MeCP2 regulates the timing of critical period plasticity that shapes functional connectivity in primary visual cortex

Significance

Rett syndrome is a neurodevelopmental disorder caused by mutations in methyl-CpG-binding protein 2 (MeCP2). It is thought to result from altered neuronal connectivity and/or plasticity, possibly through abnormal experience-dependent synapse development, but the underlying mechanisms remain obscure. Using MeCP2-null mice, we examined experience-dependent development of neural circuits in the primary visual cortex where GABAergic interneurons regulate a critical period of neural plasticity. We provide evidence that a precocious maturation of parvalbumin⁺ GABAergic interneurons correlates with the altered timing of the critical period and deficient visual function in the absence of MECP2. Our study begins to establish a link from specific molecular changes in GABAergic neurons to critical period of circuit development and to functional alterations in a mouse model of Rett syndrome.

Abstract

Mutations in methyl-CpG-binding protein 2 (MeCP2) cause Rett syndrome, an autism spectrum-associated disorder with a host of neurological and sensory symptoms, but the pathogenic mechanisms remain elusive. Neuronal circuits are shaped by experience during critical periods of heightened plasticity. The maturation of cortical GABA inhibitory circuitry, the parvalbumin⁺ (PV⁺) fast-spiking interneurons in particular, is a key component that regulates the initiation and termination of the critical period. Using MeCP2-null mice, we examined experience-dependent development of neural circuits in the primary visual cortex. The functional maturation of parvalbumin interneurons was accelerated upon vision onset, as indicated by elevated GABA synthetic enzymes, vesicular GABA transporter, perineuronal nets, and enhanced GABA transmission among PV interneurons. These changes correlated with a precocious onset and closure of critical period and deficient binocular visual function in mature animals. Reduction of GAD67 expression rescued the precocious opening of the critical period, suggesting its major role in MECP2-mediated regulation of experience-driven circuit development. Our results identify molecular changes in a defined cortical cell type and link aberrant developmental trajectory to functional deficits in a model of neuropsychiatric disorder.

MECP2 is a methyl-CpG binding protein implicated in the regulation of neuronal chromatin architecture and gene transcription (1–4), especially in response to neural activity and experience during postnatal life (5, 6). Mutations in the MeCP2 gene cause Rett syndrome (RTT) (7), an autism spectrum disorder. Although MECP2 is broadly expressed in the developing and mature brain (8, 9), patients with RTT achieve expected early postnatal developmental milestones, but later experience a progressive loss of acquired linguistic, social, and motor skills; patients are further characterized by a selected set of postnatal symptoms including stereotyped behaviors and a myriad of sensory and neurological phenotypes (10). MeCP2 mutations likely impact gene regulation in multiple neuronal (11) and glial cell types (12, 13) and in multiple brain regions; the initial impact may further result in compensatory changes, including maladaptive synaptic plasticity (8, 14), leading to altered developmental trajectory of circuit formation and function. Although numerous mouse models of genetic etiology have been established (15), it has been difficult to identify primary molecular and cellular changes and trace their impact on circuit alterations that underlie behavioral deficits.

Brain circuits are shaped by experience during restricted time windows of early postnatal life. These “critical periods” are crucial for experience-dependent formation of synaptic connections and wiring of functionally related neuronal pathways that underlie sensory, motor, cognitive, social, and language abilities (16). The postnatal maturation of GABA-mediated inhibition and cortical interneurons is crucial for controlling the timing of critical period plasticity in rodent primary visual cortex (17, 18). There is increasing evidence that dysfunction of GABA signaling is associated with autism disorders and RTT (19–21), raising the possibility that deficits in critical period plasticity may underlie some of the symptoms in these disorders. However, it remains unclear how MeCP2 mutations impact gene expression in any defined cells and circuits that alter subsequent developmental trajectories, leading to specific functional deficits.

Using a MeCP2-null mouse model (22), here we examined experience-dependent development of neural circuits in the primary visual cortex (V1) where the maturation of a defined class of GABA interneurons regulates a critical period of plasticity that shapes visual perception. In V1 of MeCP2^−/y mice, vision onset triggered a significant up-regulation of several key GABAergic markers, including GABA synthetic enzymes GAD67 and GAD65, vesicular GABA transporter, parvalbumin (PV), and components of perineuronal nets (PNNs); several of these are associated with the PV class GABA interneurons. These changes are correlated with an early enhancement of GABA transmission within the PV interneuron network and an accelerated maturation of cortical activity propagation patterns. We further demonstrate an early onset and closure of the critical period and deficient visual function: binocular neurons in mature V1 remain mismatched with regard to inputs from the two eyes. Importantly, a reduction of GAD67 level is sufficient to rescue the precocious onset of the critical period, suggesting a significant role for GAD67 in MECP2-mediated regulation of experience-driven V1 circuit development. Together, our studies reveal the impact of MeCP2 mutation on specific gene expression in identified cell types and developmental stages and trace altered neural circuit wiring to functional deficits.

Results

Accelerated Maturation of Parvalbumin Interneuron Network in V1 of MeCP2^−/y Mice.

Dani et al. (23) suggested a shift in balance of excitation and inhibition (E/I) in favor of inhibition in mature cortex of MeCP2^−/y mice, although only subtle changes in GABAergic marker expression were detected in cortical homogenates (21). To determine the impact of MeCP2 mutation on cortical GABAergic neurons during early postnatal development with cellular and laminar resolution, we examined a panel of GABA markers that directly impact inhibitory transmission. Shortly after eye opening (postnatal day 15, P15), we detected a significant increase in GAD67, the activity-regulated and rate-limiting GABA synthetic enzyme, and the calcium-binding protein parvalbumin, especially in the infragranular layers of MeCP2^−/y V1 (Fig. 1 A–F). This increased GAD67 expression was more prominent in PV-expressing cells, most of which mature into fast-spiking basket interneurons (24). There was no change in the number of PV interneurons as quantified using a Rosa26-loxpSTOPloxp-H2BGFP reporter activated by a PV-T2A-Cre allele that labels PV⁺ nuclei (Fig. S1 A–C), but the level of PV immunofluorescence was increased in a subset of these neurons. These results indicate that more GABAergic neurons express higher levels of GAD67 and PV protein in MeCP2^−/y V1, compared with WT V1. Increased GAD67 was not only detected in the somata of PV interneurons but also in their presynaptic terminals (or puncta), forming a ring-like structure (Fig. 1 D–F) (18). In addition, immunofluorescence puncta for GAD65 (Fig. 1 G–I), the other GABA synthetic enzyme, specifically localized to presynaptic terminals, and vesicular GABA transporter (VGAT) were also increased (Fig. 1 J–L). Together, these results suggest increased presynaptic vesicles and/or GABA loading, especially in PV interneurons, in MeCP2^−/y V1.

Fig. 1.

