Development of zwitterionic hydrogel thin film coatings of cochlear implant electrode arrays that reduce intracochlear inflammation and electrode impedance

Abstract

Cochlear implants (CIs) have revolutionized the treatment of sensorineural hearing loss, yet patient outcomes remain highly variable due to biological responses within the cochlea. A critical challenge is the foreign body response (FBR) triggered by CI biomaterials, which can lead to inflammation, fibrosis, and increased electrode impedance, ultimately impairing auditory function. Zwitterionic hydrogels have been shown to provide a highly lubricious anti-fouling surface that minimizes protein and cell adsorption. This study evaluates a carboxybetaine methacrylate (CBMA) zwitterionic hydrogel coating to reduce the FBR to cochlear implants. Using a large animal (sheep) model, human CI electrode arrays were coated with a thin film CBMA hydrogel via one step UV photografting and photopolymerization with a custom mold. Electrophysiological measurements at the time of implantation demonstrate significantly reduced total impedance, polarization impedance, and access resistance in CBMA-coated CIs. Reductions in total impedance and access resistance were maintained after 4 weeks in vivo. High-resolution Xray Microscopy imaging confirmed intracochlear placement without translocation or tip fold-over and a trend to reduced neo-ossification in CBMA-coated implants. Histological analysis revealed significantly decreased cellular infiltration, macrophage infiltration, and fibrotic tissue deposition within the cochlea surrounding CBMA-coated implants compared to uncoated controls. This work highlights the potential of durable thin film zwitterionic coatings to enhance CI performance and preserve residual hearing by mitigating insertional trauma and attenuating chronic inflammation and fibrosis.

Graphical Abstract

1. Introduction

Over the past 30 years, cochlear implants (CIs) have evolved dramatically, transforming from experimental devices into a widely accepted clinical solution for individuals with severe-to-profound sensorineural hearing loss, with over 600,000 implants performed over the past 30 years.¹ Technological advancements in electrode array design, signal processing, and surgical techniques have substantially enhanced speech perception and auditory outcomes. Despite these improvements, variability in patient performance persists, often linked to biological factors within the cochlea. One major limitation is intracochlear inflammation, which can impair neural survival, increase fibrosis, and reduce the electrode-neuron interface efficacy, ultimately compromising long-term implant function and auditory outcomes.²

CI electrode arrays consist of platinum contacts and input wires embedded in a poly(dimethyl siloxane) (PDMS) carrier. The PDMS enables flexibility and long-term mechanical stability, helps resist wire breaks, and seals circuitry from body fluids. However, insertion of a CI into the cochlea can cause immediate structural damage as well as a vigorous inflammatory, fibrotic response to the foreign body in humans and animals.^2–10 Host reactions following implantation of biomaterials include injury, blood-material interactions, provisional matrix formation, acute inflammation, chronic inflammation, granulation tissue development, and fibrous capsule formation. This foreign body response (FBR) begins immediately upon implantation, where trauma related to insertional forces is correlated with an increase in proinflammatory cytokines and the presence of free radicals.¹¹ Nonspecific proteins such as fibrinogen and albumin adsorb to the surface of the biomaterial. Neutrophils release proinflammatory cytokines which further recruit macrophages and lymphocytes to the area in an attempt to digest the biomaterial. Finally, as the response develops, fibroblasts are recruited and encase the biomaterial in a fibrotic capsule.^12,13(Fig. 1) This response is largely dependent on the surface properties and fouling potential of the CI biomaterials.¹³

Figure 1.

Schematic: (A) Foreign body response on cochlear implants. (B) Zwitterionic monomers (CBMA) and non-zwitterionic cross-linker (PEGDMA) simultaneously photopolymerized and photografted to the cochlear implant surface yielding (C) ultra-low biofouling coating and reduced fibrotic encapsulation. (D) Coated cochlear implants are inserted in sheep for 4 weeks.

In the case of CIs the fibrous sheath that encases the electrode array increases impedances and negatively impacts signal resolution.^14–16 The cellular response to biomaterials can corrode metals and degrade polymeric materials, thereby reducing the functional longevity of the implant,^17–19 and in rare cases, a robust FBR can lead to extrusion of the electrode array outside the cochlea.^19–24 The inflammatory FBR may also be toxic to the neurosensory structures of the inner ear^25,26 and has been implicated in reduction of residual acoustic hearing in hearing preservation cochlear implantation.^2,27,28 Importantly, histological and electrophysiological data suggest that delayed loss of acoustic hearing following CI is likely driven by the inflammatory FBR.^7,10,29,30 This FBR coupled with trauma during implantation, results in structural and functional damage to delicate cochlear tissues leading to significantly decreased performance and loss of residual hearing.^2,7–10,31

