Axon hyperexcitability in the contralateral projection following unilateral optic nerve crush in mice

Abstract

Optic neuropathies are characterized by degeneration of retinal ganglion cell axonal projections to the brain, including acute conditions like optic nerve trauma and progressive conditions such as glaucoma. Despite different aetiologies, retinal ganglion cell axon degeneration in traumatic optic neuropathy and glaucoma share common pathological signatures. We compared how early pathogenesis of optic nerve trauma and glaucoma influence axon function in the mouse optic projection. We assessed pathology by measuring anterograde axonal transport from retina to superior colliculus, current-evoked optic nerve compound action potential and retinal ganglion cell density 1 week following unilateral optic nerve crush or intraocular pressure elevation. Nerve crush reduced axon transport, compound axon potential and retinal ganglion cell density, which were unaffected by intraocular pressure elevation. Surprisingly, optic nerves contralateral to crush demonstrated 5-fold enhanced excitability in compound action potential compared with naïve nerves. Enhanced excitability in contralateral sham nerves is not due to increased accumulation of voltage-gated sodium channel 1.6, or ectopic voltage-gated sodium channel 1.2 expression within nodes of Ranvier. Our results indicate hyperexcitability is driven by intrinsic responses of αON-sustained retinal ganglion cells. We found αON-sustained retinal ganglion cells in contralateral, sham and eyes demonstrated increased responses to depolarizing currents compared with those from naïve eyes, while light-driven responses remained intact. Dendritic arbours of αON-sustained retinal ganglion cells of the sham eye were like naïve, but soma area and non-phosphorylated neurofilament H increased. Current- and light-evoked responses of sham αOFF-sustained retinal ganglion cells remained stable along with somato-dendritic morphologies. In retinas directly affected by crush, light responses of αON- and αOFF-sustained retinal ganglion cells diminished compared with naïve cells along with decreased dendritic field area or branch points. Like light responses, αOFF-sustained retinal ganglion cell current-evoked responses diminished, but surprisingly, αON-sustained retinal ganglion cell responses were similar to those from naïve retinas. Optic nerve crush reduced dendritic length and area in αON-sustained retinal ganglion cells in eyes ipsilateral to injury, while crush significantly reduced dendritic branching in αOFF-sustained retinal ganglion cells. Interestingly, 1 week of intraocular pressure elevation only affected αOFF-sustained retinal ganglion cell physiology, depolarizing resting membrane potential in cells of affected eyes and blunting current-evoked responses in cells of saline-injected eyes. Collectively, our results suggest that neither saline nor sham surgery provide a true control, chronic versus acute optic neuropathies differentially affect retinal ganglion cells composing the ON and OFF pathways, and acute stress can have near-term effects on the contralateral projection.

Keywords: degeneration; glaucoma; hyperexcitability; retinal ganglion cells; traumatic optic neuropathy.

Graphical abstract

Figure 1

Unilateral crush induces axonal hyperexcitability in the contralateral nerve. (A, left) IOP elevation following a single unilateral injection of polystyrene microbeads (1.5 μl, n = 14) or saline (1.5 μl, n = 14; P < 0.05). (A, right) Mean IOP is elevated by 31% in microbead-injected eyes compared with saline controls (P < 0.001). Mean IOP corrected from a published regression based on cannulation measurements (red dotted line, see Reitsamer et al.) resulted in a 5% increase in microbead IOPs and an 8% decrease in saline compared with our TonoPen XL measurements. (B) Longitudinal optic nerve sections immunolabelled for GFAP (middle panels) and counterstained with DAPI (blue) in (left) sham and (right) crush nerves. Intact CTB tracing indicates robust anterograde transport in sham nerves. In crushed nerves, deficits in CTB fluorescence indicates injury site (dashed lines). Scale bar = 200 µm. (C) CTB transport (top panels) to the SC and corresponding intensity heat maps (bottom) following (left) microbead injection and (right) ONC. Circles indicate optic discs. Scale bar = 500 µm. (D) Percentage of intact CTB transport to the SC is similar between eyes subjected to ocular hypertension (OHT) and saline injection (P = 0.55, n = 4). Transport is significantly diminished 1 week (1Wk) post ONC compared with sham controls (P < 0.0001, n = 8). (E) Representative current-evoked optic nerve compound action potential (CAP) traces from (left) naïve, saline, 1Wk OHT, (right) sham and 1Wk ONC . (F) Optic nerve CAP responses of naïve, saline and 1Wk OHT were statistically similar (P ≥ 0.67, n ≥ 5). ONC significantly reduced CAP responses compared with naïve (P = 0.007, n ≥ 10) and sham nerves (P < 0.0001, n = 17). Sham optic nerve CAP responses dramatically increased compared with naïve (P < 0.0001, n ≥ 10) and saline optic nerves (P < 0.0001, n ≥ 4). We normalized optic nerve CAP area to recording pipette resistance. Statistics: (A) Unpaired t-test. (D and F) One-way ANOVA and Tukey’s post hoc test. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data are presented as mean ± SEM.

