Materials Today
ResearchNontransient silk sandwich for soft, conformal bionic links
Graphical abstract
Introduction
Electrical interfaces to the neural tissues provide a compelling platform for the marriage between electrical engineering, material science, and biology. Indeed, bioelectronics is a rapidly expanding field [1] with profound impact on neurophysiological applications. Many diagnostic and therapeutic procedures in bioelectronic medicine are used to treat neurological disorders such as that of the peripheral nervous system and the brain [2]. Many such examples and applications require stable in vivo monitoring and/or modulation of neural signals via implantable electrode technologies. Such needs have led to relentless efforts to improve the electrode interfaces in terms of geometry, materials, and biological footprint, enabling smaller and flexible designs [3], [4]. Existing bionic links for the nerves and the brain fall into two broad categories: penetrating probes that leave a biological footprint in the target tissues and thin-film surface arrays that are non-penetrating and are deployed on the tissue surfaces [4]. Flexible electrode interfaces that can seamlessly integrate into the biological milieu to obtain high-fidelity electrical readouts without eliciting physical damage to the target tissues has emerged as a critical necessity to form reliable bioelectronic interfaces.
Over the last decade, there has been an ongoing push to make the electrode interfaces flexible, exemplified by the onset of plastic polymeric films in the development of surface electrode interfaces, for both the nerves [5], [6] and the brain [7], [8], [9], [10]. Although they are biocompatible, the dry and static nature of the plastic substrates are quite foreign to the biological tissue, which is generally soft, wet, and dynamic [3], [11] (Supplementary Table 1). There is also discomfort associated with the deployment of plastic interfaces, due to possible cracks that can either develop at the suture holes or across the interface body. In vivo deployment of these technologies are fundamentally limited by their brittle nature, even as the critical requirement for soft compliant interfaces comes to the forefront. Design strategies for such soft interfaces must address challenges associated with the material mismatch between the soft biological tissues and the substrate and superstrate layers of the bionic links. Although many flexible plastic realizations exist [8], [10], [12], [13], [14], [15], [16], [17], there have been no definitive solutions for incorporating soft conformal substrates that can couple intimately to the curvilinear tissue surfaces, and at the same time eliciting little or no deleterious effects on the target tissues during deployment, an interesting challenge in itself.
The above need translates towards a material that provides a soft flexible electrode interface, is mechanically robust and biocompatible, shows great affinity to the wet surfaces of the biological tissues, and remains robust and stable for long periods of time. Materials that were originally not developed for the soft tissue interfaces have become attractive candidates for softer yet robust electrode substrates. However, the challenge is to choose materials with desirable properties (such as robustness, softness and biocompatibility), develop innovative strategies for their incorporation into the neural interface design, and realize soft biocompatible interfaces that are easily deployable in vivo, and remain functional for a long period of time. Here, we choose silk fibroin, a well-known naturally produced material. The properties of silk such as excellent biocompatibility, robust mechanical properties, and effortless control of its degradation lifetime ranging from a few seconds to years are well documented [18], [19]. Apart from these traditional reasons, there exists an invaluable but unexplored set of properties as listed below, that we exploit for neural electrode technology.
- •
Hydrated water-processed silk is soft and highly flexible [20], [21], showing excellent affinity to wet surfaces. In vivo conditions, for example, nerve and cortical surfaces, are wet, which serve as the target sites of application for hydrated silk.
- •
Moreover, the programmable lifetime of silk [7], coupled with its adhesive property while being moist [22], provide promising avenues for integrating the thin form-factor silk-based devices with the tissue surfaces, as exemplified later by the conformal wrapping of silk bionic links around thin nerves.
The excellent biocompatibility of silk fibroin has facilitated the development of resorbable devices for in vivo applications. [7], [13], [22], [23], [24], [25], [26] However, incorporating silk fibroin in the development of soft, flexible and non-dissolvable silk electrode arrays that can seamlessly and intimately couple to the wet surfaces of the neural tissues has not been reported. Existing silk-based devices can be broadly categorized into pristine silk micro- and nanostructures [27], [28], [29], [30], [31], [32], [33], [34], silk-metal constructs [22], [35], [36], [37] and resorbable silk layered supports [7], [23], [25]. Current device-level realizations for in vivo deployment are almost exclusively bioresorbable architectures [7], [23], [25], relying on the tendency of silk fibroin to resorb into the biology with programmable dissolution kinetics. Such “built to disappear” construction dramatically limits the scope of use for silk fibroin in the design of stable neural interfaces. In this light, the nontransient nature of the silk layers that serve as substrates in the fabrication process flow incorporating pristine silk becomes critically important. Even though there have been demonstrations of sustainability of bare silk film implants in an animal model [37], functional non-dissolvable silk interfaces to the tissues have not been realized yet.
Here, we introduce materials and device design strategies to realize silk electrode arrays as bionic links for direct electrical interfacing to neural tissues such as the peripheral nerve and the cortex. The platform builds on new technologies that exploit silk fibroin as the organic foundation of a new class of implantable silk-based devices that are nontransient, flexible and conformal to small tissue geometries. The key enabling concepts include SILK-SEAL and QUICK-SILK as defined below.
