. 9:521. Neuroeng. 2:5. doi: 10.3389/neuro.16.005.2009 doi: 10.3389/fnins.2015.00521 Garma, L. D., Ferrari, L. M., Scognamiglio, P., Greco, F., and Santoro, F. Chan, Y., Skreta, M., McPhee, H., Saha, S., Deus, R., and Soleymani, L. (2018). (2019). Inkjet-printed PEDOT: PSS multi-electrode arrays for low-cost in vitro Solution-processed wrinkled electrodes enable the development of stretchable electrophysiology. Lab Chip. 19, 3776–3786. doi: 10.1039/C9LC00636B electrochemical biosensors. Analyst 144, 172–179. doi: 10.1039/C8AN01637B Genzer, J., and Groenewold, J. (2006). Soft matter with hard skin: From skin Charkhkar, H., Frewin, C., Nezafati, M., Knaack, G. L., Peixoto, N., wrinkles to templating and material characterization. Soft Matter. 2, 310–323. Saddow, S. E., et al. (2014). Use of cortical neuronal networks for in doi: 10.1039/b516741h vitro material biocompatibility testing. Biosens. Bioelectron. 53, 316–323. Gómez, N., Cuadras, J., Butí, M., and Navarro, X. (1996). Histologic assessment doi: 10.1016/j.bios.2013.10.002 of sciatic nerve regeneration following resection and graft or tube repair in the Chen, A., Lee, E., Tu, R., Santiago, K., Grosberg, A., Fowlkes, C., et al. mouse. Restor. Neurol. Neurosci. 10, 187–196. (2014). Integrated platform for functional monitoring of biomimetic heart Gordon, T. (2016). Electrical stimulation to enhance axon regeneration after sheets derived from human pluripotent stem cells. Biomaterials 35, 675–683. peripheral nerve injuries in animal models and humans. Neurotherapeutics 13, doi: 10.1016/j.biomaterials.2013.10.007 295–310. doi: 10.1007/s13311-015-0415-1 Chen, A., Lieu, D. K., Freschauf, L., Lew, V., Sharma, H., Wang, J., et al. Greco, F., Fujie, T., Ricotti, L., Taccola, S., Mazzolai, B., and Mattoli, V. (2013). (2011). Shrink-film configurable multiscale wrinkles for functional alignment Microwrinkled conducting polymer interface for anisotropic multicellular Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 13 February 2021 | Volume 9 | Article 615218 Ferrari et al. All-Polymer Printed Regenerative Cuff Electrodes alignment. ACS Appl. Mater. Interfaces 5, 573–584. doi: 10.1021/am3 Nemani, K. V., Moodie, K. L., Brennick, J. B., Su, A., and Gimi, B. (2013). In 01908w vitro and in vivo evaluation of SU-8 biocompatibility. Mater. Sci. Eng. C 33, Greco, F., Ventrelli, L., Dario, P., Mazzolai, B., and Mattoli, V. (2012). Micro- 4453–4459. doi: 10.1016/j.msec.2013.07.001 wrinkled palladium surface for hydrogen sensing and switched detection Nguyen, D., Sa, S., Pegan, J. D., Rich, B., Xiang, G., McCloskey, K. E., et al. (2009). of lower flammability limit. Int. J. Hydrogen Energy 37, 17529–17539. Tunable shrink-induced honeycomb microwell arrays for uniform embryoid doi: 10.1016/j.ijhydene.2012.04.145 bodies. Lab Chip. 9, 3338–3344. doi: 10.1039/b914091c Green, R., and Abidian, M. R. (2015). Conducting polymers for neural Rivnay, J., Owens, R. M., and Malliaras, G. G. (2013). The rise of organic prosthetic and neural interface applications. Adv. Mater. 27, 7620–7637. bioelectronics. Chem. Mater. 26, 679–685. doi: 10.1021/cm4022003 doi: 10.1002/adma.201501810 Rodri, F. J., Ceballos, D., Schu, M., Valero, A., Valderrama, E., Stieglitz, T., et al. Green, R. A., Matteucci, P. B., Hassarati, R. T., Giraud, B., Dodds, C. (2000). Polyimide cuff electrodes for peripheral nerve stimulation. J. Neurosci. W. D., Chen, S., et al. (2013). Performance of conducting polymer Methods 98, 105–118. doi: 10.1016/S0165-0270(00)00192-8 electrodes for stimulating neuroprosthetics. J. Neural Eng. 10:016009. Rodríguez-Hernández, J. (2015). Wrinkled interfaces: taking advantage of doi: 10.1088/1741-2560/10/1/016009 surface instabilities to pattern polymer surfaces. Prog. Polym. Sci. 42, 1–41. Groenendaal, L., Jonas, F., Freitag, D., Pielartzik, H., and Reynolds, doi: 10.1016/j.progpolymsci.2014.07.008 J. R. (2000). Poly (3, 4-ethylenedioxythiophene) and its Rosberg, H.