Inner Limiting Membrane Peel Extends In Vivo Calcium Imaging of Retinal Ganglion Cell Activity Beyond the Fovea in Non-Human Primate

Abstract

Purpose: Adaptive optics scanning light ophthalmoscopy (AOSLO) paired with intravitreal injection of a viral vector coding for the calcium indicator GCaMP has enabled visualization of neuronal activity in retinal ganglion cells (RGCs) at single cell resolution in the living eye. However, the inner limiting membrane (ILM) restricts viral transduction to the fovea in humans and non-human primates, hindering both therapeutic intervention and physiological study of the retina. To address this issue, we explored peeling the ILM before intravitreal injection to expand calcium imaging beyond the fovea in the living primate eye.

Methods: Five eyes from three Macaca fascicularis (aged 3-10 years; 2 males, 1 female) that were immune suppressed with cyclosporine, underwent vitrectomy and ILM peel centered on the fovea prior to intravitreal delivery of 7m8:SNCG:GCaMP8. RGC responses to visual flicker were evaluated using AOSLO calcium imaging 1 to 6 months after intravitreal injection.

Results: Calcium activity was observed in RGCs throughout the ILM peeled area in all eyes, representing a mean eight-fold increase in accessible recording area relative to a representative control eye. RGC responses in the ILM peeled and control eyes were comparable and showed no significant decrease over the 6 month period after the procedure. In addition, we demonstrated that activity can be recorded directly from the retinal nerve fiber layer.

Conclusions: Peeling the ILM is a viable strategy to expand viral access to RGCs for gene therapy, and when paired with GCaMP imaging has the potential to advance visual neuroscience, preclinical evaluation of retinal function, detection of vision loss, and assessment of therapeutic interventions.

Figure 1.

Overview of Study. (A) Following a pars plana vitrectomy, the ILM was stained with brilliant blue dye. (B) Mechanical force lifts the ILM in a circular pattern around the fovea using forceps. (C) Intravitreal injection delivers viral vector to the retina. (D) A blue autofluorescence (BAF) confocal scanning laser ophthalmoscope (cSLO) confirmed the presence of GCaMP in a pattern matching the surgical peel two months post injection. (E) Functional assessments were conducted at retinal imaging locations using a blue 488-nm excitation source in combination with a red LED flashing stimulus. (F) Individual RGCs were segmented and (G) Their responses to the LED stimulus tabulated.

Figure 2.

ILM peeling enables AAV-mediated fluorescence beyond the fovea in macaque. (A) 30°cSLO BAF images centered on the fovea, taken in each eye of M1 including the (A) control eye OS and (B) the ILM peeled eye OD. GCAMP Fluorescence intensity contour (visualized in the bottom, line denoted by a white dashed line) is normalized to the maximum value and smoothed. AOSLO Imaging locations are labeled for each eye. (C) Box plot of individual RGC z-scores at each imaging location, grouped by retinal eccentricity. Group responses for M1 OD are 6.4 ± 4.4 (n = 77) (ON = 18; OFF = 59); 11.52 ± 7.5 (n = 85) (ON = 36; OFF = 49); 6.4 ± 3.9 (n = 40) (ON = 14; OFF = 26). Group responses for the control eye (M1 OS) are 10.0 ± 6.6 (n = 62) (ON = 42; OFF = 20); 10.9 ± 7.6 (n = 73) (ON = 45; OFF = 28); respectively. Taken at 17 weeks after intravitreal injection in OS and 21 weeks OD.

Figure 3.