Figure 2

Reduced expression of NaV1.6 and paranode morphology in experimental nerves. Confocal micrographs of longitudinal optic nerve sections in mice subjected to saline- and microbead injection or sham surgery and ONC. (A) Immunostaining of Caspr1-labelled paranodes and Na_V1.6 within nodes of Ranvier (scale bar = 15 µm). Insets are higher magnification micrographs showing example node-paranode complexes (scale bar = 5 µm). (B) Immunostaining of Caspr1-labelled paranodes and Na_V1.2 (scale bar = 15 µm). Insets are higher magnification micrographs showing representative node-paranode complexes (scale bar = 5 µm). (C, left) Compared with nerves from naïve mice, NaV1.6 immunolabelling decreased in nerves from all experimental animals (P < 0.008, n ≥ 592 nodes, n ≥ 3 animals). After 1Wk ONC, Na_V1.6 intensity also significantly diminished compared with optic nerves from 1Wk OHT mice (P = 0.0027, n ≥ 49, n ≥ 3 animals). (C, right) We did not detect a significant difference in NaV1.2 expression between naïve, saline, sham and OHT nerves. NaV1.2 expression significantly increased in sham nerves compared with nerves from saline-injected eyes (P = 0.0104). Crush significantly reduced NaV1.2 localization compared with naïve and sham nerves (P ≤ 0.0236). Naïve: n = 77 nodes, saline: n = 26 nodes, sham: n = 65 nodes, OHT: 29 nodes, ONC: 0 nodes. Animal n ≥ 3. (D, left) We did not detect a significant difference in node length between conditions (P ≥ 0.201). (D, right) Optic nerve axon paranode length of naïve and saline-injected eyes were similar (P > 0.99). Compared with naïve, paranode length significantly decreased in sham, OHT and ONC nerve (P ≤ 0.0205, n ≥ 49). OHT significantly reduced paranode extent relative to nerves from saline-injected eyes (P = 0.0006, n ≥ 509). Animal n ≥ 0.3 Statistics: (C, D) Kruskal–Wallis one-way ANOVA, Dunn’s post hoc. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data are presented as mean ± SEM.

Figure 3

Optic nerve crush depolarizes RMP and decreases light-evoked responses of αON-S RGCs. (A) Confocal micrograph of Alexa 555 filled naïve αON-S RGC. Retinas were immunolabelled for SMI-32. Orthogonal rotation shows αON-S RGC dendrites ramify in the ON sublamina of the inner plexiform layer defined by choline acetyltransferase (ChAT). Arrow indicates RGC axon. Scale bars = 50 µm. (B) Reconstructed and skeletonized αON-S RGCs from saline, OHT, sham and ONC retinas. Scale bars = 50 µm. (C) ONC depolarized αON-S RGCs RMP compared with like cells from naïve (P = 0.05, n ≥ 16) and sham retinas (P = 0.03, n ≥ 11). (D, left) Averaged histograms (3 ms bins) of light-evoked response of naïve, saline and sham αON-S RGCs. (D, right) ONC reduced light responses of αON-S RGCs compared with OHT (right). (E) Compared with like cells from naïve retinas, light responses (mean, integrated and peak) of αON-S RGCs from sham and saline retinas were similar (P > 0.99, n ≥ 12). ONC significantly reduced light-induced responses (mean, integrated, peak) of αON-S RGCs compared with naïve (P < 0.0001, n ≥ 17). Peak light response is significantly greater in sham αON-S RGCs compared with saline αON-S RGCs (P = 0.02, n ≥ 12). After 1Wk, there is no difference in mean, integrated, or peak light responses for OHT αON-S RGCs compared with saline αON-S RGCs (P > 0.99, n ≥ 16). Statistics: (C) One-way ANOVA and Tukey’s post hoc. (E) Kruskal–Wallis one-way ANOVA, Dunn’s post hoc. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data presented as mean ± SEM.

Figure 4

Optic nerve crush diminishes light-evoked responses of αOFF-S RGCs. (A) Confocal image of Alexa 555 filled αOFF-S RGC in a naïve retina immunolabelled against SMI-32. Orthogonal rotation shows αON-S RGC dendrites project in and beyond the OFF sublamina of the inner plexiform layer as defined by ChAT. Arrow shows the RGC axon. Scale bars = 50 µm. (B) Example αOFF-S RGCs, reconstructed and skeletonized, from saline, OHT, sham and ONC retinas. Scale bars = 50 µm. (C) OHT significantly depolarized αOFF-S RGCs RMP compared with like cells of naïve and saline retinas (P ≤ 0.023, n ≥ 8). (D, left) Averaged histogram of spontaneous and light-evoked spiking of αOFF-S RGCs from naïve, saline and sham retinas. (D, right) ONC reduced light responses of αOFF-S RGCs compared with OHT. (E) Light responses (mean, integrated and peak) of αOFF-S RGCs from naïve, saline and sham eyes were similar (P ≥ 0.98, n ≥ 12). ONC significantly blunted mean (P ≤ 0.0001), integrated (P ≤ 0.001) and peak (P ≤ 0.0003) light-driven responses for ONC αOFF-S RGCs compared with naïve and sham αOFF-S RGCs (n ≥ 10). Statistics: (C, E) One-way ANOVA and Tukey’s post hoc. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data are presented as mean ± SEM.