-E., Carlsson, K. S., and Dahlin, L. B. (2005). Prospective study derivatives: past, present, and future. Adv. Mater. 12, 481–494. of patients with injuries to the hand and forearm: costs, function, and doi: 10.1002/(SICI)1521-4095(200004)12:7<481::AID-ADMA481>3.0.CO;2-C general health. Scand. J. Plast. Reconstr. Surg. Hand Surg. 39, 360–369. Hu, H., and Larson, R. G. (2006). Marangoni effect reverses coffee-ring depositions. doi: 10.1080/02844310500340046 J. Phys. Chem. B 110, 7090–7094. doi: 10.1021/jp0609232 Sharma, H., Nguyen, D., Chen, A., Lew, V., and Khine, M. (2011). Unconventional Khodagholy, D., Doublet, T., Quilichini, P., Gurfinkel, M., Leleux, P., Ghestem, A., low-cost fabrication and patterning techniques for point of care diagnostics. et al. (2013). In vivo recordings of brain activity using organic transistors. Nat. Ann. Biomed. Eng. 39, 1313–1327. doi: 10.1007/s10439-010-0213-1 Commun. 4, 1–7. doi: 10.1038/ncomms2573 Shum, A. M. Y., Che, H., Wong, A. O. T., Zhang, C., Wu, H., Chan, C. W. Y., Kozai, T. D. Y., Catt, K., Du, Z., Na, K., Srivannavit, O., Razi-ul, M. H., et al. (2015). et al. (2017). A micropatterned human pluripotent stem cell-based ventricular Chronic in vivo evaluation of PEDOT/CNT for stable neural recordings. IEEE cardiac anisotropic sheet for visualizing drug-induced arrhythmogenicity. Adv. Trans. Biomed. Eng. 63, 111–119. doi: 10.1109/TBME.2015.2445713 Mater. 29:1602448. doi: 10.1002/adma.201602448 Lacour, S. P., Courtine, G., and Guck, J. (2016). Materials and technologies Simon, D. T., Gabrielsson, E. O., Tybrandt, K., and Berggren, M. (2016). Organic for soft implantable neuroprostheses. Nat. Rev. Mater. 1, 1–14. bioelectronics: bridging the signaling gap between biology and technology. doi: 10.1038/natrevmats.2016.63 Chem. Rev. 116, 13009–13041. doi: 10.1021/acs.chemrev.6b00146 Lago, N., Ceballos, D. F. J., Rodriguez, Stieglitz, T., and Navarro, X. (2005). Srinivasan, A., Tahilramani, M., Bentley, J. T., Gore, R. K., Millard, D. C., Long term assessment of axonal regeneration through polyimide regenerative Mukhatyar, V. J., et al. (2015). Microchannel-based regenerative scaffold for electrodes to interface the peripheral nerve. Biomaterials 26, 2021–2031. chronic peripheral nerve interfacing in amputees. Biomaterials 41, 151–165. doi: 10.1016/j.biomaterials.2004.06.025 doi: 10.1016/j.biomaterials.2014.11.035 Lago, N., Yoshida, K., Koch, K. P., and Navarro, X. (2007). Assessment Stieglitz, T., and Meyer, J. U. (1999). Implantable microsystems. Polyimide-based of biocompatibility of chronically implanted polyimide and platinum neuroprostheses for interfacing nerves. Med. Device Technol.10:28. intrafascicular electrodes. IEEE Trans. Biomed. Eng. 54, 281–290. Stríteský, S., Marková, A., Víteček, J., Šafaríková, E., Hrabal, M., Kubáč, L., et al. doi: 10.1109/TBME.2006.886617 (2018). Printing inks of electroactive polymer PEDOT: PSS: The study of Lew, V., Nguyen, D., and Khine, M. (2011). Shrink-induced single-cell plastic biocompatibility, stability, and electrical properties. J. Biomed. Mater. Res. A microwell array. JALA 16, 450–456. doi: 10.1016/j.jala.2011.06.003 106, 1121–1128. doi: 10.1002/jbm.a.36314 Lundborg, G., Dahlin, L. B., Danielsen, N., Gelberman, R. H., Longo, F. M., Valero-Cabré, A., Tsironis, K., Skouras, E., Perego, G., Navarro, X., and Neiss, Powell, H. C., et al. (1982). Nerve regeneration in silicone chambers: influence W. F. (2001). Superior muscle reinnervation after autologous nerve graft of gap length and of distal stump components. Exp. Neurol. 