Starburst amacrine cells amplify optogenetic visual restoration through gap junctions

Abstract

Ectopic induction of optogenetic actuators, such as channelrhodopsin, is a promising approach to restoring vision in the degenerating retina. However, the cell type-specific response of ectopic photoreception has not been well understood. There are limits to obtaining efficient gene expression in a specifically targeted cell population by a transgenic approach. In the present study, we established a murine model with high efficiency of gene induction to retinal ganglion cells (RGCs) and amacrine cells using an improved tetracycline transactivator-operator bipartite system (KENGE-tet system). To investigate the cell type-specific visual restorative effect, we expressed the channelrhodopsin gene into RGCs and amacrine cells using the KENGE-tet system. As a result, enhancement in the visual restorative effect was observed to RGCs and starburst amacrine cells. In conclusion, a photoresponse from amacrine cells may enhance the maintained response of RGCs and further increase or improve the visual restorative effect.

Keywords: gene therapy; optogenetics; retina; starburst amacrine cells; visual restoration.

Graphical abstract

Figure 1

Gene induction of RGCs and SACs in M4-Figure 1YC and RGCs in 5B-YC. In the M4-YC mouse retina, we identified the expression of YC (green) in RGCs and amacrine cells within sections (A–D) In the 5B-YC mouse retina, we identified the expression of YC (green) in RGCs in sections (E–H). Co-expression of the RGC marker RBPMS in flat-mounted retinas of M4-YC(I) and 5B-YC(J). Percentage of YC-positive cells in RBPMS-positive (K) or DAPI-positive cells (L) and RBPMS-positive cells in YC-positive cells (L) in both lines from confocal flat-mounted GCL (n = 3 retinas each). Regions were chosen in each quadrant, and we obtained RBPMS, DAPI-positive, YC-positive, and co-labeled cells. Co-expression of the SAC marker ChAT in flat-mounted retinas of M4-YC in INL (M–O) and GCL (Q–S). Percentage of ChAT-positive cells in YC-positive cells in M4-YC mice from INL (P) and GCL (T) (n = 3 retinas each). Error bars represent the SEMs. INL, inner nuclear layer; ONL, outer nuclear layer. Scale bar, 50 μm in (E); 100 μm in (B–D), (H), (I), (K), (M), (Q), and (R); 400 μm in (N); and 500 μm in (F), (G), (O), and (P).

Figure 2

M4-ChR2 mouse shows a higher visual restoration effect (A–C) Raster plots and peri-stimulus time histogram (PSTH) of light stimulation from single RGCs of MNU-treated tetO-ChR, M4-ChR2, and 5B-ChR2 mice. Light intensity was 13.6 log photons/cm²/s, and the duration was 1.0 s. (D, E) The averaged rate histogram after filtering with the Gaussian function from M4-ChR2 and 5B-ChR2 mice. At least 10 trials were conducted for each cell. The gray areas around the averaged traces represent the SEM. (F) Maintained-to-peak ratio of the spiking responses recorded. The maintained time frame is 0.4–1.0 s from light stimulation as shown in H (n = 7 retinas, 164 cells in MNU-treated M4-ChR2 mice, and n = 4 retinas, 127 cells in MNU-treated 5B-ChR2 mice). Error bars represent SEMs. ∗∗∗p < 0.001. Student’s 2-tailed t test. (G) Schematic image of VEP measurement. (H) Representative VEP traces from MNU-injected and control mice. (I) The average amplitude of the VEPs in the control tetO-ChR mice (n = 4), MNU-treated tetO-ChR mice (n = 8), M4-ChR2 mice (n = 14), and 5B-ChR2 mice (n = 12) at 10 weeks of age. It was stimulated with a light stimulus intensity of 100-ms pulses of white LED 4,000 cds/m². Signals were low-pass filtered at 300 Hz and averaged over the 60 trials. (J) Schematic image of the LDT testing. Mice were tested in a 30 × 45 × 30-cm box containing equally sized bright (200 lux at ground level) and dark chambers connected by a 5 × 5-cm opening, across which the mice could move freely. Visible mice feel uneasy in bright places, so staying time in the bright half gets shorter. (K) Percent time in the bright half at 10 min of control (tetO-ChR2 mice) (n = 4), MNU-injected tetO-ChR2 mice (n = 8), MNU-injected M4-ChR2 mice (n = 15), and MNU-injected 5B-ChR2 mice (n = 8) measured from LDT testing. (L) Schematic view of the OKR system. The images of the right or left eyes are captured by a CCD camera placed on the same side. During measurement, the contralateral eyes are covered with aluminum foil. Visual stimulation is presented on three LCD monitors around the mouse, the head of which is fixed in the middle. (M) The average eye velocities of the control tetO-ChR2 mice (n = 3), MNU-injected tetO-ChR2 mice (n = 3), M4-ChR2M4-ChR2 mice (n = 9), 5B-ChR2 mice (n = 5), rd1;tetO-ChR2 mice (n = 8), rd1;M4-ChR2M4-ChR2 mice (n = 10), and rd1;5B-ChR2 mice (n = 5) measured from the OKR system at 10 weeks of age. All error bars represent the SEMs. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. One-way ANOVA testing.