Many strategies have been investigated for mitigating the FBR on implanted biomaterials including coating biomaterials with hydrogel coatings which mask the biomaterial surface from the host tissues.³² Zwitterionic hydrogels have drawn recent attention due to their unique structural moieties that have a positive and negative charge (zwitterion) within the same molecule. These hydrogels offer excellent biomimetic and biocompatible properties due to their high water content, porous structure, and compressive modulus which resembles native tissues and organs.³³ For example, carboxybetaine methacrylate (CBMA) is a zwitterionic molecule with a large dipole moment capable of attracting multiple water molecules, and when used as the molecular backbone for a hydrogel creates a coating that is extremely hydrophilic. This CBMA-coating imbues significant lubricity to the surface of the CI, and provides a water layer that inhibits fouling of proteins, cells, and bacteria to the material surface.^34–38

Our group has engineered novel photografted hydrogel coatings based on ultra-low fouling zwitterionic systems comprised of CBMA that minimize protein, cellular, and bacterial adhesion (Fig. 1).^39–42 These non-biodegradeable zwitterionic hydrogel thin films are covalently grafted to the biomaterial surface and resist biofouling for both in vitro and in vivo systems.^39–42 The thin film zwitterionic hydrogel coating has also demonstrated long-term in vivo durability and significant lubricity.⁴³ Herein we describe a novel method for applying CBMA hydrogel coatings on human CI electrode arrays and assess electrophysiological and histological responses in a large animal model.

2. Materials and Methods:

Benzophenone, acetone, paraformaldehyde, and hydroxy-4’-(2-hydroxyethoxy)-2-methylpropuophenone (HEPK) were obtained from Sigma-Aldrich (St. Louis, MO). 3-[[2-(Methacryloyloxy)ethyl]dimethylammonio]propionate (CBMA) was purchased from TCI Chemicals (Portland, OR). Poly(ethylene glycol) dimethacrylate (PEGDMA) with an average spacer molecular weight of 400 (Polysciences, Warrington, PA) was used as the cross-linking molecule for the zwitterionic thin films. Dimethylsiloxane-acetoxy terminated ethylene oxide block copolymer (75% non-siloxane) (Gelest, Morrisville, PA) was used as a surfactant dispersing agent. Anti-αSMA primary antibody, 488 goat antirabbit secondary antibody, 633 goat antimouse secondary antibody, and glass coverslips were obtained from Thermo Fisher (Waltham, MA). Anti-Iba-1 primary antibody was obtained from Fujifilm Wako Corporation, (Osaka, Japan) for macrophage labeling. An Omnicure S1500 lamp (Lumen Dynamics, Mississauga, Canada) was used for photocuring and photografting. Imaging was performed with both light and confocal microscopes (Leica, Wetzlar, Germany).

2.1. Coating process

All zwitterionic thin film pre-polymer solutions were composed of 31.5wt% zwitterionic monomer CBMA, 3.5wt% cross-linking molecule PEGDMA, 0.8wt% surfactant, 0.05wt% HEPK photoinitiator, and 64.15wt% distilled water. This prepolymer solution was used to coat cochlear implant devices for in vivo experiments. Functional human cochlear implant electrode arrays (HiFocus SlimJ) were provided by Advanced Bionics (San Diego, CA). The surface grafting agent was coated by placing the electrode array in a 50 g/L benzophenone (surface grafting agent) acetone solution, soaked for 1 hour, and dried for 20 minutes in vacuum. To maintain consistent coating thickness across the tapered geometry of the CI electrode array, a custom mold was fabricated using micro-additive printing with the assistance of ProtoStudios (Iowa City, IA) (Fig. 2). 40μL pre-polymer solution was added to the central void before inserting the electrode array. Samples were simultaneously photografted and photopolymerized via broad spectrum UV irradiation for 10 min at 30 mW/cm² under an inert nitrogen atmosphere to prevent oxygen inhibition. The sample was then thoroughly washed using sterile phosphate buffered saline (PBS) to remove any residual solvent or unreacted monomer and initiator from the nascent hydrogel coating. Samples were stored in sterile PBS until implantation. To determine coating thickness, coated CI electrode arrays were soaked in 1g/L fluorescein dye solution for 5 minutes, washed in PBS, cut in cross section, and imaged using light and confocal microscopy.

Figure 2.

Custom mold for uniform coating. (A) The mold is closed via two rubber o-rings at each end and the channel is filled with 40μL zwitterionic monomer solution. Channel is stained with fluorescein dye solution for visualization. (B,C) Cochlear implants dipped in fluorescein dye solution for 5 minutes and imaged with light microscopy, CBMA hydrogel coating absorbs fluorescein dye (C) whereas uncoated electrode array does not (B). Both PDMS and electrode faces are coated with CBMA coating. (D) Cross section of CBMA-coated electrode array with fluorescein dye-embedded coating distinguishing the outer surface of the electrode array as seen on confocal microscopy. (E) Green fluorescent protein and brightfield channels overlayed demonstrating the coating in situ. Hydrated coating exhibits uniform 25μm thickness across the tapered architecture of the electrode array.