76, 361–375. or poly-L-lactide-ǫ-caprolactone (PLC) tube implantation in comparison doi: 10.1016/0014-4886(82)90215-1 to silicone tube repair. J. Neurosci. Res. 63, 214–223. doi: 10.1002/1097- Malhotra, B. D., and Ali, M. A. (2017). Nanomaterials for Biosensors: Fundamentals 4547(20010115)63:2<214::AID-JNR1014>3.0.CO;2-D and Applications. Amsterdam: William Andrew. Veraart, C., Grill, W. M., and Mortimer, J. T. (1993). Selective control of muscle Márton, G., Tóth, E. Z., Wittner, L., Fiáth, R., Pinke, D., Orbán, G., et al. activation with a multipolar nerve cuff electrode. IEEE Trans. Biomed. Eng. 40, (2020). The neural tissue around SU-8 implants: a quantitative in vivo 640–653. doi: 10.1109/10.237694 biocompatibility study. Mater. Sci. Eng. C Mater. Biol. Appl. 112:110870. Wallman, L., Zhang, Y., Laurell, T., and Danielsen, N. (2001). The geometric doi: 10.1016/j.msec.2020.110870 design of micromachined silicon sieve electrodes influences functional McNeal, D. R. (1976). Analysis of a model for excitation of myelinated nerve regeneration. Biomaterials 22, 1187–1193. doi: 10.1016/S0142-9612(00) nerve. IEEE Trans. Biomed. Eng. BME-23, 329–337. doi: 10.1109/TBME.1976. 00342-2 324593 Wong, W. S., and Salleo, A. (2009). Flexible Electronics: Materials and Meyer, C., Stenberg, L., Gonzalez-Perez, F., Wrobel, S., Ronchi, G., Udina, Applications. Boston, MA: Springer Science and Business Media. E., et al. (2016). Chitosan-film enhanced chitosan nerve guides for doi: 10.1007/978-0-387-74363-9 long-distance regeneration of peripheral nerves. Biomaterials 76, 33–51. Yannas, I. V., and Hill, B. J. (2004). Selection of biomaterials for peripheral doi: 10.1016/j.biomaterials.2015.10.040 nerve regeneration using data from the nerve chamber model. Biomaterials 25, Moldovan, M., and Krarup, C. (2004). Mechanisms of hyperpolarization 1593–1600. doi: 10.1016/S0142-9612(03)00505-2 in regenerated mature motor axons in cat. J. Physiol. 560, 807–819. doi: 10.1113/jphysiol.2004.069443 Conflict of Interest: The authors declare that the research was conducted in the Navarro, X., Calvet, S., Butí, M., Gómez, N., Cabruja, E., Garrido, P., et al. (1996). absence of any commercial or financial relationships that could be construed as a Peripheral nerve regeneration through microelectrode arrays based on silicon potential conflict of interest. technology. Restor. Neurol. Neurosci. 9, 151–160. doi: 10.3233/RNN-1996-9303 Navarro, X., and Kennedy, W. R. (1991). The effects of autologous Copyright © 2021 Ferrari, Rodríguez-Meana, Bonisoli, Cutrone, Micera, Navarro, nerve transplants on motor and sudomotor reinnervation by Greco and del Valle. This is an open-access article distributed under the terms regenerative axons. Brain Res. 565, 181–187. doi: 10.1016/0006-8993(91) of the Creative Commons Attribution License (CC BY). The use, distribution or 91648-K reproduction in other forums is permitted, provided the original author(s) and the Navarro, X., Verdú, E., and Buti, M. (1994). Comparison of regenerative and copyright owner(s) are credited and that the original publication in this journal reinnervating capabilities of different functional types of nerve fibers. Exp. is cited, in accordance with accepted academic practice. No use, distribution or Neurol. 129, 217–224. doi: 10.1006/exnr.1994.1163 reproduction is permitted which does not comply with these terms. Frontiers in Bioengineering and Biotechnology | www.frontiersin.org 14 February 2021 | Volume 9 | Article 615218