Correction of amblyopia in cats and mice after the critical period

Abstract

Monocular deprivation early in development causes amblyopia, a severe visual impairment. Prognosis is poor if therapy is initiated after an early critical period. However, clinical observations have shown that recovery from amblyopia can occur later in life when the non-deprived (fellow) eye is removed. The traditional interpretation of this finding is that vision is improved simply by the elimination of interocular suppression in primary visual cortex, revealing responses to previously subthreshold input. However, an alternative explanation is that silencing activity in the fellow eye establishes conditions in visual cortex that enable the weak connections from the amblyopic eye to gain strength, in which case the recovery would persist even if vision is restored in the fellow eye. Consistent with this idea, we show here in cats and mice that temporary inactivation of the fellow eye is sufficient to promote a full and enduring recovery from amblyopia at ages when conventional treatments fail. Thus, connections serving the amblyopic eye are capable of substantial plasticity beyond the critical period, and this potential is unleashed by reversibly silencing the fellow eye.

Keywords: amblyopia; cat; metaplasticity; monocular deprivation; mouse; neuroscience; ocular dominance plasticity; tetrodotoxin.

Figure 1.. Temporary inactivation of one retina potentiates visual responses to stimulation of the non-inactivated eye in awake, late-adolescent mice.

(A) Schematic of mouse brain showing recording site in binocular V1 (V1b) and tetrodotoxin (TTX) injection site into contralateral eye. (B) Experimental timeline showing ocular manipulation and visual evoked potential (VEP) recording session times. (C) Longitudinal measurements of VEP magnitude for stimulation of the contralateral, inactivated eye. Filled bars denote mean peak-to-peak magnitude and open circles denote individual biological replicates. Average VEP waveforms are shown above for each time point. Asterisk denotes a statistically significant difference (p < 0.05) and ns denotes p > 0.05. Analyses were performed using one-way repeated measures ANOVA (F = 12.95, p = 0.0013). For comparisons of interest, Sidak’s post hoc tests were performed with correction for multiple comparisons (baseline vs. 1 hr, p < 0.0001; 1 hr vs. 7 days, p = 0.0008; baseline vs. 7 days, p > 0.9999). (D) Same as C for non-inactivated, ipsilateral eye. Analyses were performed using one-way repeated measures ANOVA (F = 9.361, p = 0.0092) followed by Sidak’s post hoc tests (baseline vs. 1 hr, p < 0.0001; 1 hr vs. 7 days, p < 0.0001; baseline vs. 7 days, p = 0.0121). (E) VEP magnitude over time normalized to baseline. Blue trace denotes inactivated contralateral eye; yellow trace denotes non-inactivated ipsilateral eye; open circles denote mean; error bars denote SEM; blue and yellow asterisks denote significant differences (p < 0.05) compared to hypothesized value of 100% for the contralateral and ipsilateral eyes, respectively (ns denotes p > 0.05). Analyses were performed using a one-sample t-test for the contralateral eye (compared to baseline: 1 hr, p < 0.0001; 1 day, p = 0.7475; 2 days, p = 0.2739; 2 days, p = 0.0976; 7 days, p = 0.7856) and a one-sample Wilcoxon test for the ipsilateral eye (compared to baseline: 1 hr, p = 0.0078; 1 day, p = 0. 0078; 2 days, p = 0.0156; 2 days, p = 0. 0156; 7 days, p = 0. 0156), with test identity selected based on outcome of Shapiro-Wilk normality test. All data in this figure is for phase-reversing sinusoidal grating stimulation at a spatial frequency of 0.2 cycles per degree (cpd).

Figure 1—figure supplement 1.. Histological verification of electrode position in mouse binocular primary visual cortex (V1).

Top, Nissl-stained coronal slice of mouse brain showing track left by an electrode implanted in layer 4 of V1. Green arrows denote the boundaries of binocular V1 (V1b). Bottom, cartoon of coronal slice from mouse brain corresponding to micrograph above. The monocular and binocular segments of V1 are shaded in blue and green, respectively.

Figure 2.. Fellow eye inactivation in mice promotes stable and complete recovery of vision in both eyes following long-term monocular deprivation (MD).