Increased expression of key components of GABA transmission in MeCP2^−/y V1 at P15. (A–C) GAD67 (green) and PV (red) expression in the soma of GABAergic neurons in infragranular layers was higher in MeCP2^−/y compared with WT. Immunofluorescence was carried out without Triton to preserve GAD67 in cell soma. (Scale bar, 50 μm.) Arrowheads point to soma with high GAD67 and PV expression quantified in C (number of high-intense soma/0.6 mm²). Somata with higher level expression of GAD67 or PV were identified by a threshold method in both genotypes (Materials and Methods). (D–L) Increased synaptic localization of GAD67 (D–F), GAD65 (G–I), and VGAT (J–L) in MeCP2^−/y compared with that in WT. Asterisk indicates somata; arrowheads indicate perisomatic puncta. Note the much more prominent puncta-ring structures surrounding soma in the MeCP2^−/y panel. Immunofluorescence was carried out with Triton to better reveal synaptic localization of these proteins. (Scale bar, 10 μm.) Cumulative distribution graphs show higher percentage of brighter GAD67 (F), GAD65 (I), and VGAT (L) puncta in MeCP2^−/y V1 (red) compared with WT (black).

Fig. S1.

MeCP2-null line was crossed with a reporter line expressing nuclear GFP (Rosa26-loxpSTOPloxp-H2BGFP) reporter allele activated by a PV-T2A-Cre allele that selectively labels PV⁺ nuclei. GFP was expressed in PV⁺ neurons as shown by double labeling with anti-PV (A, Inset, red) in WT and MeCP2^−/y (A and B) cortices. (C) No difference in the number of PV⁺ cells was detected between the two genotypes at P12 and P15 in the visual cortex (n = 4–6 sections).

During the critical period of visual cortical development (25), which spans around 3 wk after eye opening, GABA⁺ PV interneurons undergo substantial maturation in synaptic innervation and physiological properties, concomitant with increased expression and synaptic localization of GAD67, GAD65, and PV (18, 24, 26). In V1 of MeCP2^−/y mice, we found that increased expression of GAD67, GAD65, vGAT, and PV persisted to at least the fifth week (Fig. 2 A–I). Interestingly, there was an increase in MECP2 expression in V1 during this period (Fig. S2). Together these results indicate that, upon the onset of visual experience, MeCP2 deficiency results in a precocious and persistent increase in synaptic localization of molecular machineries for GABA synthesis (GAD67 and GAD65), release (VGAT), and release regulation (PV) in V1, especially in PV interneurons.

Fig. 2.

Increased expression of key components of GABA transmission in MeCP2^−/y V1 at P30. (A–C) GAD67 (green) and PV (red) expression in the soma of GABAergic neurons in infragranular layers was higher in MeCP2^−/y compared with WT. (Scale bar, 50 μm.) Arrowheads point to soma with high GAD67 and PV expression quantified in C (number of high-intense soma/0.6 mm²). (D–L) Increased synaptic localization of GAD67 (D–F) and VGAT (G–I) in MeCP2^−/y compared with that in WT. Arrowheads indicate perisomatic puncta. (Scale bar, 10 μm.) Cumulative distribution graphs show higher percentage of brighter GAD67 (F) and VGAT (I) in MeCP2^−/y V1 (red) compared with WT (black).

Fig. S2.

Increased MECP2 expression during critical period of plasticity in WT V1. MECP2 expression was assessed by confocal imaging at one setting (P30) for comparison during critical period of plasticity. MECP2 expression is very low at P12 with increasing expression at P15 (A) and reaches adult levels after P30 (B). (C) Histogram shows higher frequency of brighter MECP2 at P30, compared with P15 (red) and P12 (black) from representative V1 sections, suggesting a time course for MECP2 function with the maturation of the visual cortex.

Another highly characteristic feature of V1 maturation is the progressive increase of PNNs following eye opening (27). PNNs consist of specialized extracellular matrix proteins such as chondroitin sulfate proteoglycans and interdigitate with synaptic contacts (28). The developmental increase of PNNs is regulated by visual experience (29), and PNNs in mature V1 mainly surround the soma and proximal dendrites of PV interneurons (Fig. 3) (30). Mature PNN is inhibitory for experience-dependent plasticity: its increase in developing V1 correlates with the termination of the critical period and its removal in adult V1 restores plasticity (27, 31, 32). Shortly after eye opening (P15), PNNs in WT V1 were still highly immature: wisteria floribunda agglutinin (WFA), which binds the glycosaminoglycan chains of chondroitin sulfate proteoglycans signals, was weak and diffuse, and rather few neurons were surrounded by PNNs (Fig. 3 A and B). In P15 MeCP2^−/y V1, however, WFA signals were already distinct, and PV cells were surrounded by PNNs especially in infragranular layers (Fig. 3 C and D) (numbers of PNN/0.6 mm² in WT were 13.0 ± 3.2 and in MeCP2^−/y: 31.1 ± 4.1, P value = 0.0009). At P30, PNNs in WT V1 began to take on their characteristic appearance, but in relatively small numbers (Fig. 3 E and F), in line with the progression of a normal critical period; in MeCP2^−/y littermates, however, there were significantly more well-formed PNNs in V1 (Fig. 3 G and H) (numbers of PNN/0.6 mm² in WT: 8.8 ± 1.5 and in MeCP2^−/y: 22 ± 3.6, P value = 0.001. Because of significant differences in level and pattern of PNNs at P15 and P30, different threshold criteria were used for defining and quantifying PNN; thus PNN quantifications are not comparable across ages). Taken all together, these results suggest a precocious maturation of GABAergic interneurons, especially of PV⁺ neurons, after vision onset in MeCP2^−/y V1.

Fig. 3.

Precocious maturation of perineuronal nets (PNNs) in MeCP2^−/y mice. PNNs were identified by lectin WFA (wisteria floribunda agglutinin conjugated to biotin) staining, which labels extracellular matrix components. By P15, weak and diffuse labeling of WFA (green) in WT (A and B) compared with well-formed PNNs, especially around PV⁺ neurons (red) in infragranular layers [layers 5 (L5) and 6 (L6) marked at Top] of MeCP2^−/y V1 (C and D). By P30, WFA staining begins to delineate more well-formed structures in WT V1 (E and F), whereas PNNs in MeCP2^−/y V1 are already well-formed and numerous (G and H). Arrows point to mature nets that encapsulate PV⁺ neurons. Lower panels are higher magnification images of Top Insets. (Scale bars, 100 μm for A, C, E, and G and 50 μm for B, D, F, and H.)

Enhancement of GABA Transmission Within the PV Interneuron Network in MeCP2^−/y V1.

In addition to providing perisomatic inhibition to pyramidal neurons (PyNs), PV cells form extensive connections among themselves through both electrical coupling and reciprocal GABAergic synapses (33, 34). This highly connected network of fast-spiking PV cells plays a key role in the detection of temporal activity patterns within the neocortex (35) and in the generation of network oscillations (34, 36). During postnatal development, substantial maturation of intrinsic and synaptic properties, including GABA transmission among PV cells, converts the PV cell network from a slow to fast signaling device for regulating cortical activity patterns (26, 37). To determine the physiological impact of increased GADs and vGAT levels in PV cells in MeCP2^−/y V1, we measured a series of physiological parameters. The intrinsic properties of layer V PV neurons were very similar between WT and MeCP2^−/y V1 (Fig. S3). We then measured miniature inhibitory postsynaptic current (mIPSC) in both PV and PyNs in layer V of V1 at P15. Whereas there was no significant change in mIPSC frequency in either cell population (Fig. S4 C and F), we found a significant increase in mIPSC amplitude in PV cells (Fig. S4 A and B), consistent with increased GAD67 and GAD65 expression in MeCP2^−/y V1 interneurons. Interestingly, there was no increase in mIPSC amplitude in PyNs (Fig. S4 D and E).