2.2. Sheep surgery

All animal surgeries were performed in accordance with approved IACUC protocol (#1092431), including proper anesthesia, analgesia, sterile technique, and post-surgical monitoring. Sheep surgery followed similar workflow to protocols described previously.⁴⁴ Briefly, three female Suffolk-Dorsett sheep were used; male sheep were excluded as they can be combative risking cochlear implant device placement, and complicate housing postoperatively. Sheep were anesthetized and placed prone, clipped free of wool posterior and superior to both ears, and prepped and draped in sterile fashion. The temporal bone was exposed to identify surface landmarks including the temporal line, the mastoid tip, and the mastoid portion. Under an operating microscope, a standard mastoidectomy with facial recess approach was performed and the round window was exposed. After incision of the round window membrane, electrode arrays either coated or uncoated (randomly assigned to right or left ears), were inserted into the round window by the same otologist-neurotologist (MH). The round window was patched with a muscle graft to prevent leakage of perilymph. Surgical wounds were closed in a routine surgical manner with sutures and staples. Animals were monitored postoperatively for vestibular disturbances, facial nerve dysfunction, and wound complications. Animals were anesthetized 4 weeks postoperatively and electrophysiological measurements were repeated.

2.3. Electrophysiology

Impedance measurements were made intraoperatively immediately after electrode array placement, and 4 weeks postoperatively just prior to harvest. Sheep did not experience electrical stimulation outside of testing sessions. Measurements were made using BEDCS 1.18.338 software (Bionic Ear Data Collection System, Advanced Bionics, LLC) connected to a clinical programming interface (CPI) and Platinum Series sound processor. Following conditioning, complex impedance was measured. Voltage waveforms were collected for each electrode. Stimulus and recording parameters used were similar to those previously described by Saoji et al.⁴⁵ Stimuli were 64 μA biphasic pulses, 161 μs per phase. Access impedance was derived from the voltage recorded at 18 μs divided by the stimulus current. Total impedance was derived from the voltage recorded at 161 μs. Polarization impedance was calculated as the difference between total and access impedance. Impedance values from intracochlear electrodes is reported in kiloohms (kΩ). When possible, electrically evoked compound action potentials (eCAP) recordings were used to estimate the location of each electrode as intracochlear or extracochlear. Positive eCAP measurements are assumed to occur when stimulating electrodes are intracochlear. eCAPs were recorded through Advanced Bionics Neural Response Imaging (NRI) tools with standard recording parameters. Due to recording artifacts, eCAP results were of poor technical quality and threshold data cannot be reliably reported. Intracochlear placement of at least 8 electrodes was confirmed either by intraoperative eCAP measurements or as visualized on post mortem 3D Xray microscopy (XRM) images.

2.4. Tissue preparation and histology

After electrophysiology measures were recorded, animals were euthanized and cochleae were collected with implants in place and fixed in 4% paraformaldehyde. Samples were dehydrated to 70% ethanol for 3D XRM imaging. Cochleae were imaged both with the implant in place when possible and after the implant was removed. 3D Slicer software was used for segmentation of neo-ossification based on consistent thresholding of Hounsfield units across specimens. Samples were then rehydrated to 1x PBS before decalcification in 0.2M ethylendiaminetetraacetic acid (EDTA) for 6 weeks. After 6 weeks, samples were sectioned into 30μm tissue sections before undergoing immunostaining and hematoxylin and eosin (H&E) staining. Based on the variability of insertion depth across animals, cochlear sections corresponding to the apical-most electrode were not assessed, rather sections corresponding to the cochlear base approximating the round window opening and the ascending basal cochlear turn in a mid modiolar section were selected for analysis. Primary antibodies used were anti-Iba1 for macrophage labeling as well as anti-αSMA for fibrosis diluted to 1:400 in blocking buffer. Hoechst 3342 was used for nuclear staining.

For histological analysis of nuclei, macrophage, neurons, and fibrosis, 30-μm-thick sections from cochleae were immunolabeled, to visualize macrophage (anti-Iba1 antibody with Alexa 488 conjugated secondary antibody), cell nuclei (Hoechst 3342) and fibrotic tissue (anti-α-SMA antibody with Alexa 633 conjugated secondary antibody). Fluorescently labeled cochlear sections were imaged on a Leica Stellaris 5 confocal system using a 20 × (0.70 NA) objective, 0.75x digital zoom, z-axis-spacing of 1μm, and constant exposure/gain settings throughout the experiment. Sections were labeled for quantification of different cell types and tissues with the data from 3 adjacent sections. Image analysis was performed in IMARIS (Oxford Instruments, UK) image analysis software; cell counts, and quantitation of volumetric analyses were found from maximum intensity z-projections of 3D confocal image stacks.

