Postsynaptic neuronal activity promotes regeneration of retinal axons

Abstract

The wiring of visual circuits requires that retinal neurons functionally connect to specific brain targets, a process that involves activity-dependent signaling between retinal axons and their postsynaptic targets. Vision loss in various ophthalmological and neurological diseases is caused by damage to the connections from the eye to the brain. How postsynaptic brain targets influence retinal ganglion cell (RGC) axon regeneration and functional reconnection with the brain targets remains poorly understood. Here, we established a paradigm in which the enhancement of neural activity in the distal optic pathway, where the postsynaptic visual target neurons reside, promotes RGC axon regeneration and target reinnervation and leads to the rescue of optomotor function. Furthermore, selective activation of retinorecipient neuron subsets is sufficient to promote RGC axon regeneration. Our findings reveal a key role for postsynaptic neuronal activity in the repair of neural circuits and highlight the potential to restore damaged sensory inputs via proper brain stimulation.

Keywords: CP: Neuroscience; axons; central nervous system; distal injury; neural activity; optic; postsynaptic neurons; regeneration; retinal ganglion cells; sensory systems; vision.

Figure 1:. Injury to the distal optic tract of adult mice

(A) Schematic of distal injury (red dotted line) to the optic tract showing an anterograde tracer cholera toxin subunit b (Ctβ)-conjugated to Alexa-Fluor-488 injected into the eye to label RGCs and their axonal projections to central visual targets: dorsal lateral geniculate nucleus (dLGN); pretectum; superior colliculus (SC). The grey dotted rectangle shows the magnified regions shown in C and D. (B) Experimental timeline to assess the distal injury and its effect on RGC death. (C, D) Schematic of sagittal (C) and dorsal (D) views of the optic tract and RGC projections into central visual targets. (E) Sagittal sections demarcating the lesion area immunostained for astrocytes (GFAP, green) and microglia (IBA1, magenta). (F-I) Sagittal (F) and dorsal (G) views of normal uninjured RGC axon projections into the pretectum and SC. Sagittal (H) and dorsal (I) views of RGC axon projections 2 weeks after distal injury. The red dotted line indicates the lesion site. (J-M) Whole-mount retinas from sham-uninjured (J) and injured-saline-control (K), injured-CNO-activity (L) mice labeled with an RGC marker (RBPMS). Quantification of RGCs two weeks after distal injury from the contralateral eyes (M). Ordinary one-way ANOVA: p = 0.5857; Tukey’s multiple comparisons test: p = 0.9102, p = 0.7882, p = 0.5675. N = 5 animals/group for sham and activity; N = 4 animals for control. Error bars indicate SEM. Scale bars: 100 μm.

Figure 2:. Non-specific stimulation of the distal optic tract promotes RGC axon regeneration

(A) Schematic of chemogenetic stimulation of neurons in the pretectum relative to the lesion site (grey dotted line). (B-E) Ctβ labeled RGC axon projections (green) in the pretectum, and SC and mCherry labeled neurons in the pretectum expressing hM3Dq-mCherry (blue). White dotted rectangles in B and D show magnified regions in C and E, respectively. (F) Schematic of chemogenetic stimulation: clozapine-N-oxide (CNO) binds modified hM3Dq to increase neural activity within cells. (G) Experimental timeline for stimulating neurons in the pretectum following distal injury. (H-I) Representative images showing Ctβ labeled RGC axon projections in the pretectum and SC. Sagittal sections of the brain with pre-injury Ctβ label (green), post-injury Ctβ label (magenta) from control (H), and neural activity groups (I). The white dotted line indicates the lesion site. (J-K) White rectangles in I and J show magnified regions in J and K, respectively. (L-M) Images in J and K were processed to identify regenerating axons only labeled with post-injury Ctβ. (N) Quantification of the pixel density of regenerating axons (analyzed as shown in P). Individual data points in each graph represent the sum of pixel density from one animal. Mann-Whitney test: ****p<0.0001. N = 10 animals (control), 14 animals (activity). (O) Quantification of the pixel density of spared axons labeled with both pre-injury and post-injury Ctβ (analyzed as shown in P). Mann-Whitney test: n.s. p = 0.8408. N = 10 animals (control), 14 animals (activity). (P) Pre-injury Ctβ label (green) and post-injury Ctβ label (magenta) injected into the eyes of mice can be processed to distinguish regenerating versus spared axons. (Q) Quantification of regeneration as a function of distance. The average pixel density of all animals at each point on the x-axis is plotted. The grey dotted line shows the lesion site, the blue bar indicates the hM3Dq-injection site. Paired t-test to compare the group as a whole: ****p<0.0001. Multiple Mann-Whitney tests to compare control and activity at each individual distance: p = 0.022, 0.259, 0.095, 0.0038, 0.00038, 0.0059, 0.0058, 0.0047, 0.0024, 0.0059, 0.0058, 0.0069. N = 10 animals (control), 14 animals (activity). Error bars indicate SEM. Scale bars: 100 μm. See also Figure S1-S4.