Fig. S3.

Intrinsic properties of layer V PV neurons are similar between WT and MeCP2^−/y at P15. Features such as (A) axial resistance (Ra), (B) membrane resistance (Rm), (D) action potential (AP) threshold, (E) AP amplitude, (F) after hyperpolarization, and (G) AP half-width are not significantly different in PV neurons between the genotypes. Membrane potential (Vm, C) was significantly different, although the mean changes in the values were minimal. Data presented are from 23 WT and 25 MeCP2^−/y PV neurons.

Fig. S4.

Cell-type–specific increases in miniature inhibitory synaptic transmission in MeCP2^−/y V1 at P15. (A–C) Increase in the amplitude and no change in the frequency of miniature IPSCs in PV interneurons in MeCP2^−/y V1. (A) Examples of miniature IPSC events in PV interneurons in WT (black) and MeCP2^−/y V1 (red). Cumulative probability histogram and averages show significant increase in the amplitude of miniature IPSCs (P = 0.001) (B) and no significant change in frequency (C) between WT and MeCP2^−/y PV cells. (D) Examples of miniature IPSC events in PyNs in WT (black) and MeCP2^−/y V1 (red). Cumulative probability histogram and averages show no significant change in the amplitude (E) and frequency (F) of miniature IPSCs between WT and MeCP2^−/y pyramidal cells at P15. Data representation is mean ± SEM with Student’s t test.

As fast-spiking interneurons, PV cells communicate among each other and transmit inhibitory output to pyramidal neurons typically by firing bursts of action potentials (APs) (34). We therefore performed paired recordings between PV neurons (PV–PV) and PV to PyNs (PV–PyNs) in layer V at P15. Importantly, when we stimulated PV neurons with a train of action potentials at 20 Hz, we found a significant reduction in synaptic depression in MeCP2^−/y V1 compared with that in WT V1 (Fig. 4 A–C). This reduction likely results from an enhancement of GABA transmission between PV–PV cells or a change in release probability. This increase in inhibitory transmission dynamics is likely physiologically relevant as PV cells typically fire in the frequency range of 10–40 Hz in vivo (34). On the other hand, there was no such change in inhibitory transmission from PV to PyNs (Fig. 4 D–F). An enhancement of GABA transmission among PV cells at P15 would suggest an earlier physiological maturation of these cells which, as a more coherent network, would likely provide more robust inhibitory control over the spatiotemporal flow and processing of neural activity in V1.

Fig. 4.

Enhancement of GABA transmission among PV cells in MeCP2^−/y V1 at P15. (A, Top) A schematic of whole-cell patch recording between two PV neurons in layer V of wild-type V1 (black). (Bottom) Example IPSCs in a PV neuron (Upper trace) evoked by a train of APs (20 Hz) in a presynaptic PV neuron (Lower trace), averaged from 20 trials. (B) Same recording configuration as in A but in PV neurons of MeCP2^−/y V1 (red). (C) Analysis of the short-term dynamics of GABA transmission among PV neurons showed significant reduction in synaptic depression in MeCP2^−/y V1 (red), compared with WT (black) (t test, **P < 0.01, and **P < 0.001). (D–F) Similar analysis of GABA transmission from PV neurons to PyNs in layer V did not show significant changes in short-term synaptic dynamics between WT and MeCP2^−/y.

Altered Development of Spatiotemporal Activity Propagation in MeCP2^−/y V1.

From eye opening to the peak of the critical period (fourth postnatal week), there is significant maturation in the spatiotemporal propagation of intracortical activity within rodent V1 (38). Using voltage sensitive dye imaging (VSD), we first confirmed in WT mice, as previously shown in rats, that there is a significant reduction in cortical activation across all layers between P15 and P30 after stimulating the white matter (WM) (Fig. 5) and layer IV (Fig. S5). At P30, the amplitude of peak activation of V1 was significantly reduced, compared with P15, as measured by the time course of VSD signal intensity (Fig. 5 G–J, Left) and by the input/output relationship of the VSD signal at increasing stimulation intensities (Fig. 5 G–J, Right). In MeCP2^−/y V1 at P15, following white matter stimulation, there was a significant reduction in the spatial spread of voltage signaling across all layers compared with those of WT littermates. The observed differences in the peak of the voltage signals was likely due to the reduced amplitude of peak activation in MeCP2^−/y V1. The overall patterns of vertical and horizontal propagation of voltage signal were not significantly different between genotypes (Figs. S6 and S7). These data indicate a significant decrease in V1 activation, but no alteration of the direction of propagation over time in MeCP2^−/y V1. The spatiotemporal profile of signal propagation in P15 MeCP2^−/y mice appeared more similar to that of WT V1 at P30, suggesting an accelerated maturation of the cortical circuit operation. The enhancement of GABA transmission dynamics (39), within PV cell network in particular, in P15 MeCP2^−/y V1 likely contributes to the earlier spatial restriction in activity propagation.

Fig. 5.

Accelerated maturation of neural activity propagation through cortical layers after white matter stimulation in MeCP2^−/y V1. Decreased activation of signal through cortical layers in MeCP2^−/y V1 after white matter stimulation, shown with representative sample VSD images for WT (A–C) and MeCP2^−/y V1(D–F) at 0, 5, and 10 ms. The maximum transition in voltage signal is 1%∆F/F (deep red color); the minimum is −1%∆F/F (dark blue). The gray area in the scale bar indicates 0.02%∆F/F, the detection threshold. White boxes: regions of interest quantified in G–J, Left. (G–J, Left) Time course of activation in WT L2/3 (G), L4 (H), L5 (I) and L6 (J) show significant decrease in amplitude at P30 (black), compared with P15 (76). Voltage signals were measured as the transition in fluorescence from the baseline (∆F/F). The time course of activation in MeCP2^−/y at P15 (green) across all layers is similar to WT-P30 and significantly reduced compared with WT-P15. (Right) Input/output relationship of the VSD signal at increasing stimulation intensities across genotypes and time points suggests precocious maturation in the patterns of cortical activation in MeCP2^−/y. Asterisks indicate statistically significant differences.

Fig. S5.

Accelerated maturation of neural activity propagation through cortical layers after layer 4 stimulation in MeCP2^−/y V1. Decreased activation of signal through cortical layers in MeCP2^−/y V1 after layer 4 stimulation, shown with representative sample VSD images for WT (A–C) and MeCP2^−/y V1 (D–F) at 0, 5, and 10 ms from stimulus delivery. The maximum transition in voltage signal is 1%∆F/F (deep red color); the minimum is −1%∆F/F (dark blue). The gray area in the scale bar indicates 0.02%∆F/F, the detection threshold. White boxes show regions of interest quantified in G–J, Left. (G–J, Left) Time course of activation in WT L2/3 (G), L4 (H), L5 (I), and L6 (J) show significant decrease in amplitude at P30 (black), compared with P15. Voltage signals were measured as the transition in fluorescence from the baseline (∆F/F). The time course of activation in MeCP2^−/y at P15 (green) across all layers is similar to WT-P30 and significantly reduced compared with WT-P15. (Right) Input/output relationship of the VSD signal at increasing stimulation intensities is presented across genotypes and time points. These data suggest a precocious maturation of patterns of cortical activation in MeCP2^−/y. Asterisks indicate statistically significant differences.