Figure 3

Acetylcholine and gap junctions were involved in the maintained response (A, D, G, J) Mean ± SEM of exemplar cell response firing rate recorded during normal Ames' medium superfusion (left), in synaptic block (middle), and after washout (right). MNU-treated M4-ChR2 mice with L-AP4 block (n = 3 retinas, 47 cells) (A), MNU-treated 5B-ChR2 mice with L-AP4 block (n = 3 retinas, 45 cells) (D), MNU-treated M4-ChR2 mice with MFA block (n = 3 retinas, 85 cells) (G), and MNU-treated M4-ChR2 mice with atropine block (n = 3 retinas, 33 cells) (J). The gray areas around the averaged traces represent the SEM. (B, C, E, F, H, I, K, L) Averaged normalized peak firing rate and maintained rate. The maintained time frame is 0.4–1.0 s from light stimulation. Light intensity was 13.6 log photons/cm²/s. All error bars represent the SEM. ∗∗∗p < 0.001. One-way ANOVA and Tukey’s test.

Figure 4

Induction efficiency into SACs tended to affect visual restoration (A) The rAAV2-CAG-ChR2-tdimer2-WPRE expression cassette. ITR, inverted terminal repeat; WPRE, woodchuck hepatitis virus posttranscriptional regulatory element. (B–E) Co-expression of the RGC marker RBPMS (B) and SAC marker ChAT (D) and tdimer2 in flat-mounted retinas of rAAV2-CAG-tdimer2-WPRE treated retinas in rd1 mice. (C, E) Percentage of YC-positive cells in M4-YC and 5B-YC mice and tdimer2-positive cells in rAAV-treated mice in RBPMS-positive (C) or ChAT-positive cells (E) from confocal flat mounted retina (n = 3 retinas each). Regions were chosen in each quadrant, and we obtained RBPMS, ChAT-positive, YC/tdimer2-positive, and co-labeled cells. (F, G) Average VEP traces (F) and quantification of its amplitudes (G) from control (tetO-ChR2) (n = 3), rd1:tetO-ChR2 (n = 6), rd1;M4-ChR2 (n = 12), rd1;5B-ChR2 (n = 12) and rAAV-treated rd1:tetO-ChR2 mice (n = 12) at 10 weeks of age. It was stimulated with 100-ms pulses of white LED 4,000 cds/m² light stimulus intensity. Signals were low-pass filtered at 300 Hz and averaged over the 60 trials. (H) Percent time in bright half at 10 min in control (tetO-ChR2) (n = 3), rd1:tetO-ChR2 (n = 6), rd1;M4-ChR2 (n = 12), rd1;5B-ChR2 (n = 12), and rAAV-treated rd1:tetO-ChR2 mice (n = 12) measured from LDT testing. (I–K) Correlation between transfection efficiency into SACs (tdimer2-positive cells/ChAT-positive cells in INL) and maintained to peak rate (I) (n = 24), VEP amplitude (J) (n = 24), and percent time in the bright half in LDT testing (K) (n = 12). All error bars represent the SEMs. INL, inner nuclear layer; GCL, ganglion cell layer. Scale bars, 50 μm in (B), (D), n.s., not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. unpaired t test (E), Games-Howell test (C, G, H), and Pearson’s correlation coefficient (I–K).