Effect of shape and coating of a subretinal prosthesis on its integration with the retina

Abstract

Retinal stimulation with high spatial resolution requires close proximity of electrodes to target cells. This study examines the effects of material coatings and 3-dimensional geometries of subretinal prostheses on their integration with the retina. A trans-scleral implantation technique was developed to place microfabricated structures in the subretinal space of RCS rats. The effect of three coatings (silicon oxide, iridium oxide and parylene) and three geometries (flat, pillars and chambers) on the retinal integration was compared using passive implants. Retinal morphology was evaluated histologically 6 weeks after implantation. For 3-dimensional implants the retinal cell phenotype was also evaluated using Computational Molecular Phenotyping. Flat implants coated with parylene and iridium oxide were generally well tolerated in the subretinal space, inducing only a mild gliotic response. However, silicon-oxide coatings induced the formation of a significant fibrotic seal around the implants. Glial proliferation was observed at the base of the pillar electrode arrays and inside the chambers. The non-traumatic penetration of pillar tips into the retina provided uniform and stable proximity to the inner nuclear layer. Retinal cells migrated into chambers with apertures larger than 10 mum. Both pillars and chambers achieved better proximity to the inner retinal cells than flat implants. However, isolation of retinal cells inside the chamber arrays is likely to affect their long-term viability. Pillars demonstrated minimal alteration of the inner retinal architecture, and thus appear to be the most promising approach for maintaining close proximity between the retinal prosthetic electrodes and target neurons.

Fig. 1.

SEM of three-dimensional implant structures. A. Implant with an array of pillars at three densities, with center-to-center distances of 60 μm, 40 μm and 20 μm. B. High magnification SEM of the pillar array. Pillars are 10 μm in diameter and 65 μm in height. All scale bars in this figure are 100 μm. C. Two microfabricated layers of the chamber structures prior to adhesion to the basal membrane. Chamber sizes are 40 and 20 μm, and aperture sizes are 20 and 10 μm. D. High magnification view of the chamber array. The apertures can be seen clearly in the center of the chambers.

Fig. 2.

A. Implantation tool prior to loading the implant between the grooves of the fork. Insert: a diagram of the front view of the implant loaded in the grooves of the fork. B. View of the tool with a fork fully withdrawn, demonstrating the central road that prevents the implant from sliding back when the fork is withdrawn.

Fig. 3.

A. Wild type rat retina. B. RCS rat retina 45 days post-natal (P45). C. RCS rat retina with a flat SU-8 implant in the subretinal space 6 weeks post-op. A fibrotic seal running along the length of the implant is denoted with the left arrow. A region of gliosis separating the implant from the INL by 40 μm is shown by the right arrow. Scale bar is 50 μm.

Fig. 4.

Comparison of typical flat implants with different coatings 6 weeks post-op. A. SiO₂ coating appears to induce significant fibrosis over the implant; B. IrOx causes a mild gliosis above to the implant, pointed by an arrow; Parylene-C coating allows the INL to settle down very close to the implant. INL is separated from the upper surface of the implant by only 15–30 μm. Scale bar is 50 μm.

Fig. 5.

Edges of the implants with silicon oxide (A), IrOx (B), and parylene (C) coatings. A massive encapsulating membrane in front of the silicon-oxide implant and a very fine membrane at the edge of the IrOx sample are denoted by the arrows. The gap (*) between the implant and the tissue in A is an artifact of histological preparation. Scale bar is 50 μm.

Fig. 6.

The chamber structure implanted into P45 RCS rat sub-retinally for six weeks. The three chambers on the right have 20 μm apertures, the two on the left are 10 μm. A. Typical histology shows cell bodies migrating through wider apertures while only processes migrate through 10 μm apertures. Artifactual folds are marked with a *. B. γGE signatures: CMP identifies ON-cone bipolar, OFF-cone/rod bipolar cells as well as glycinergic and GABAergic amacrine cells migrating through the larger apertures that retain their adult small molecular phenotypes. Scale bar is 50 μm. C. TQE signatures: CMP identifies Müller cells that migrate through the apertures and appear to hypertrophy inside the chambers. Scale bar is 50 μm. γAC: GABAergic GAC: glycinergic amacrine cell, MC: Müller cell, ON: ON-cone bipolar cell, OFF: OFF-cone/rod bipolar cell.

Fig. 7.

Retinal blood vessels migrate into the chambers (pointed by the arrows), as seen at 6 weeks post-op. Folds (dark horizontal curves inside the chambers) are denoted with *.

Fig. 8.

Pillar implant in subretinal space of RCS rat 6 weeks post-op. A. Histology of the area with 40 μm pillar spacing. Artifactual folds from sectioning are marked with a *. B. Histology of the area with 20 μm pillar spacing. C. CMP results with 40 μm pillar spacing. γGE signatures identify ON-cone bipolar, OFF-cone/rod bipolar cells as well as glycinergic and GABAergic amacrine cells that retain their adult small molecular phenotypes. D. Similar CMP results with 20 μm pillar spacing. E. τQE signatures of CMP with 60 μm pillar spacing demonstrating neurons retaining their normal phenotype, with Müller cells becoming hypertrophic and filling remnant space in-between the implant pillars. Note the novel plexiform layer shown in the box representing likely sprouting from contributing bipolar, amacrine and horizontal cells. F. Similar CMP results with 20 μm pillar spacing. γAC: GABAergic GAC: glycinergic amacrine cell, MC: Müller cell, ON: ON-cone bipolar cell, OFF: OFF-cone/rod bipolar cell. GAA: GABAergic GLA: glycinergic amacrine cell, MC: reactive Müller cell layer, ON: ON bipolar cell, OFF: OFF bipolar cell.