Fig. S6.

Interlaminar and intralaminar spatiotemporal patterns of circuit activation elicited by white matter (WM) stimulation. (A) Interlaminar spatiotemporal patterns of circuit activation. (Top) Time course of the percent of ΔF/F measured in a region of interest (ROI) spanning from the pia to the WM. At 2.5 ms following stimulation of the WM, L6 showed the largest level of activation. At 5 ms, the optical signal had spread into L5 and L4. L2/3 became maximally activated 10 ms after the stimulus. Analysis of the temporal sequence of activation showed that the overall vertical pattern of signal propagation was similar in WT and MeCP2^−/y mice. Arrows indicate the peak intensity plotted in the bar plot below. (Bottom) Peak values of %ΔF/F at different time points following WM stimulation. Significant difference in the peak of the voltage signals between age groups was observed. Asterisks indicate significant differences. *P < 0.05; **P < 0.01. (B) Intralaminar spatiotemporal patterns of circuit activation shown as %ΔF/F measured in ROIs within L2/3 and L4 at 10 ms after WM stimulation. (Inset) Average Gaussian half-width measured in L2/3 and L4 in both age groups. The horizontal propagation of cortical activation within each layer was also quantified by measuring the half-width of the Gaussian fitting the profile of the change in fluorescence for each pixel along a line parallel to the pia. Following WM stimulation, the half max width (HMW) of the %ΔF/F signal were not significantly different between genotypes or time points within all layers of slices.

Fig. S7.

Interlaminar and intralaminar spatiotemporal patterns of circuit activation elicited by stimulation of L4. (A) Interlaminar spatiotemporal patterns of circuit activation. (Top) Time course of %ΔF/F measured in an ROI spanning the cortical mantle from the pia to the WM. Arrows indicate the peak intensity plotted in the bar graph below. (Bottom) Peak values of %ΔF/F at different time points following L4 stimulation. Although the overall pattern of signal propagation was similar between WT and MeCP2^−/y V1, the reduction of signal intensity in MeCP2^−/y V1 was more profound with L4 stimulation. The peak activation of P15 MeCP2^−/y V1 was significantly reduced starting at 5 ms after the delivery of the stimulus, suggesting a decreased efficiency in the propagation of the activity to the superficial layers. Asterisks indicate significant differences. *P < 0.05; **P < 0.01. (B) Intralaminar spatiotemporal patterns of circuit activation shown as distribution of %ΔF/F measured in ROIs within L2/3 and L4 10 ms following L4 stimulation. (Inset) Averaged Gaussian half-width measured in L2/3 and L4 for all age groups. Although the pattern of activation was similar between WT and MeCP2^−/y V1, L4 stimulation unveiled a significant reduction in intralaminar spread of cortical activity in MeCP2^−/y V1. Asterisks indicate significant differences *P < 0.05; **P < 0.01.

A Precocious Critical Period of Plasticity in V1 of MeCP2^−/y Mice.

Following eye opening, visual experience stimulates the maturation of receptive field properties and guides the functional wiring of V1 circuits during a critical period (25). Importantly, this critical period does not start immediately at vision onset but is initiated more than a week later in rodents (40). Maturation of GABAergic inhibition during this “precritical period” in the third postnatal week is necessary to initiate the critical period, which peaks in the fourth week and closes shortly after (16). A precocious increase of GABA transmission in V1 and precocious maturation of PV cells trigger an earlier onset (18, 41), whereas an accelerated rise of PNNs results in an earlier closure of the critical period (27). We compared the timing of ocular dominance (OD) plasticity in WT and MeCP2^−/y mice using a brief monocular deprivation (MD) protocol. As expected, MD in WT mice from P26 induced a significant OD shift (Fig. 6 A, B, and E, Left), whereas MD from P15 to P20 had no effect (Fig. 6 A and D, Left). However, in MeCP2^−/y mice, MD from P15 to P20 induced a significant OD shift (Fig. 6 A and D, Right) but had no effect at the normal peak of the critical period (i.e., from P26; Fig. 6 C and E, Right). Taken together, both the onset and closure of the critical period for OD plasticity were accelerated in MeCP2^−/y mice. Furthermore, in ∼P30 MeCP2^−/y mice, the evoked (Fig. S8A) and spontaneous (Fig. S8B) firing rates in vivo were not significantly different from WT controls, suggesting that the accelerated maturation of PV interneurons did not grossly alter activity levels in MeCP2^−/y V1.

Fig. 6.

Precocious critical period in MeCP2^−/y mice. (A) Schema representing the experimental design for the ocular dominance (OD) plasticity experiments and precocious critical period in MeCP2^−/y V1. (B and C) Representative cortical response magnitude to visual stimuli through the contralateral (Contra) and ipsilateral eye (Ipsi) in control and monocularly deprived WT (B) and MeCP2^−/y (C) V1 during the normal critical period (P26–P32). (D and E) Quantification of ocular dominance index (ODI). ODI score was calculated as (Contra − Ipsi)/(Contra + Ipsi) for every responsive pixel. (D) In WT and MeCP2^−/y V1, MD was performed from P15 until intrinsic imaging of OD at P20/21. Whereas WT V1 did not show OD plasticity at this time point (Left, white circles), MeCP2^−/y V1 displayed significant ODI shift (Right, red circles). (E) When MD was performed from P26 to P31–P33, WT V1 showed OD plasticity (Left), whereas MeCP2^−/y V1 did not (77).

Fig. S8.

No change in firing rates of neurons between WT and MeCP2^−/y V1. Both evoked (A) and spontaneous (B) neuronal activity were not significantly different at P30 between WT (black) and MeCP2^−/y (red) for contralateral (evoked: P = 0.13, spontaneous: P = 0.12) and ipsilateral (evoked: P = 0.86, spontaneous: P = 0.93) eye stimulation.

Binocular Matching of Orientation Preference Is Disrupted in V1 of MeCP2^−/y Mice.

A major function of critical period plasticity is to promote the experience-dependent matching of orientation tuning through the two eyes (42). Shortly after eye opening, V1 neurons exhibit orientation tuning and respond to stimulus to both eyes, but the orientation preference is binocularly mismatched. This initial mismatch is then progressively corrected through visual experience during the critical period, such that individual binocular neurons become tuned to similar orientations through the two eyes (42), a feature presumably important for normal binocular vision. In MeCP2^−/y mice, monocular orientation tuning profiles of V1 neurons were similar to those in WT mice (Fig. 7 A–F), indicating that the basic receptive field properties of V1 neurons are normal in the absence of MeCP2. However, the binocular matching of orientation preference failed to occur in MeCP2^−/y mice (Fig. 7 A–D and G–L). The degree of matching in ∼P30 MeCP2^−/y mice was significantly worse than age-matched WTs (which had reached the adult level by P30, Fig. 7L) and in fact remained at a level preceding the critical period (∼P20 WT mice) and similar to that observed after long-term visual deprivation from birth (42). Therefore, the precocious critical period in MeCP2^−/y mice, which reduces the influence of visual experience on the developing V1, is correlated with and possibly contributes to aberrant wiring of V1 circuits and deficient visual perception.