- •
SILK-SEAL involves integration of two thin (15 μm thick) dry silk layers serving as the substrate and the insulation, housing a conducting layer in between, to yield a silk ‘sandwich’ sensor. We refer to this concept as the SILK-SEAL as it enables green assembly of two silk layers using moist silk as an adhesive.
- •
QUICK-SILK employs a bilayer bandage of resorbable silk and curable medical-grade elastomer to facilitate quick conformal deployment and securing of the resulting silk devices on the tissue surfaces in a suture-free manner. This technology serves as a band-aid to ‘paste’ the flexible silk arrays onto the target tissue surfaces.
The results presented here indicate that silk, in its hydrated state, can serve as a soft, yet stable material to design sustainable implantable sensors for seamless integration into the biological milieu, hinting at a new breed of soft and excellently biocompatible electrical sensors. In addition to their established nontransient nature, water-annealed silk layers, when deployed with ultrathin interconnect metal layer and flexible components for electrical connection, yield devices with levels of mechanical flexibility necessary for conformal contact with the neural tissues. Comprehensive in vitro tests (3 months) establish the reliability of silk fibroin layers for use in neural interface design such as to record, stimulate or serve as scaffold for regenerating nerve. In addition, the silk arrays reported here provide accurate and reproducible measurements of signals similar to neural activity for acute use (h). Concept validations and evaluation of design strategies include the acute capture of electroneurograms (ENG) and micro-electrocorticograms (ECoG) in anesthetized animal models, where silk fibroin serves, both as the nontransient laminate on top of the nerve and cortical surfaces supporting recording electrode sites and as a resorbable backing facilitating intimate electrical contacts with the target tissues.
Section snippets
Nontransient silk electrode arrays for neural interfacing
Figure 1, Figure 1b show the schematic diagrams of thin flexible construction of the silk electrode arrays based on a selectively exposed gold (Au) conducting layer embedded between two nontransient silk layers. A thin layer of Au (thickness 300 nm) is deposited on dry silk substrates (15 μm thick) to yield conductive electrode patterns on silk. Another layer of silk (15 μm thick) electrically isolates the connection traces from surrounding biofluids and adjacent tissues. The sensing sites
Conclusion
In this work, we presented non-resorbable, soft, ultra-flexible silk bionic links in the form of electrode arrays for interfacing with the neural tissues (peripheral nerve and the cortex). The silk bionic links were enabled by novel silk-based concepts: SILK-SEAL and QUICK-SILK. The SILK-SEAL technique facilitated the fabrication of the soft water-stable implantable multi-layered silk sandwich sensors, a first of its kind using silk fibroin (Supplementary Tables 10 and 11), an FDA approved
Aqueous silk fibroin
Aqueous solution of silk fibroin is prepared using previously reported methods [18]. Silk fibers are boiled in 0.02 M sodium carbonate solution for 30 min to obtain degummed silk fibroin. After drying it overnight, it is dissolved in 9.3 M solution of lithium bromide and then dialyzed to obtain 4–6% (wt/vol) silk solution.
Evaluating the mass of water-annealed silk under accelerated soak
Aqueous silk (2 mL of 5% wt/vol) is poured into small weighing boats (square base with sides 30 mm) and left to dry overnight at ambient conditions, to obtain silk film
Author contributions
A.C.P. and N.V.T. conceived the idea. A.C.P., A.B., Y-H.L., B.L. and N.V.T designed the research. A.C.P. fabricated the devices. A.C.P. and B.L. conceived and performed bench tests and analysis. A.C.P. and B.L. performed electrical, mechanical and bioactivity characterization for the devices. A.C.P., A.B., Y-H.L. and N.V.T. performed in vivo experiments and analyzed the data. A.C.P., A.B., B.L. and N.V.T. wrote and edited the manuscript.
Disclosure statement
The authors have a corresponding patent application (10201801877U), which is currently pending.
Data availability
The raw data required to reproduce these findings forms part of an ongoing study. Supplementary material accompanies this paper and provides processed data referenced by the main article.
Acknowledgments
The work is supported by National Research Foundation (NRF-CRP10-2012-01). We thank Marshal Dian Sheng Wong and Li Jing Ong, Singapore Institute for Neurotechnology for their excellent technical assistance with setting up the recording apparatus; Gil G. L. Gammad, Singapore Institute for Neurotechnology for his excellent assistance with animal handling and surgery; Dr. Dihan Hasan (Group of Chengkuo Lee, Department of Electrical & Computer Engineering, NUS) for his inputs and excellent support
References (54)
Acta Biomater.
(2010)Curr. Opin. Biomed. Eng.
(2018)Colloids Surf., B
(2018)Biomaterials
(2007)Acta Biomater.
(2012)Exp. Neurol.
(1981)Neurobiology of Disease
(2015)- et al.
Nature
(2016) Nat. Rev. Drug Discovery
(2014)Adv. Mater.
(2014)