(A) Left, schematic of mouse brain showing recording site in V1b, as well as sites of deprivation and inactivation in the contralateral and ipsilateral eyes, respectively. Right, timeline showing experimental manipulations and recording session times. (B) Longitudinal measurements of visual evoked potential (VEP) magnitude for stimulation of the contralateral eye at 0.2 cycles per degree (cpd) for three littermate treatment groups: Sham, sham contralateral eyelid suture P26-47 and fellow eye saline at P47; MD, contralateral eyelid suture P26-47 and fellow eye saline at P47; MD then TTX, contralateral eyelid suture P26-47 and fellow eye TTX at P47. Filled bars denote mean peak-to-peak magnitude and open circles denote individual biological replicates. Asterisk denotes a significant difference in VEP magnitude (p < 0.05) compared to before treatment (0 weeks). Analyses were performed using a two-way repeated measures ANOVA (treatment × time, F_(8,156)=6.925, p < 0.0001) followed by Dunnett’s multiple comparisons tests (Sham, 0 vs. 1, 2, 3, 4 weeks: p = 0.4963, 0.9774, 0.9986, 0.9999; MD, 0 vs. 1, 2, 3, 4 weeks: p = 0.5051, 0.5160, 0.9756, 0.9640; MD then TTX, 0 vs. 1, 2, 3, 4 weeks: p = 0.0052, 0.0024, 0.0007, 0.0010). (C) Contralateral VEP magnitude over time for stimulation at 0.2 cpd, with Sham in black (dashed), MD in gray, and MD then TTX in blue. Open circles denote mean, error bars denote SEM; average waveforms shown above plot; gray and blue asterisks denote significant difference compared to Sham (p < 0.05) for the MD and MD then TTX groups, respectively, computed using Dunnett’s multiple comparisons tests (MD vs. Sham at 1, 2, 3, 4, 5 weeks: p = 0.0050, 0.0414, 0.0356, 0.0025, 0.0077; MD then TTX vs. Sham at 1, 2, 3, 4, 5 weeks: p = 0.0011, 0.9806, 0.6524, 0.5640, 0.8190). (D) Contralateral VEP magnitude across different spatial frequencies just after opening deprived eye but before treatment (0 weeks) and 4 weeks after opening deprived eye and inactivating fellow eye. Symbols/colors same as C. Post hoc comparisons performed using Dunnett’s multiple comparisons tests (0 weeks, Sham vs. MD at 0.05, 0.2, 0.4 cpd: p = 0.0054, 0.0050, 0.0002; 0 weeks, Sham vs. MD then TTX at 0.05, 0.2, 0.4 cpd: p = 0.0006, 0.0011, < 0.0001; 4 weeks, Sham vs. MD at 0.05, 0.2, 0.4 cpd: p = 0.0311, 0.0077, 0.0157; 4 weeks, Sham vs. MD then TTX at 0.05, 0.2, 0.4 cpd: p = 0.9319, 0.8190, 0.6413). (E–G) Same as B–D but for ipsilateral (fellow, inactivated) eye, with yellow denoting the MD then TTX condition in F–G. Analyses were performed using a two-way repeated measures ANOVA (treatment × time, F_(8,156)=1.464, p = 0.1745), with the absence of a significant interaction suggesting that there were no differences over time that could be attributed to treatment.

Figure 3.. Reverse occlusion (RO) improves deprived eye vision following long-term monocular deprivation (MD), but the recovery is not lasting.

(A) Schematic and timelines for experiment comparing fellow eye inactivation to RO following 3 weeks MD. (B–G) Same format as Figure 2B–G, but for two littermate treatment groups: MD then TTX, contralateral eyelid suture P26-47 and fellow eye TTX injection + sham suture at P47 + sham opening at P54; MD then RO, contralateral eyelid suture P26-47 and fellow eye saline + RO at P47-54. MD then TTX group is blue for C–D and yellow for F–G. MD then RO group is gray (dashed) for C–D and F–G. Except where otherwise noted, all data in this figure are for phase-reversing sinusoidal grating stimulation at a spatial frequency of 0.2 cycles per degree (cpd). Analyses in B and E were performed using two-way repeated measures ANOVA tests (contralateral eye, treatment × time, F_(4,108)=6.363, p = 0.0001; ipsilateral eye, treatment × time, F_(4,108)=1.474, p = 0.2152), with the significant interaction for the contralateral eye motivating Dunnett’s multiple comparisons tests (MD then TTX, 0 vs. 1, 2, 3, 4 weeks: p = 0.0057, 0.0065, 0.0437, 0.0042; MD then RO, 0 vs. 1, 2, 3, 4 weeks: p = 0.0035, 0.4119, 0.5738, 0.9831). Dunnett’s post hoc tests were also used for analyses in C (MD then TTX vs. MD then RO at 0 and 1 weeks: p > 0.9999, at 2, 3, 4 weeks: p = 0.5361, 0.5446, 0.0004) and D (0 weeks, MD then TTX vs. MD then RO at 0.05, 0.2, 0.4 cpd: p = 0.9917, >0.9999, 0.9998; 1 week, MD then TTX vs. MD then RO at 0.05, 0.2, 0.4 cpd: p = 0.9995, >0.9999, >0.9999; 4 weeks, MD then TTX vs. MD then RO at 0.05, 0.2, 0.4 cpd: p = 0.0041, 0.0004, 0.0255).

Figure 4.. Fellow eye inactivation promotes stable functional and structural recovery in cats following long-term monocular deprivation (MD).