Fig. 7.

Deficient binocular matching of orientation tuning in MeCP2^−/y V1. (A–D) Polar plots of orientation tuning curves of individual binocular neurons in V1. (A and B) An example WT neuron showing orientation selective responses through the contralateral (Contra, A) and ipsilateral (Ipsi, B) eye, and matched orientation tuning between the two eyes. (C and D) An example MeCP2^−/y neuron showing sharp orientation tuning monocularly, but mismatch between the two eyes. (E and F) The mean monocular orientation selectivity index (OSI) (E) and orientation selectivity tuning width (OS width) (F) were not different between the two genotypes. (G and H) Correlation of contralateral and ipsilateral orientation preference in WT littermate controls (G: 54 cells, 7 mice, P32–P37) and MeCP2^−/y mice (H: 52 cells, 12 mice, P32–P37). Higher percentage of binocular cells in the WT V1 (I) have similar orientation preferences between the two eyes (ΔO), compared with MeCP2^−/y (J). (K and L) Cumulative distribution (K) and the mean (L) of the absolute values of ΔO in WT (black line) and MeCP2^−/y (red line).

Rescue of Critical Period Onset by GAD67 Reduction in MeCP2^−/y Mice.

Among the several molecular alterations in V1 of MeCP2^−/y mice, GAD67 likely has the most significant effect on cellular/vesicular GABA contents and directly impacts the strength of GABA transmission (43). To determine whether the precocious rise in GAD67 expression after eye opening in MeCP2^−/y mice is a key event in triggering an earlier onset of the critical period, we reduced GAD67 expression by crossing a germ-line gad1^+/− heterozygous allele into the MeCP2^−/y background. GAD67 level in P15 MeCP2^−/y; Gad^+/− V1 was reduced to a level comparable to those in WT V1 (Fig. 8G). Importantly, reduction of GAD67 expression in MeCP2^−/y mice restored PNNs and PV expression to levels comparable to those in WT V1 (Fig. 8 A–G). Furthermore, MD from P15 to P19 induced no OD plasticity in MeCP2^−/y; Gad^+/− V1, similar to WT mice and in sharp contrast to MeCP2^−/y (Fig. 8H). These results indicate that the accelerated rise of GAD67 expression after vision onset is a key mechanism through which MeCP2 deficiency leads to altered gene expression and a precocious critical period.

Fig. 8.

GAD67 down-regulation rescues precocious onset of critical period in MeCP2^−/y V1. Effects of GAD67 down-regulation on PNN and PV expression in MeCP2^−/y V1. In V1 of compound mutants MeCP2^−/y; Gad1^+/− at P15 (D–F), down-regulation of GAD67 reduces the expression of PNN and PV in deeper layers 5 and 6 (L5 and L6) compared with those in MeCP2^−/y V1 (A–C). Arrowheads point to cells with high-PV (B and E) and PNN (A and D) expression that were quantified in G. (G) Quantification of the impact of Gad1 germ-line heterozygote on GAD67, PNN, PV expression in V1 (quantified as high-intensity immunofluorescence soma/0.6 mm²) in WT or MeCP2^−/y background. Gad1^+/− reduces GAD67 levels in MeCP2^−/y mice and restores PNN and PV expression in MeCP2^−/y mice to levels similar to those in WT mice. Asterisks denote statistically significant changes, whereas “ns” denotes no significant change. (H) Quantification of ocular dominance index (ODI) in MeCP2^−/y and MeCP2^−/y; Gad^+/− V1 following monocular deprivation from P15 to P20/21. Whereas MeCP2^−/y V1 shows precocious OD plasticity (Left, black/red circles), MeCP2^−/y; Gad^+/− V1 no longer exhibits plasticity at this early time point (Right, black/red squares).

Discussion

Although autism spectrum disorders are thought to primarily affect social functioning, there is increasing evidence that unusual sensory processing is at least a concomitant and possibly a cause of many of the behavioral symptoms (44, 45). Here, we use a well-established model of experience-dependent plasticity to discover the impact of MECP2 deficiency in identified cell types and developmental stages, and trace altered neural circuit plasticity to functional deficits.

It is now well established, largely from studies in rodent primary visual cortex, that the maturation of GABAergic inhibition mediated by a fast-spiking PV interneuron network is a crucial mechanism for triggering the onset and regulating the progression of the critical period for functional wiring in V1 (16, 46). Upon eye opening at ∼P13, visual experience stimulates substantial and progressive activity-dependent maturation of PV cell morphology, connectivity, and physiological properties (18, 24, 43, 47). Visual acuity of MeCP2^−/y mice is comparable to that of WT mice at the time of eye opening, as measured by an optomotor task (21). However, 2–3 d after vision onset (by P15), we found substantial molecular changes (e.g., in GAD67, GAD65, vGAT, PV, and PNN), especially in PV cells, which are indicative of physiological changes. As a rate-limiting GABA synthetic enzyme, GAD67 expression is highly sensitive to input and network activity and directly impacts the magnitude GABAergic transmission (43, 48). Because GAD67 is expressed in all cortical GABAergic neurons and yet appears to manifest cell-type–specific regulation (49), it is crucial to examine activity regulation of GAD67 transcription with laminar and cell-type resolution. A recent study using qRT-PCR on cortical homogenates did not detect changes in GAD67 in adult MeCP2^−/y V1 (21), likely due to the dilution effect by other non-PV interneurons. In our immunohistochemistry study, improved somatic GAD67 staining was achieved by omitting Triton in the protocol to better retain soluble GAD67 in the soma (50). It should be noted that, in addition to mediating cytosolic GABA synthesis, which influences cellular GABA levels, GAD67 and GAD65 form complexes that directly interact with vGAT on synaptic vesicles, thereby facilitating the coupling from GABA synthesis to vesicular packaging (51, 52). The key role of GAD67, GAD65, PV, and vGAT in regulating vesicular GABA content and release (26, 51–53) suggest that alterations in their levels may impact the strength and/or dynamics of GABA transmission (54), especially during fast spiking in PV cells.

In addition to the progressive innervation and inhibitory control of pyramidal neurons, a major aspect of PV cell maturation is the formation of extensive connections among PV cells themselves through both electrical coupling and reciprocal GABAergic synapses (33, 36, 37). The extensive and strong GABA signaling among PV cells is crucial for the formation of a coherent PV network for detecting precise temporal activity patterns (34, 35) and for providing robust and effective inhibitory output in controlling cortical activity patterns (34). Therefore, PV interneurons not only provide certain “levels of inhibition” to PyNs, but also the spatiotemporal pattern of inhibitory control, which is shaped by GABA transmission among PV cells. During the postnatal period in both hippocampus and neocortex, PV interneurons undergo substantial maturation in morphology, connectivity, and intrinsic and synaptic properties (18, 26, 37, 55). A major component of this maturation is the enhancement of GABA signaling among PV cells, which contribute to their conversion from a slow to fast signaling network (37, 55, 56). Importantly, we found a significant reduction in synaptic transmission between PV cells at the 20-Hz range, likely due to a change in release probability, in MeCP2^−/y V1 (Fig. 4 A–C). These results are likely physiologically relevant as PV cells are fast spiking and typically fire burst-of-action potentials in the frequency range of 10–40 Hz in vivo (57). Together, these results suggest that, whereas the increase of GAD levels on vesicular GABA content may be quite modest, the concerted increase of several components that impact GABA release machinery (including vGAT and PV) can lead to a selective enhancement of GABA transmission dynamics among PV cells. We suggest that GABA transmission among PV cells in a fast-spiking regime, similar to their physiological operation, is particularly sensitive to levels of molecular machinery for GABA synthesis and release.