Figure 5

Optic crush enhances current-evoked spiking in αON-S RGCs of the sham nerve. (A, left) αON-S RGCs of sham nerves produce higher spike rates than like cells from naïve and saline eyes to depolarizing currents up to 140 pA (n ≥ 7). (A, right) Current-evoked spike rates of αON-S RGCs from ONC eyes are similar to αON-S RGCs from OHT eyes (right) (n ≥ 13). (B) αON-S RGCs from naïve and saline eyes produce similar mean current-evoked spike rates (P = 0.12). Mean current-evoked spike rate is significantly greater in sham αON-S RGCs compared with naïve αON-S RGCs (P = 0.002). (C, left) αOFF-S RGCs from naïve, saline and sham eyes produce similar spike rates in response to depolarizing currents up to 180 pA (n ≥ 12). (C, right) ONC reduced responses to depolarizing current in αOFF-S RGCs versus OHT αOFF-S RGCs (P ≤ 0.01, n ≥ 6). (D) Mean current-evoked spike rate is decreased for saline αOFF-S RGCs compared with naïve αOFF-S RGCs (P < 0.01). OHT increases αOFF-S RGC spike rate compared with saline αOFF-S RGCs (P < 0.0001). Conversely, ONC αOFF-S RGCs have a significantly lower mean current-evoked spike rate compared with sham αOFF-S RGCs (P < 0.0001). Statistics: (A, C) Two-way repeated measures ANOVA with Bonferroni’s post hoc. (B) Kruskal–Wallis one-way ANOVA, Dunn’s post hoc. (D) One-way ANOVA and Tukey’s post hoc. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data presented as mean ± SEM.

Figure 6

Unilateral optic nerve crush produces bilateral pro-degenerative responses in αON-S RGCs. (A) Representative confocal micrographs of SMI-32 immunolabelled whole-mount retinas. Scale bars = 25 µm. (B) Somatic SMI-32 intensity increased in sham αON-S RGCs compared with like cells from naïve (P = 0.04, n ≥ 7) and saline retinas (P < 0.001, n ≥ 9). ONC significantly reduced SMI-32 expression in αON-S RGCs compared with sham αON-S RGCs (P = 0.01, n ≥ 9). (C) Soma areas are larger in sham αON-S RGCs compared with naïve (P ≤ 0.02) and saline αON-S RGCs (P = 0.006, n ≥ 9). ONC decreased αON-S RGC soma area relative to sham (P = 0.02, n ≥ 9). (D, left) Sholl analysis indicates similarity in dendritic complexity of αON-S RGCs from naïve, saline and sham eyes (P = 0.16, n ≥ 9). (D, right) ONC significantly reduced dendritic intersections of αON-S RGCs compared with like cells from OHT eyes (P < 0.05). (E) αON-S RGCs from naïve, saline and sham retinas possess similar dendritic arbour morphologies based on branch points, dendritic length and field area (P ≥ 0.22, n ≥ 9). Total dendritic length is significantly reduced in αON-S RGCs following ONC compared with naïve (P < 0.05) and sham αON-S RGCs (P = 0.01, n ≥ 9). Dendritic field area is reduced by ONC in αON-S RGCs compared with naïve αON-S RGCs (P < 0.05, n ≥ 9). Statistics: (B, C and E) One-way ANOVA and Tukey’s post hoc. (D) Two-way repeated measures ANOVA with Bonferroni’s post hoc test. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data are presented as mean ± SEM.

Figure 7

Optic nerve crush piques mechanisms controlling dendritic branch points in αOFF-S RGCs. (A) Example confocal images of SMI-32 immunolabelled whole-mount retinas. Scale bars = 25 µm. (B) ONC reduced SMI-32 accumulation in αOFF-S RGCs relative to sham αOFF-S RGCs (P < 0.05, n ≥ 11). (C) ONC decreased αOFF-S RGC soma area compared with like cells from sham eyes (P = 0.01, n ≥ 11). (D, left) Sholl analysis indicated the number of dendritic intersection is similar for αOFF-S RGCs from naïve, saline and sham eyes (P = 0.96, n ≥ 6). (D, right) The number of dendritic intersections is also similar for αOFF-S RGCs from ONC and OHT eyes (P = 0.48, n ≥ 10). (E) There is no difference in the number of branch points, dendritic length or dendritic field area of saline and sham αOFF-S RGCs compared with naïve (P ≥ 0.99, n ≥ 8). ONC significantly reduced the number of branch points in αOFF-S RGCs compared with like cells from naïve (P = 0.05, n ≥ 8) and sham retinas (P = 0.003, n ≥ 11). Statistics: (B) One-way ANOVA and Tukey’s post hoc. (C and E) Kruskal–Wallis one-way ANOVA, Dunn’s post hoc. (D) Two-way repeated measures ANOVA with Bonferroni’s post hoc test. Significance indicators: *<0.05, **<0.01, ***<0.001, ****<0.0001. Data are presented as mean ± SEM.