First, the outline of the scala tympani was traced on cochlear sections from each sheep in the cochlear base and in the basal turn of mid-modiolar sections to establish the volume. The number of macrophages (Iba1), and nuclei (Hoechst 3342) were counted using an automated counting system in IMARIS image analysis software. The density of nuclei and macrophages was calculated per 10⁵μm³. The area of fibrosis was assessed by volumetric quantification of α-SMA in the basal cochlea and in mid-modiolar sections referenced to the volume of the scala tympani. Statistical analysis was performed with GraphPad Prism software. 2-way ANOVA with Sidak correction was used for comparisons.

3. Results

3.1. Coatings

Coating composition was selected carefully based on previous publications on zwitterionic hydrogel coatings.^41,43,46 The zwitterionic monomer, crosslinker, photoinitiator, and surfactant concentrations all enabled the single step photografting and photopolymerization process. Zwitterionic hydrogels have been shown to be tunable in their stiffness based on crosslink density, allowing for the interface of the biomaterial and the surrounding tissue to have comparable moduli.⁴¹ The relatively low crosslink density chosen for this study provided a hydrogel coating that was optimized for a combination of lubricity, antifouling, and antifibrotic properties while still maintaining durability, flexibility, and modulus.

The custom mold was designed with a tapered geometry mimicking CI electrode array but fabricated to be slightly larger to allow pre-polymer solution to surround the array at the appropriate thickness. The mold also allowed transmission of UV light, enabling facile photografting and photopolymerization of a uniform zwitterionic thin film to the electrode array surface. The cross-linked coatings were successfully photopolymerized and photografted to the electrode array surface via one-step curing process with 10 minutes under full spectrum UV irradiation. The coating was approximately 25μm in thickness from the proximal marker to the distal tip in its hydrated form via microscopy. Coverage of both PDMS and platinum substrates was achieved (Fig. 2). No delamination of hydrogel coatings was observed upon removal of the coated electrode arrays, indicating that they withstood sterilization, operative handling, and implantation in the cochlea for four weeks.

3.2. Impedances

Having developed a process that enabled uniform coating of human electrode arrays, we performed a pilot experiment in sheep to validate electrode functionality and assess histological changes. Sheep were implanted bilaterally with one cochlea receiving a standard implant with an uncoated electrode array and the opposite cochlea implanted with a coated electrode array. Complex impedances were measured at the time of implantation and after 4 weeks incubation in vivo. All implanted electrodes, both uncoated and coated, remained functional throughout the study period. Polarization impedance, access resistance, and total impedance was significantly lower with CBMA-coated cochlear implants in saline and immediately after implantation. Access resistance and total impedance remained significantly lower in CBMA-coated cochlear implants after 4 weeks (Fig. 3).

Figure 3.

(A) Schematic of complex impedance. Biphasic electrical pulses are applied across an electrode and the resulting voltage can be used to generate polarization impedance and access resistance components. Total impedance (B), polarization impedance, (C) and access resistance (D) across electrodes at the time of implantation (immediate) and after 4 weeks.

Because the zwitterionic hydrogel coating covers both the electrode contacts and the surrounding PDMS carrier, it is important to verify that the observed impedance reduction is not an artifact of conductive bridging between adjacent electrodes. To address this, the electrode interaction matrix was measured in 1x PBS for arrays coated with sulfobetaine methacrylate (SBMA), a zwitterionic polymer that is structurally analogous to the CBMA (Supplemental Fig. 1). In these measurements, the off-diagonal elements of the interaction matrix (which reflect the degree of electrical coupling between non-stimulated electrodes) were comparably low for both coated and uncoated arrays, and the spatial decay pattern from the stimulating electrode was preserved. This indicates that the zwitterionic hydrogel coating does not create a conductive path between electrode contacts and that the impedance reduction reflects a change in the electrode-electrolyte interface rather than inter-electrode shorting.^47–49

3.3. Intracochlear inflammation:

Both uncoated and CBMA-coated electrode arrays were inserted until resistance was met, but not forced within the cochlea. While the sheep cochlea has many similarities to the human cochlea and has been accepted as an appropriate model for cochlear implantation, the scala tympani volume is less than in the human cochlea,⁵⁰ and the architecture tapers more rapidly towards the apex, limiting cochlear implantion to the basal and mid regions of the cochlea. When possible, 3D XRM demonstrated that the electrode arrays were implanted within the scala tympani to the depth of at least 8 electrodes, accommodating at least 180 degrees of the first turn of the cochlea. There was no evidence of translocation of electrode arrays into the scala media, nor was there tip fold over. The volume of neo-ossification was over an order of magnitude greater in uncoated CIs (mean 12.24×10⁻³ mm³) as compared to CBMA-coated (0.795 ×10⁻³ mm³), but this difference did not reach statistical significance (Supplemental Fig. 2). The neoossification was found in the basal cochlea, with the majority of neoossification most concentrated at first turn of the cochlea where the electrode array first accommodates to the cochlear curvature (Fig. 4).

