Increased Connexin36 Phosphorylation in AII Amacrine Cell Coupling of the Mouse Myopic Retina

Abstract

Myopia is a substantial public health problem worldwide. In the myopic retina, distant images are focused in front of the photoreceptors. The cells and mechanisms for retinal signaling that account either for emmetropization (i.e., normal refraction) or for refractive errors have remained elusive. Gap junctions play a key component in enhancement of signal transmission in visual pathways. AII amacrine cells (ACs), coupled by connexin36, segregate signals into ON and OFF pathways. Coupling between AII ACs is actively modulated through phosphorylation at serine 293 via dopamine in the mouse retina. In this study, form deprivation mouse myopia models were used to evaluate the expression patterns of connexin36-positive plaques (structural assay) and the state of connexin36 phosphorylation (functional assay) in AII ACs, which was green fluorescent protein-expressing in the Fam81a mouse line. Single-cell RNA sequencing showed dopaminergic synapse and gap junction pathways of AII ACs were downregulated in the myopic retina, although Gjd2 mRNA expression remained the same. Compared with the normal refractive eye, phosphorylation of connexin36 was increased in the myopic retina, but expression of connexin36 remained unchanged. This increased phosphorylation of Cx36 could indicate increased functional gap junction coupling of AII ACs in the myopic retina, a possible adaptation to adjust to the altered noisy signaling status.

Keywords: amacrine cell; ganglion cell; gap junction (connexin); myopia; retina.

FIGURE 1

Enrichment pathways and plots of AII amacrine cells (ACs). (A) Enrichment plot of dopaminergic synapse. NES, normalized enrichment score; FDR, false discovery rate q value, <0.25 is statistically significant. (B) Enrichment plot of gap junction. (C) Retinal Gjd2 mRNA expression in treated eyes and fellow eyes after form deprivation for 2 days, no significant difference exists. (D) tSNE analysis of AII AC clusters in treated and follow eyes showed that there was no significant difference in cell numbers between treated and follow eyes.

FIGURE 2

Measurement of refractive errors (RE) in mice by photorefraction and OCT. (A) Axial length (AL) of the mouse eye was measured from the surface of the cornea to the RPE layer by OCT. (B) The optical axis of the mouse eye can be precisely aligned with dynamic scan control of OCT. (C) AL measured by OCT could be used to predict RE in 8-week-old WT mice. Refractive errors of 30 eyes of WT mice (photorefraction) fitted by linear regression (red line, R² = 0.042) matched the data of AL estimated from OCT (black line, R² = 0.99). The difference between these two methods is close to 0 (blue line, R² = 0.1%). Calculation: RE = 38 – 1.544/(1.03 × AL).

FIGURE 3

Defocused images altered the signaling responses in RGCs. Perievent rasters histogram of OFF (A–C) α-RGC and ON (D–F) α-RGC responses (middle images) to 0.5-s flashed 0.002 cycles/degree light stimuli. Light spot was maintained in 125 μm when defocused –10 D (B,E), whereas light intensity decreased from I = 5.1 × 10⁴ to 5.0 × 10⁴R*/rod/s. OFF α-RGC light responses decreased after a defocused image was projected (B). Peristimulus time histogram of the OFF α-RGC showed that the probability of spikes decreased from 1.28 to 0.61 (–10-D defocus), then back to 1.21 after refocusing. ON α-RGC (D) decreased from 1.05 to 0.7 (–10-D defocus) and then back to 1.05 after refocusing. (G) and (H) showed OFF α-RGC (A–C) and ON α-RGC (D–F) with Neurobiotin filling after recording. Lower images show the ganglion cell branched in OFF and ON layer with anti-ChAT labeling (blue). Scale bar: (A–F) 75 μm, (G,H) 20 μm.

FIGURE 4

Ser293 antibody labeling patterns in mouse retina. (A–C) Labeling pattern of phospho-Ser293 antibody in vertical section of mouse retina. (A) Ser293-P antibody-labeled (red) abundant punctate structures in the inner plexiform layers (IPL) and some in the outer plexiform layer (OPL). (B) Labeling with monoclonal Cx36 antibody (green) shows the labeling of Cx36 also in the OPL and IPL. (C) Merged image of (A) and (B) shows multiple plaques identified as Ser-293-P colocalized with Cx36 antibody labeling (yellow). (D) In the negative control of Cx36 KO mouse retina, labeled punctate of Cx36 antibody and Ser293-P were absent in both OPL and IPL. Scale bar is 20 μm.