(A) Recording and visual stimulation setup for measuring visually evoked responses non-invasively in cats. (B) Methodology for computing visually evoked responses from scalp surface field potential. Left, example raw field potential time series recorded from V1 during presentation of phase-reversing visual stimuli at a range of spatial frequencies. Gray lines denote timing of phase reversals. Right, data on left shown in the frequency domain following discrete Fourier transform (DFT). Blue arrows point to peaks in spectral power at 2 Hz (the phase reversal frequency) and six harmonics, representing visually evoked responses. Black arrows point to the frequencies used for control power measurements. (C) Total power at the visually driven (phase-reversing) frequency and its harmonics (blue) vs. control (black) across a range of spatial frequencies. Values computed from same recording shown in B. (D) Timeline showing experimental manipulations and recording session times. (E) Total power of visually evoked responses for one animal (C474, right primary visual cortex [V1]) viewing visual stimuli through the deprived left eye (blue) vs. the fellow right eye (yellow). Arrows denote time of tetrodotoxin (TTX) injections. Error bars, SEM. Spatial frequency, 0.1 cycles per degree (cpd). (F) Total power of visually driven responses for the same animal shown in E across a range of spatial frequencies before (top), during (middle), and after (bottom) inactivation of the fellow eye with TTX. Deprived left eye shown in blue, and fellow right eye shown in yellow. Black symbols denote control frequencies as in C. (G) Ocular dominance indices calculated for four cats before MD, after MD (but before inactivation), during inactivation, and after the fellow eye was no longer inactivated. Data are shown for both right (closed circles) and left (open circles) V1. Positive and negative values indicate response biases toward the left and right eyes, respectively, while the dashed line at 0 indicates balanced V1 responses expected from a neurotypical animal. As a reference, indices for the right hemisphere of C474 were calculated from data shown in E. Asterisks denote significant differences (p > 0.05) from the hypothesized value of 0, analyzed using Wilcoxon signed-rank tests (pre-MD, p = 0.9453; post-MD, p < 0.0001; fellow eye inactivated, p < 0.0001; post-TTX, p = 0.8236). (H) Left, low magnification image of the lateral geniculate nucleus (LGN) stained for Nissl substance after 3 weeks MD (top), or 3 weeks MD followed by fellow eye inactivation (bottom). Arrow indicates lamina A1, ipsilateral to the deprived eye. Middle and right, high magnification images from deprived (middle) and non-deprived (right) A1 layers for the same conditions shown on left. (I) Deprivation effect, stereological quantification of neuron soma size within deprived and non-deprived A and A1 layers, for normally reared cats, cats undergoing 3 weeks of MD, and cats undergoing 3 weeks of MD followed by fellow eye inactivation. Asterisk denotes significant difference (p > 0.05) from the hypothesized value of 0%, analyzed using a one-sample t-test (normal, p = 0.1917; MD, p = 0.0100; MD then TTX, p = 0.5716). (J) Average soma size for neurons in LGN layers downstream of the deprived left eye (L) or the fellow right eye (R) for normally reared animals compared to those undergoing 3 weeks of MD followed by fellow eye inactivation. There was no significant difference between the groups (Welch’s ANOVA, W = 1.426, p = 0.2881). Open and closed symbols indicate values drawn from laminas A1 and A, respectively. Black lines indicate mean and SEM for each group.

Figure 4—figure supplement 1.. Time-domain visual evoked potentials (VEPs) from cat scalp surface field potential.

(A) Bottom, example phase reversal-aligned EEG for 120 trials within a recording session where the 0.1 cycles per degree (cpd) grating stimulus was presented (interleaved with blocks of grating stimuli at other spatial frequencies). The EEG is normalized to a 50 ms window prior to each phase reversal at time = 0. Top, VEP computed as the average across all trials shown below. (B) Top, VEP waveforms for grating stimuli presented at (from left to right): 0.05, 0.1, 0.5, 1, 2 cpd, and gray screen. Bottom, peak VEP magnitude during the 300 ms window following the phase reversal, quantified as a function of spatial frequency. Data for 0.1 cpd is the same as shown in A, and data used in this plot is the same as example of frequency-domain analysis shown in Figure 4C.

Figure 4—figure supplement 2.. Trajectory of deprivation-driven ocular dominance shift and inactivation-mediated recovery in individual cats.

Ocular dominance indices (ODIs) over time for four cats subjected to long-term monocular deprivation (MD) followed by fellow eye inactivation. ODIs are computed from scalp surface potential recordings during monocular viewing of grating stimuli at 0.1 cycles per degree (cpd). Arrows denote time of tetrodotoxin (TTX) injections. For each plot, the dashed line at 0 indicates a balanced ODI typical of a visually normal animal, whereas values of 1 and –1 indicate complete dominance by either the left or right eye, respectively. Error bars, SEM.