Figure 4.

Fig. 4 Cochlear fibrosis as imaged via X-Ray Microscopy in uncoated (left) and CBMA-coated (right) CI’s. (A) Axial slice of cochleae demonstrating presence of neossification in the basal cochlea (arrow) as compared to (B) CBMA-coated CI. (C,D) 3D reconstructions of cochleae made semi-transparent with implants in situ (purple). This CBMA-coated CI demonstrates an area of redundant electrode array within the basal cochlea (star). (E,F) 3D representation of Neo-ossification (red). Despite redundancy of the implant, there is minimal fibrosis seen. Scale bar is 2mm.

For each sheep, intracochlear tissue inflammatory responses were measured via immunostaining for cell nuclei (Hoechst 3342), macrophages (IBA-1), and fibrotic tissue (a-SMA). Hematoxylin and eosin staining was also performed. Intra-animal (uncoated vs CBMA-coated) and inter-animal (across all animals) comparisons are displayed according to anatomic location, with the basal turn of the cochlea in mid modiolar sections (Fig. 5) and a longitudinal section of the cochlear base (Fig. 6). Significant variability of inflammatory response was observed after implantation across animals, with sheep #1 mounting the most robust inflammatory response. When combining the data across all three animals, a global effect can be appreciated, especially in the basal cochlea at the level of the round window (Supplemental Fig. 3). Here a statistically significant decrease was observed between CBMA-coated and uncoated implants in cell nuclei density, macrophage density, as well as proportion of scala tympani volume occupied by fibrotic tissue (Fig. 6).

Figure 5.

Basal cochlear turn of mid-modiolar sections comparing uncoated (left) and CBMA-Coated (right) CIs. (A,B) H&E staining with inflammatory infiltrate occupying the scala tympani (star) as compared to scala tympani in CBMA-coated specimen. (C,D) Immunostaining for fibrosis with α-SMA (red), macrophages with Iba1 (green) and cell nuclei with Hoechst 3342 (blue). (E, F, G) 2-way ANOVA comparisons across ears within the same animal as well as summary data across all animals for uncoated (red) and CBMA-coated (blue) implants. Representative images from Sheep 1 where significance was reached in all categories. Scale bars represent 500μm.

Figure 6.

Representative images of basal cochleae at the level of the round window opening comparing uncoated (left) and CBMA-coated (right) CIs. (A) H&E staining with inflammatory infiltrate occupying the scala tympani (star) as compared to scala tympani in CBMA-coated specimen (B). (C,D) Immunostaining for fibrosis with α-SMA (red), macrophages with Iba1 (green) and cell nuclei with Hoechst 3342 (blue) for uncoated (C) and CBMA-coated (D) CIs. (E,F,G) 2-way ANOVA comparisons across ears within the same animal as well as summary data across all animals for uncoated (red) and CBMA-coated (blue) implants. Scale bars represent 500μm.

4. Discussion

Intracochlear inflammation following cochlear implantation affects residual acoustic hearing, increases impedances, and limits the overall effectiveness of the device.² Two major factors that trigger the inflammatory response include insertional trauma and the foreign body response to the implanted biomaterials. Here a thin film zwitterionic hydrogel coating was permanently applied to functional human cochlear implant electrode arrays; CI impedance measures and the inflammatory response in a large animal model were then assessed.

4.1. CBMA coating process

The dramatic antifouling behavior of zwitterionic hydrogel coatings has been extensively studied on flat surfaces. Many neural stimulators, however, have complex geometry with a wide variety of shapes. Coating three dimensional structures poses a significant challenge in translating these materials from the laboratory to clinical use. Cochlear implants present a particular challenge due to their size, conical, tapering architecture, and limited space for insertion, thus requiring thin film coatings.