Figure 3:. Selective stimulation of retinorecipient cells promotes regeneration

(A) Cre-dependent Flex-hM3Dq injected into the NOT of Syt17::Cre mice increases neural activity in Cre⁺ cells postsynaptic to RGCs. (B) Experimental timeline to stimulate NOT cells posts distal-injury. (C, F) Coronal sections showing Cre+ cells in the NOT that receive Ctβ labeled RGC input expressing hM3Dq (blue, F). (D-E, G-H) Representative images of coronal sections of the brain labeled with pre-injury Ctβ (green) and post-injury Ctβ (magenta) from control (D, G) and neural activity groups (E, H). Images in D and E are shown processed to identify “regenerating” axons (G, H, respectively). (I) Quantification of the pixel density of “regenerating” axons. Individual data points in represent the sum of pixel density from one animal. Mann-Whitney test: *p = 0.033. N = 3 animals (control) and 7 animals (activity). (J) Quantification of the pixel density of spared axons. Individual data points in represent the sum of pixel density from one animal. Mann-Whitney test: n.s. p >0.9999. N = 3 animals (control) and 7 animals (activity). (K) Quantification of the pixel density of “regenerating” axons as a function of distance.The average pixel density of all animals at each point on the x-axis is plotted. The blue bar in K represents the hM3Dq injection site. The grey dotted line indicates the lesion site. Paired t-test to compare the group as a whole: **p = 0.0079. Multiple unpaired t-tests to compare control and activity at each individual distance: p = 0.14, 0.34, 0.040, 0.025, 0.031, 0.048, 0.095, 0.318 . N = 3 animals (control) and 7 animals (activity). Error bars indicate SEM. Scale bars: 100 μm. (L-N) Retinal whole-mounts from control (L) and activity (M) groups show GFP+ RGCs labeled via retrograde tracing from the NOT. (M’) The white box in M is magnified in M’. (N) Quantification of GFP+ RGCs from both groups. Unpaired t-test: *p = 0.02. N = 4 animals (control) and 3 animals (activity). See also Figure S5-S7.

Figure 4:. Increasing neural activity rescues deficit in optomotor response caused by distal injury

(A) Schematic of the OptoDrum used to measure optomotor response or OMR (Striatech Inc.). The fully-automated software overlays the red, yellow, and green dots (right) to denote head, body, and tail positions. (B) Image captured from a video shows a mouse observing low spatial frequency (0.056 cyc/deg) drifting gratings. (C) Experimental timeline showing two recordings of optomotor response before and after injury. (D-E) Optomotor response for the threshold of spatial frequency tracked by each animal in control and activity groups before injury (‘pre-injury, light grey) and after injury (‘post-injury, dark grey) (D). Two-way ANOVA: **p = 0.0081, n.s. p = 0.5696, 0.1543. Quantification of the defect percentage in control and activity groups after injury plotted. Pre-injury responses were scored as 100, and post-injury responses were calculated as a percentage of the pre-injury response for each animal in each group (E). Two-way ANOVA: *p = 0.0107. N = 5 animals/group. Error bars indicate SEM. (F-G) The length of time each animal tracked the moving stimuli is shown as the tracking duration for both control and activity groups, before injury (“pre-injury,” light grey) and after injury (“post-injury,” dark grey) (F). Two-way ANOVA: **p = 0.0036, n.s. p = 0.5429, 0.8157. Quantification of the defect percentage in tracking in control and activity groups after injury plotted. Pre-injury responses were scored as 100, and post-injury responses were calculated as a percentage of the pre-injury response for each animal in each group (G). Two-way ANOVA: *p = 0.0415. N = 5 animals/group. Error bars indicate SEM.