Although a simple prediction from an increased GABA transmission among PV cells might be a reduced PV cell firing and thus enhanced network activity, our in vivo recordings at ∼P30 did not show alterations in firing rates in MeCP2^−/y V1 (Fig. S8). Alternatively, the enhancement of GABA transmission among PV cells at P15 suggests an early maturation of these cells that, together with their electrical coupling and rebound spiking (34), may provide a more coherent inhibitory network in regulating neural activity level and patterns within V1. Indeed, the propagation of neural activity in MeCP2^−/y V1 increases in spatial precision compared with that of WT V1 at P15 and resembles that of P30 WT V1 (Fig. 5 and Figs. S5–S7). The accelerated maturation of PV neurons is unlikely due to abnormal sensory input to V1, as the initial synapse formation, elimination, and connectivity between retinal ganglion cells and relay neurons in the lateral geniculate nucleus were similar between WT and MeCP2^−/y until the fourth postnatal week (58).

An accelerated maturation of the PV network has been shown to trigger an early onset of the critical period (18, 59). On the other hand, once initiated, critical period plasticity rises, peaks, and terminates with a rather fixed schedule (16), in part defined by several molecular markers including those that mark PNNs (29). Thus, our detection of much enhanced PNN at P15 and P30 further hints at an early closure of the critical period. Our direct measurement of the critical period in MeCP2^−/y mice confirmed both these predictions. A recent study using conditional deletion of MeCP2 in PV cells during the third to fourth postnatal week reported an absence of critical period plasticity (60). This study revealed a cell autonomous role of MeCP2 in PV cell physiology and function. Due to the presence of MECP2 protein in ∼85% of PV cells around P20, the onset of critical period plasticity at earlier time points was not measured in these mice (60). In our study, we further demonstrate that the precocious critical period in MeCP2^−/y mice, which reduces the influence of visual experience on the developing V1, compromises experience-dependent circuit wiring and visual function: orientation tuning of individual binocular neurons in mature MeCP2^−/y V1 remain unmatched between the two eyes. This result is consistent with a recent study that demonstrated the impact of a precocious critical period on binocular vision (61). A previous study reported a deficit in glutamatergic synapse maturation and more protracted OD plasticity in V1 in adult MeCP2 heterozygous females (62). As MeCP2^+/− female brains consist of a mosaic of MECP2-expressing and MECP2-null cells due to random X-chromosome inactivation, results obtained from these mutant mice are likely not directly comparable with those from MeCP2^−/y (null) males. Taken together, our results and previous studies (60, 62) suggest that cell specific, germ-line, and mosaic deletions of MeCP2 have different impacts on the developmental trajectory of cortical circuit assembly and function; these should be taken into consideration when discussing their relevance to Rett syndrome.

It is of interest to note that at the retinogeniculate synapse of MeCP2^−/y mice, the initial phase of synapse formation proceeds normally (between P9 and P21) but the subsequent experience-dependent phase of synapse remodeling goes awry (58). Given our finding of an early V1 deficit, it is possible that aberrant V1 to LGN influence may contribute to the failure of experience-dependent remodeling of retinogeniculate synapses. Our results suggest that MeCP2 mutations affect basic aspects of visual function, which might impact acquisition of proper social-behavioral abilities. In fact, Djukic et al. (63) reported recently that Rett patients have more difficulty reading emotional expressions than their counterparts and these problems are linked to atypical eye scanning.

Our results suggest that Gad1 is among the earliest genes misregulated in V1 of germ-line MeCP2^−/y mice, which likely sets in motion a cascade of cellular and circuitry alterations. Importantly, we demonstrate that a reduction of GAD67 expression using Gad1 heterozygote is sufficient to rescue the expression pattern of multiple GABAergic markers and the early onset of the critical period. This result suggests a causal role of GAD67 in the functional impact of MECP2-mediated gene regulation in experience-driven circuit development. It is possible that the early impact of MECP2 deficiency on GAD67 expression is an up-regulation; this may enhance inhibitory transmission, reduce cortical network activity (23), and promote GABAergic synapse formation (21, 47). On the other hand, because GAD67 transcription is highly sensitive to input activity (48), long-term reduction of activity levels could result in progressive down-regulation and possible silencing of Gad1 transcription, as a component of circuit-level compensatory changes. Because alterations of GAD67 expression is implicated in numerous neuropsychiatric disorders, understanding the primary cause and subsequent compensative and maladaptive changes of GAD67 expression in specific cells, circuits, and at defined developmental stages will be critical to elucidate commonly shared circuit level pathogenic mechanisms in these disorders.

GABAergic interneurons are basic components of cortical microcircuits that are conserved across cortical areas. The mechanisms that underlie experience-triggered and MECP2-regulated PV interneuron maturation and critical period timing may also apply to other cortical areas and functional modalities. Hierarchical organization of sensory-motor pathways may develop through a cascade of critical periods as circuit maturation progresses from lower to higher cortical areas (64, 65). Deficient motor, social, and emotional behaviors are characteristic to MeCP2^−/y mice and patients with RTT. Whereas we demonstrate here that MECP2 is necessary to promote experience-dependent alignment of functionally related inputs (e.g., binocular matching onto individual V1 neurons) for proper vision, similar mechanisms may operate in other cortical areas to achieve proper integration of behaviorally relevant sensory-motor inputs or of neuronal pathways that carry appropriate social–emotional information during their corresponding critical periods.

Materials and Methods

All procedures were approved by the Cold Spring Harbor Laboratory and Northwestern University Institutional Animal Care and Use Committee (IACUC) and conducted in the MeCP2^−/y mouse line generated by A. Bird’s laboratory (22). PV⁺ neurons were visualized by crossing MeCP2 heterozygotes into the HG (Rosa26-loxpSTOPloxp-H2BGFP) line (66). Double mutants for MeCP2 and Gad1 were generated by crossing MeCP2 heterozygote females with gad1^+/− males. All control animals were WT littermates for the mutant mice.

Generation of gad1^+/− Mice.

Using homologous recombination in ES cells, a cassette containing destabilized GFP cDNA (D2GFP) was inserted at the translation initiation codon (ATG) of the Gad1 gene. The goal was to generate a Gad1 gene transcription reporter allele, but the same allele is also a gene knockout. This design was essentially the same as the widely used Gad1-GFP knockin allele (67). Targeted ES clones were identified by PCR and Southern blotting. Positive ES clones were injected into C57BL/6 mice to obtain chimeric mice following standard procedures. Chimeric mice were bred with C57BL/6 mice to obtain germ-line transmission. D2GFP expression was weak and restricted to GABAergic neurons throughout the mouse brain, indicating successful gene targeting. The colony is maintained as heterozygotes because homozygotes are lethal.