Simultaneous photopolymerization and photografting to the surface of CI materials enables significant spatial and temporal control of hydrogel synthesis. Other commonly used methods for coating CIs such as dip, spray, and deposition have been used as well.^51,52 The use of a mold designed for a specific implant allows for precise thickness control and coating coverage defined by the mold parameters. Dip coating methods often have challenges with coating thickness for a thin cylindrical shape with beading of monomer solution resulting in lack of coating uniformity.⁵³ This becomes especially relevant with precurved electrode arrays which can aggregate monomer solution unevenly across the maximal and minimal radii of curvature. Spray and deposition coating methods have shown variable coverage and porosity,^54,55 which can lead to increased cellular adhesion. Other methods such as galvanostatic polymerization allow grafting from electrode surfaces resulting in a polymer that does not adequately coat the remainder of the PDMS housing.^56,57 This method could be combined with other coating methods to result in complete coverage of the medical device, however this would likely require multiple steps and increased complexity of process. Due to the ease of scanning and additive manufacturing, the design of molds for various CI arrays is low cost, easily reproducible, scalable, and can be applied to a variety of electrode designs including both straight and pre-curved arrays. Mold design allows for specification of coating thickness, coverage, and uniformity, as well as the spatial control allotted by photopolymerization.^40,42

The method described here provides a relatively simple, scalable, and inexpensive method to reliably provide a durable thin film coating that covers the entire electrode array. A three dimensional mold that allows UV light transmission was developed that provided consistent, uniform, thin film coatings. The custom additive manufacturing mold is inexpensive and can offer a wide variety of customizability, including tailoring coating thickness and architecture. The mold is produced via additive manufacturing, and can be easily reformatted to accommodate the dimensions of each unique electrode array design. The coatings are simultaneously photografted creating a covalent attachment to the PDMS surface and photopolymerized to create the bulk cross-linked hydrogel, allowing for a one-step curing and grafting process with 10 minutes under UV irradiation.

Furthermore, the zwitterionic coating reduces nonspecific protein adsorption and cell adhesion, acting to shield the host from the CI biomaterials including PDMS and platinum to mitigate the foreign body response.⁴⁰ When compared to PDMS, macrophages and fibroblasts adhere preferentially to platinum surfaces.²⁷ Additionally, histological studies have demonstrated significantly thicker fibrotic capsules in areas with exposed platinum as opposed to PDMS.^27,40 Importantly, the coating on these CI electrode arrays is of uniform 25μm thickness from the proximal marker to the distal tip, covering both PDMS and platinum substrates (Fig. 2).

With the use of light for curing, these coatings can be spatially controlled by masking regions from the UV irradiation, directing coating coverage to only extend to the depth marker on the electrode array. Further experimentation with photomasks for creating patterns and other micro-architectures are currently under investigation and have yielded differential cellular growth cues.⁵⁸ The CBMA coating is not biodegradeable and remains in place even 1 year after subcutaneous implantation in mice.⁴⁰ Thus, it is expected to provide a prolonged effect to minimize the foreign body response after cochlear implantation.

4.2. CBMA-coatings of electrode arrays are durable and reduce electrode impedances

Having established a method to apply the thin film coatings, we used functional human CIs in a preliminary, pilot study with a large animal model to further validate the durability and efficacy of the zwitterionic coatings. Importantly, all implanted electrode arrays, both coated and uncoated, maintained their functionality as determined by electrode impedances over the study period. This continued functionality indicates that the coating process and the coating itself does not impair the function of the electrode array, at least over the 4-week study period. We have previously shown that these coatings remain durable when applied to precurved human electrode arrays and implanted subcutaneously in mice for one year, consistent with their expected resistance to biodegradation.⁴⁰

Several animal studies have correlated the presence of inflammatory cells, fibrous tissue, and new bone formation with increased electrical resistivity and suggest that the insertion trauma and the foreign body response contribute to increased electrical impedances after cochlear implantation.^59,60 The scala tympani of the sheep tapers at a greater rate than in humans, which does not allow for a complete insertion of a human CI electrode array.^44,50 It is reasonable to expect that sheep cochlea exhibit similar variation in cochlear anatomy as human cochleae, but this has not been studied directly. In this study electrode arrays were inserted until resistance was met by the surgeon. At the time of insertion, cochlear electrophysiology was measured, including eCAP and complex impedance and the 8 distal electrodes were used for analysis.

As a biphasic impulse is generated by the electrode, the resulting voltage measured can provide information regarding the electrode properties in space.⁶¹ Complex impedances divide the total impedance measure into two components: polarization impedance and access resistance. Polarization impedance is considered to reflect the electrode-electrolyte interface, whereas access resistance is considered to reflect the tissue environment around the electrode (Fig 3A).¹⁶ In our study, polarization impedances were significantly lower in coated samples at the time of insertion; this effect became less pronounced after 4 weeks. It is known that providing electrical stimulus can “condition” the electrode until a quasi stable electrical impedance level is reached.⁶² This change is likely due to adsorption and deposition of chemical species and nonspecific proteins on the electrode surface as a result of reduction-oxidation reactions in response to electrical current. Zwitterions have been shown to provide a conductive advantage known as the “zwitterion effect” thought to be due to the inherent dipole allowing for freeing of electrical charges within solution.^63–65 Recent studies have used zwitterionic coatings on electrodes to reduce protein and cell adhesion and improve electrode sensing for in vivo models.⁶⁶ Furthermore, these coatings have been shown to remain stable and retain their antifouling properties after prolonged electrical stimulation.^67,68 The polarization impedances of uncoated electrodes decreased after 4 weeks, while the polarization impedances of the CBMA-coated electrodes remained largely unchanged. CBMA-coated CI’s are hydrophilic and provide consistent contact of conductive electrolyte on the surface of the electrode, potentially allowing for more efficient charge transfer. It is possible that the zwitterionic hydrogel coating provides a more controlled, electrically favorable interface with the surrounding media which acts to stabilize the electrode surface without the need for extensive conditioning. Further study is underway characterizing these potentially transformative electrochemical properties of zwitterionic hydrogel coatings.