FIGURE 5

Phospho-Ser293 and monoclonal anti-Cx36 antibody labeled prominently in IPL. (A) GFP-labeled AII amacrine cell in Fam 81 mouse retina was targeted and injected with Neurobiotin to show single AII AC’s soma–dendritic morphology in Z stack by 3D reconstruction image. (B) AII AC double labeled with phospho-Ser293 (red) and monoclonal anti-Cx35/36Cx36 (blue) puncta located predominately on the arboreal dendrites (blue) strata three to five layers. (C) Single layer of AII AC in the lobular layer. (D) Single layer of AII AC in the arboreal dendrite layer. Phospho-Ser293 puncta are also colocated on the arboreal dendrites (red) strata layer 5. (E) Coupled AII ACs displayed with Neurobiotin injection. Only cell bodies coupled via gap junction showed clearly. (F) To show the dendritic gap junctional coupling in AII ACs, two nearby AII ACs were injected with Neurobiotin. (G) Dendrites between the two nearby AII ACs injected with Neurobiotin in the S5 layer. (H) Labeled with monoclonal anti-Cx35/36 (blue). Cx36 puncta localized on dendrites of AII ACs (white arrows). Scale bar: (A–D) 10 μm, (E,F) 5 μm, (G,H) 2 μm.

FIGURE 6

Phospho-Ser293 antibodies specifically recognizes Cx36 in the whole mount mouse retina. (A–D) Confocal stack sections in stratum 5 of the IPL of the myopic mouse retina: mCx36 labeled in red and its phosphorylated form, Ser293-P (green), are present with similar punctate labeling. The magnified areas show the merged images of phosphorylated Cx36, reflected by yellow color. (E–H) Phospho-Ser293 antibody recognizes Cx36 in the control mouse retina. Images are 2-μm deep stacks. Scale bar: (A–C,E–G) 5 μm, (D,H) 2 μm.

FIGURE 7

Quantification of phosphorylation of Cx36 gap junctions in AII amacrine cells in mouse myopic retinas. For S293, the density of phosphorylation reflected by detectable Ser293-P labeling (A), the percentage of Ser293-P of Cx36 phosphorylated rate (B), and the size of Ser293-P plaque (C) were significantly increased in myopic retinas compared to controls. The density of mCx36 labeling (A) and the size of mCx36 plaque (C) did not differ between myopic and control mouse retinas. The data are presented as averages. Error bars are SEM. Significance is based on Wilcoxon signed ranks test, where *0.01 < P < 0.05, not statistically significant P > 0.05.

FIGURE 8

Quantitative comparison of trace coupling between AII amacrine cell changes following application of agonist and antagonist of dopamine D₁ receptors. (A) Flat mount view of a group of tracer coupled AII ACs in the mouse retina following injection of one cell with Neurobiotin. (B) D₁R agonist SKF38393 10 μM reduced the extent of Neurobiotin diffusion in AII AC coupling. (C) D₁R antagonist SCH23390 5 μM dramatically increased tracer coupling of AII ACs. (A–C) planes of focus are on the AII cell somata in the proximal inner nuclear layer. Scale bar: (A,B) 10 μm, (C) 50 μm. (D) Box graph showing the difference in the number of coupled AII ACs with D₁R agonist SKF38393 10 μM and D₁R antagonist SCH23390 5-μM application. There was a statistically significant difference (asterisk, P < 0.01) in the number of coupled AII ACs. (E) Box graph showing the difference in the extent of coupled AII ACs somata with the D₁R agonist SKF38393 10 μM and D₁R antagonism SCH23390 5-μM application There was a statistically significant difference (asterisk, P < 0.01) in the number of coupled AII ACs.

FIGURE 9

Illustration of the plasticity of gap junctions in AII amacrine cells in myopic and normal retinas. The diagram provides a summary of the intracellular pathways in which the plasticity of retinal gap junctions in AII ACs is affected by dopamine. In the emmetropic retina, images are focused on the retina (left) with dopamine, released from dopaminergic amacrine cells, and bound to D₁-like receptors (D₁R) activating cAMP-dependent protein kinase α (PKA) and protein phosphatase 2A (PP2A), which in turn dephosphorylates Cx36, thereby causing reduced gap junction conductance. In contrast, decreased dopamine increases the conductance of Cx36 in AII amacrine cells by increasing phosphorylation in the defocused status of the myopic retina (right).