Immunostaining.

Mice were perfused with saline and 4% (wt/vol) paraformaldehyde at different ages. The brains were removed, postfixed for 18 h in 4% (wt/vol) paraformaldehyde at 4 °C and sectioned at 60-μm thickness using vibratome. Free-floating sections were used for immunostaining using standard protocols. Sections were blocked in 10% (wt/vol) normal goat serum and 1% Triton for 2–3 h, and then incubated overnight with the following primary antibodies: GAD67 (mouse monoclonal; 1:800, Millipore), PV (mouse monoclonal; 1:1,000, Sigma), PV (rabbit polyclonal; 1:1,000, Calbiochem), GAD65 (mouse monoclonal; 1:400, Millipore), VGAT (guinea pig; 1:500, Synaptic Systems), MeCP2 (rabbit; 1:1,000, Cell Signaling), and WFA (1:500; Sigma). Sections were then incubated with appropriate Alexa Fluor dye-conjugated secondary antibodies (1:500; Molecular Probes) and mounted in Fluoromount-G (Southern Biotech). For GAD67 staining, the above mentioned protocol was used for detecting synaptic signal. To obtain GAD67 staining in soma (68), three modifications were made: (i) no Triton or detergent was used in the blocking solution or the primary antibody diluent; (ii) sections were treated with 1% sodium borohydride to reduce background; and (iii) sections were left in primary antibody for 48–60 h at room temperature. Brains were processed blind to the genotype and together at all steps in the process (perfusions, sectioning, immunostaining, and imaging) to minimize variability.

Image Acquisition and Analysis.

Visual cortices from at least three animals were used for each developmental stage and genotype. At least four sagittal sections containing the visual cortex from each animal were analyzed. Images were acquired using the Zeiss LSM510 confocal microscope and analyzed using the LSM Image Browser. To analyze GAD67⁺, PV⁺ soma, and PNNs, nonoverlapping Z-stack images (2.5-μm optical sections for 25 μm) were acquired in infragranular layers with a 20× objective (scan zoom of 0.7, area of 0.6 mm² per image). Scans from each channel were collected in the “multiple tracks” mode and subsequently merged. Maximum intensity projections of the Z stacks were obtained using the “projection” setting in the Zeiss LSM Image Browser. To count high-intensity soma and mature PNNs, the “contrast” setting in the browser was set to 100 to threshold weaker signals. GAD67⁺ and/or PV⁺ somata and mature PNNs were manually counted. Data were represented as mean ± SEM.

To analyze GAD67, GAD65, and VGAT puncta, we focused on perisomatic synapses (18) in layer V of V1. Nonoverlapping Z-stack images (1-μm optical section for 10 μm) were acquired using the 63× oil immersion objective (scan zoom of 1, area of 125 μm²). Intensity and frequency for representative single optical sections at the same Z position for each image were obtained using the “histogram” function in the browser. Differences between different groups were tested for significance using the Kolmogorov–Smirnov test (KS test). Statistical analyses and graphs were done with Prism (GraphPad Software) and Matlab (Mathworks). As the expression levels of these markers increase during development, settings appropriate for each ages were selected for quantification; thus quantifications are not comparable across ages.

Slice Physiology.

Visual cortical slices (300 µm) were prepared from P15–P16 mice. All animals were anesthetized with isofluorane before decapitation. The brain was rapidly dissected and transferred into ice-cold oxygenated artificial cerebrospinal fluid (section ACSF; composition in millimoles: 110 choline-Cl, 2.5 KCl, 4 MgSO₄, 1 CaCl2, 1.25 NaH₂PO₄, 26 NaHCO3, 11 d-glucose, 10 Na-ascorbate, 3.1 Na-pyruvate, pH 7.35, ∼300 mOsm) for 1 min. Coronal slices were dissected with a vibratome (HM 650 V, Microm) at 1–2 °C and further incubated with oxygenated ACSF (working ACSF; composition in millimoles: 124 NaCl, 2.5 KCl, 2 MgSO₄, 2 CaCl2, 1.25 NaH₂PO₄, 26 NaHCO3, 11 d-glucose, pH 7.35, ∼300 mOsm) at 34 °C for 30 min, and then transferred to ACSF at room temperature (25 °C) for >30 min before use. The recordings were located to the V1 area, using the shape of the hippocampus as the primary landmark.

Whole-cell patch recordings were made with pipettes pulled from borosilicate glass capillaries with filament (1.2 mm outer diameter, 0.69 mm inner diameter; Warner Instruments) with a resistance of 3–6 MΩ. For paired recordings and intrinsic properties experiments, pipettes were filled with a solution containing (in millimoles) 130 potassium gluconate, 15 KCl, 10 sodium phosphocreatine, 10 Hepes, 4 ATP-Mg, 0.3 GTP and 0.3 EGTA, pH 7.3 adjusted with KOH, ∼300 mOsm. Layer 5 (L5) PV cells were identified by fluorescence (PV-T2A-cre; HG mice) and PyNs by the giant typical triangular soma and apical dendrites using an upright microscope (BX51; Olympus) equipped with infrared-differential interference contrast (IR-DIC) optics and fluorescence excitation source (X-cite; EXFO). Both IR-DIC images and fluorescence images were captured with a digital camera (Microfire; Optronics). The recordings were performed at 33–34 °C in the chamber perfused with oxygenated working ACSF. Synaptic responses were evoked by presynaptic APs through soma-injected current squares (1.5 ms, 1–1.8 nA). All values were presented as mean ± SEM. Differences between groups were tested for significance using a t test. Statistical analyses and graphs were done with Prism (GraphPad Software) and Matlab (Mathworks). In Fig. 4C and Figs. S3C and S4B, *P < 0.05, **P < 0.01, and ***P < 0.001.

For mIPSC recordings, pipettes were filled with a solution containing (in millimoles) 145 CsCl, 4 NaCl, 0.5 CaCl₂, 5 Hepes, 2 ATP-Mg, and 5 EGTA, pH 7.3 adjusted with KOH, ∼296 mOsm. For mEPSC recordings, the pipettes were filled with internal solution containing (in millimoles): 20 KCl, 100 K-gluconate, 10 Hepes, 0.2% biocytin, 4 Mg-ATP, 0.3 Na-GTP, and 10 Na₂-phosphocreatinine. Layer 5 PV interneurons were identified by fluorescence (PV-T2A-cre; HG mice) and pyramidal neurons by the typical large triangular soma and apical dendrites using an upright microscope (BX51; Olympus) equipped with IR-DIC optics and fluorescence excitation source (X-cite; EXFO). Both IR-DIC images and fluorescence images were captured with a digital camera (Microfire; Optronics). mIPSC recordings were performed at room temperature (23–25 °C) in the chamber perfused with oxygenated working ACSF, which contained 1 µM TTX, 100 µM APV and 20 µM CNOX; voltage clamping at −75 mV without junction potential correction. mIPSC analysis was performed with a total of 14 WT and 12 MeCP2^−/y PV cells and 17 WT and 16 MeCP2^−/y pyramidal cells. mIPSC data were imported to miniAnalysis for frequency and amplitude inspection. All chemicals were purchased from Sigma-Aldrich and Tocris.