In human clinical practice, access resistance has been used as a biomarker for intracochlear fibrosis.^31,69–71 Access resistances were reduced at the time of implantation with CBMA-coated electrodes, suggesting that the coating alters the environment surrounding the electrode as well. This effect was sustained over the 4-week period, which likely reflects decreased inflammation surrounding the electrode array at the 4-week timepoint. The possible decrease in the inflammatory FBR is further supported by the quantified reduction of neo-ossification, inflammatory cells, and fibrotic tissue as seen by XRM and histology. Cells, fibrosis, blood, and bone formation all differentially conduct electricity and displace the more readily conductive perilymph, insulating the electrode from its neural targets.^72,73 CBMA-coatings imbue the electrode surface with an anti-fouling barrier that repels nonspecific protein adsorption and cellular adhesion, and reduces the inflammatory response, likely contributing to the significant reduction in access resistance in coated electrode arrays measured both at implantation and persistent after 4 weeks of intracochlear implantation.

4.3. CBMA-coating reduces intracochlear inflammation

XRM allows high resolution imaging of the cochlea and identification of neo-ossification. These areas of bone are segmented based on thresholding of Hounsfield units and allow for volumetric quantification of areas of neo-ossification. Post mortem studies in human cochlear implant recipients have demonstrated significant neo-ossification at the electrode interface, with 10–30% of cochlear volume occupied by fibrosis/bone.⁷⁴ Additionally there appears to be a correlation between the duration of implantation and the total volume of ossification,⁷⁵ however there is a paucity of data to suggest the time course for development of neo-ossification. Murine models have demonstrated intracochlear neo-ossification as soon as 3 weeks after cochlear implantation,⁷⁶ and there are no published data highlighting the time course for intracochlear neo-ossification in large animals such as sheep, pigs, or primates after cochlear implantation. Our study chose a 4-week timepoint to assess cochlear histology as it was expected this would offer enough time for the animal to generate a robust inflammatory response to the implanted biomaterials. Even after only 4 weeks, some tissue density consistent with bone growth within the cochlea was appreciated in the majority of our samples on XRM. This preliminary study, designed to develop a method of applying a consistent thin film coating of human CI electrode arrays, is underpowered to appreciate a significant difference between groups considering that there were relatively low rates of neoossification overall (Supplemental Fig. 2). However, our analysis revealed the majority of neo-ossification occurred predominantly at the basal cochlear turn, where the electrode array is forced to accommodate the spiral geometry of the cochlea. It is known that straight electrodes have a disproportionate impact on lateral wall, and histopathological studies have shown a significant neo-ossification burden at the ascending limb of the basal cochlea.⁹ In our study, coated electrode arrays were also advanced beyond this point but exhibited less neo-ossification as compared to uncoated controls (Fig. 4), though this difference did not meet statistical significance (Supplemental Fig. 2).

The sheep cochlea tapers more readily than the human cochlea during the approach to the apex and as such does not allow for full insertion of a human CI electrode array. The same surgeon performed all implant procedures, and advanced the electrode array until resistance was met often noted by bowing of the extracochlear portion of the array. In our study there is variability in insertion depth across specimens, which is a limitation of this study due to the anatomical differences between the human and sheep cochleae. As seen in Fig 4D, the CBMA-coated CI has an area of redundancy within the basal cochlea, secondary to over insertion into a closed space, however the amount of neoossification related to the redundancy in the basal cochlea or the more apical electrodes remains reduced as compared to uncoated electrode arrays (Fig. 4C). The high water content of zwitterionic hydrogel thin films creates a highly lubricious surface which significantly decreased insertion force and insertional force variation in coated CIs as compared to uncoated electrode arrays in a cadaveric model.³⁹ It is likely that the CBMA-coated CI is able to accommodate the basal cochlear turn with less friction, and thereby less tissue damage and inflammation upon insertion. This observation is in congruence with previous studies correlating “soft” surgical technique with reduced immediate mechanical trauma as well as subsequent inflammatory response.⁷⁷