RGCs remain responsive for at least 6 months after ILM peel. (A) AOSLO GCaMP image used to create segmentation masks at the edge of the ILM peeled area at 9° from the fovea of M1 OD. (B) Responses at 2 (6.7 ± 5.7; n = 27), 3 (6.0 ± 5.2; n = 28), and 5 (6.4 ± 4.0; n = 40) months after intravitreal injection, corresponding with 3, 4, and 6 months after ILM peel. (C) AOSLO GCaMP image within the ILM peeled area at 3.5° from the fovea of M1 OD. (D) Responses at 3 (8.9 ± 6.2; n = 59), 4 (7.8 ± 6.6; n = 68), 5 (12.5 ± 10.2; n = 87), and 6 (11.5 ± 7.5; n = 76) months after intravitreal injection. (E) AOSLO GCaMP image at 3.5° from the fovea of the control eye. (F) Responses at 1 (6.7 ± 3.5; n = 97), 2 (6.3 ± 3.2; n = 74), 3 (8.2 ± 2.9; n = 81), and 4 (10.7 ± 6.0; n = 75) months after intravitreal injection.

Figure 4.

RGC response amplitudes measured in all eyes. (A) (Left) 30° cSLO BAF images centered on the fovea for all experimental eyes, imaging locations shown with blue squares. (Right) Individual RGC z-scores in box plots for M2 OD (6.4 ± 4.4 [n = 77], 11.5 ± 7.5 [n = 85], and 6.4 ± 3.95 [n = 40]) taken 5 months after intravitreal injection. (B) M2 OS (5.4 ± 2.5 [n = 12], 4.4 ± 2.9 [n = 17], 5.2 ± 3.3 [n = 30], and 4.4 ± 3.0 [n = 24]) taken 3 months after intravitreal injection. (C) M2 OD (7.3 ± 4.6 [n = 81], 7.1 ± 6.2 [n = 77], and 6.2 ± 3.7 [n = 48]) taken 5 months after intravitreal injection. (D) M3 OS (7.1 ± 4.3 [n = 52], 4.2 ± 1.5 [n = 31], 4.2 ± 2.1 [n = 12], and 4.9 ± 3.6 [n = 20]) taken 5 months after intravitreal injection. (E) M3 OD (8.3 ± 6.4 [n = 92] and 7.1 ± 4.4 [n = 99]) taken 6 months after intravitreal injection.

Figure 5.

Optical readout of RGC activity from the RNFL. (A) 30°cSLO BAF image centered on the fovea of M1 OD (ILM peeled). (B) AOSLO GCaMP image from the nasal location shown in (A). The green dotted line represents a 3.6° × 2.7° area over which the signal was averaged to produce a response trace from the RNFL in M1 OD. (C) 30°cSLO BAF image centered on the fovea of the control eye (M1 OS). (D) The orange dotted line represents a 3.6° × 2.7° area over which the signal was averaged to produce a response trace from the RNFL in M1 OS. (E) RNFL responses to the visual stimulus in M1 OD (3.3 ± 0.02), and M1 OS, the control eye (1.4 ± 0.02).

Figure 6.

Structural imaging reveals minor morphological changes following ILM peel in (A) M1 OD, (B) M2 OS, (C) M2 OD, (D) M3 OS, and (E) M3 OD. (Left) 55° Color fundus images of white arrows indicate areas consistent with CMDSs. (Right) B-scan OCT of the foveal pit between 2 and 24 weeks after ILM peel and 1 to 142 days after intravitreal injection.

Figure 7.

Circular retinal defects correspond to the absence of ganglion cell fluorescence. (A) AOSLO calcium image of two hypofluorescent areas in M3 OS. Dendrites are visible in this region. (B) Color fundus (top left) and cSLO BAF (top right) of M3 OS showing two hypofluorescent areas that are investigated with OCT (bottom). White line indicates OCT location. (C) Montaged fluorescence AOSLO image reveals dimples in OCT correspond with the two hypofluorescent areas in AOSLO. (D) Color fundus image of a retinal location with no macular dark spots but a hypofluorescent region in the cSLO BAF (E). OCT and AOSLO of the hyperfluorescent region seen in (D). (F) Color fundus image of a retinal location with CMDS streaks but no visible RNFL dimples in OCT or (G) hypofluorescent RGC expression visible in AOSLO.