Histological analysis of cochlear tissues further demonstrates the differences between groups. As the electrode arrays were not fully inserted into the apex of the cochlea, the scala tympani of two regions were selected for histological assessment, at the cochlear base approximating the level of the round window opening (Fig. 6) and at the basal turn of the cochlea as seen in a mid-modiolar section (Fig. 5). Uncoated electrode arrays demonstrate dense infiltration of cells and fibrosis within the scala tympani, as appreciable on H&E and immunostaining. CBMA-coated CIs show significantly less overall cellular infiltration, macrophage infiltration, and fibrotic tissue deposition in the scala tympani at this location as compared with uncoated electrode arrays. This trend is also seen in the basal region of mid-modiolar sections, though at the level of mid modiolar sections, the difference failed to reach statistical significance in this small pilot study. XRM and histology can confirm that the electrode arrays were inserted to this depth as would be appreciated on a mid-modiolar section. A wide range of degrees of inflammatory responses were observed in the sheep (Supplemental Fig. 3), and thus this study is likely underpowered. Having established methods to provide a durable coating that withstands surgical processing and maintains device functionality, long term studies with a larger number of sheep may demonstrate a more substantial difference between groups.

The degree to which the significant reduction in intracochlear inflammation and fibrosis is enabled due to decreased insertion trauma or decreased biofouling to the implant surface requires deeper investigation. It has been shown in patients undergoing cochlear implantation that an early phase and/or a late phase hearing loss may occur, and that one is not necessarily predictive of the other.⁷⁸ Cochlear histology has demonstrated continued thick fibrotic capsules, multinucleated giant cells at the electrode interface and lymphocyte migration in patients even decades after implantation,⁹ suggesting a sustained FBR within the cochlea. Such evidence implies that there are two different mechanisms for intracochlear fibrosis, the short-term component reflecting perioperative factors, such as mechanical insertional trauma, and a late-term component reflecting a delayed inflammatory foreign body response to the electrode array.²⁷

Recently dexamethasone eluting electrode arrays have been developed in an effort to mitigate the inflammatory response to electrode array implantation. Early data with these electrode arrays demonstrate that reduction of the inflammatory response correlates with reduced impedances for months to a few years after implantation.^79,80 This pharmacological strategy to reduce intracochlear inflammation shows definite promise, however there are limited data characterizing the release profile of intracochlear dexamethasone from these implants, and there is the possibility of systemic effects. Furthermore, the long-term durability of a dexamethasone eluting approach remains unknown as the drug elution diminishes relatively rapidly after implantation. It is possible that blunting the initial inflammatory response in the immediate perioperative period with dexamethasone elution confers a sustained reduction of the foreign body response, but the true long term impact has not yet been adequately studied. One significant advantage of the thin film technology developed here is that it is permanent and covalently grafted to the surface of the implant, providing lifelong reduction in intracochlear inflammation and electrode impedance. However, applying any coating of the cochlear implant increases the cross sectional diameter of the implant and hydrogel coatings are classically subject to shear stresses and delamination, though coatings remained intact in this study. A direct comparison of reduction in impedances in dexamethasone eluting and zwitterionic hydrogel coated electrodes is not made in this study due to the differing technologies across device manufacturers, and this is a limitation of this study. Further, the mechanism of impedance reduction across these two technologies differ. Whereas zwitterionic coatings provide an immediate and sustained reduction in impedance that is ~50% less than uncoated arrays, dexamethasone elution only limits the rise in impedance that occurs with development of the FBR; it does not enhance the ionic conductivity at the electrode interface. Further investigation is needed to compare the reduction in intracochlear inflammation between zwitterionic hydrogel coated CI’s and dexamethasone eluting CI’s. A potential synergistic approach with dexamethasone eluting zwitterionic hydrogel coatings may be a promising avenue of investigation in the future.

5. Conclusions

Cochlear implantation leads to substantial inflammation due to both insertional trauma as well as the FBR. Thin film photografted zwitterionic hydrogels provide a highly lubricious interface that minimize insertional trauma, as well as an ultra-low biofouling surface to reduce the foreign body response. CBMA zwitterionic coatings were successfully photografted and photopolymerized to coat the surface of functional human cochlear implant electrode arrays using a custom additive manufacturing mold. This coating was durable, withstanding standard sterilization and operative handling. Additionally, the anti-fouling surface reduces electrode impedances, inflammatory cell migration, and fibrotic tissue deposition within the scala tympani in a large animal model. Zwitterionic hydrogel coatings present a promising strategy to improve performance and hearing outcomes in CI recipients.

Highlights.

• Cochlear implantation leads to a robust intracochlear inflammatory response

• Zwitterionic hydrogel coatings minimize the foreign body response

• CIs were successfully coated with a thin film zwitterionic hydrogel coating

• Coated CIs yielded lower impedances, inflammation, and fibrosis in a sheep model