VSD Imaging.

VSD imaging was performed as previously described (38, 69). Briefly, Di-4-ANEPPS (Invitrogen; absorption: 496 nm; emission: 705 nm) was dissolved in a 2:1 mixture of ethanol and 10% Chremophor-EL solution (vol/vol in ddH₂O; Sigma) and stored as stock solution with the final concentration of 3.3 mg/mL. For staining, a small volume of the stock solution was dissolved in a 1:1 mixture of FBS (Sigma) and oxygenated ACSF to a final Di-4-ANEPPS concentration of 0.2 mM. Before recording, a slice attached to an Omnipore membrane filter was covered with 100 µL staining solution and incubated for 40 min at room temperature in a recovery chamber containing oxygenated ACSF. The slice was then transferred to the recording chamber mounted on an upright microscope (Olympus; BX51WI) and the excess dye was allowed to be washed off for 10 min with constant flow of oxygenated ACSF at 32 °C. During recording, slices were illuminated with a halogen lamp (150W, TH 4–100; Olympus) controlled by a fast-speed electronic shutter (Smart Shutter with Lamda 10B controller; Sutter Instruments). The slice was activated using an extracellular unipolar stimulating electrode covered by a glass pipette (0.1 MΩ; Harvard Apparatus) and placed either in the white matter immediately below the primary visual cortex or in the center of layer IV depending on the experiment. To activate the slice, 0.2-ms pulses were delivered at a series of intensities increased by steps (0.1–0.9 mA).

The fluorescence signal emitted from the slice was recorded by activating the voltage-sensitive dye with light passing through an excitation filter (λ 530 nm) and projecting onto a dichroic mirror (λ 565 nm). The fluorescence generated by the tissue was passed through an absorption filter (λ 595 nm) and collected by a CCD camera (MiCAM02, SciMedia; Brainvision). The shutter controller, the optical signal acquisition, and the stimulus isolation unit (Warner Instruments) were synchronized by an input/output interface (MiCAM 02, SciMedia; Brainvision). A peristaltic pump (Watson-Marlow Sci 400) and a temperature controller (Warner Instruments) were used to maintain the stability of ACSF flow and temperature through the recording.

After the interested area was visually identified, images were collected with a 4× objective with high numerical aperture (NA 0.28; Olympus) and further widening the imaged field with a 0.5× lens placed on the c-mount of the CCD camera. The image resolution was set to 60 × 88 active pixels with single pixel size of 20 µm, allowing to image changes in fluorescence over an area of 1.2 mm × 1.8 mm. The response to each stimulus was recorded, collecting 256 frames at 400 Hz (one image every 2.5 ms) for a total of 640 ms. To obtain a good signal-to-noise ratio, each image sequence resulted from the average of 16 trials recorded every 15 s.The initial light intensity was set to 50% of maximum illumination to allow for the measurement of increases in activity (quantified as decreases in fluorescence and presented as green to red colors in the figure) and possible decreases in activity (quantified as increases in fluorescence and presented as light blue to dark blue in the figure). The transition in fluorescence was normalized to the baseline fluorescence signal collected 50 ms before the stimulation and quantified, and %∆F/F for each slice was calculated to compare between different conditions. The di-4-ANEPPS has a sensitivity of 10% ∆F/F changes per 100-mV transitions.

Image sequences were analyzed using Brainvision analysis software and ImageJ. Recordings were performed and analyzed blind to the genotype. Data were grouped and compared statistically after the genotype of the animal was revealed. The number of slices and animals for each group was as follows: P15 WT, 6 slices from three animals; P15 MeCP2^−/y, 8 slices from four animals; P30–P35 WT, 11 slices from three animals; and P30–P35 MeCP2^−/y, 8 slices from two animals. Data were presented as mean ± SEM. Statistical analysis was performed using SigmaPlot (Systat Software). To determine significance, the Kruskal–Wallis one-way ANOVA on ranks was performed, and if significant, was followed by Dunn’s test. P values <0.05 were considered significant.

In Vivo Physiology.

WT (n = 7) and MeCP2^−/y (n = 12) mice at ages between P32 and P37 were first sedated with an i.p. injection of chlorprothixene (5 mg/kg) and then anesthetized using urethane (0.5–1.0 g/kg in 10% saline solution, i.p.) as described previously (70). Atropine (0.3 mg/kg) and dexamethasone (2.0 mg/kg) were injected s.c. Additional urethane (0.2–0.3 g/kg) was administered as needed. The eyes of the animal were covered by silicon oil to avoid drying. The animal was placed in a stereotaxic apparatus on a heating pad. The animal’s temperature was monitored with a rectal thermo probe and maintained at 37 °C through a feedback heater control module (Frederick Haer and Company, FHC). Electrocardiograph leads were attached across the skin to monitor the heart rate continuously throughout the experiment. All experimental procedures were approved by Northwestern University Institutional Animal Care and Use Committee.

A small craniotomy (∼2 mm²) was performed at the left hemisphere to expose the cortex for recording. A total of 5–10 MΩ tungsten microelectrodes (FHC) were penetrated perpendicular to the pial surface in the binocular zone of V1 (2.8–3.3 mm lateral from the midline and 0.5–0.8 mm anterior from the lambda suture). In each animal, two to six penetrations were made with minimum spacing of 50 µm and cells recorded across all layers were included in our analysis. Electrical signals, both spikes (filtered between 0.3 and 5 kHz and sampled at 25 kHz) and field potentials (filtered between 10 and 300 Hz and sampled at 800 Hz), were acquired using a System 3 Workstation (Tucker-Davis Technologies) and the spike waveforms were further sorted offline into single units using OpenSorter (Tucker-Davis Technologies). The animals were overdosed with Euthasol (150 mg/kg pentobarbital; Virbac) at the end of recording.

Visual Stimuli and Data Analysis.

Visual stimuli and analysis followed our previous study. Briefly, visual stimuli were generated with Matlab programs (70) using the Psychophysics Toolbox extensions (71, 72). A CRT video monitor (40 × 30 cm, 60 Hz refresh rate, ∼35 cd/m² mean luminance) was placed at 25 cm in front of the animal, with its midline aligned with the animal. The stimulus was delivered through either eye separately; i.e., the stimulus was delivered through the eye contralateral to the recorded hemisphere while the ipsilateral eye was occluded, and vice versa.

Drifting sinusoidal gratings were used to determine V1 neurons’ orientation selectivity (42). The drifting direction (θ) and spatial frequency of the gratings (full contrast and temporal frequency of 2 Hz) were varied between 0° and 360° (12 steps at 30° spacing), and 0.01–0.08 cycle/degree (four logarithmic steps) in a pseudorandom order. The response to a particular stimulus condition, R, was obtained by averaging the number of spikes over the 1.5-s stimulus duration, across all trials and subtracting the spontaneous rate. The preferred direction was determined as the one that gave maximum response (R_pref), averaging across all spatial frequencies. The preferred spatial frequency (pref_SF) was the one that gave peak response at this direction. Responses across all directions at the preferred spatial frequency, R (θ), were used to calculate the preferred orientation, orientation selectivity index (OSI), and tuning width. The maximum R of the 48 conditions was used for evoked neuronal activity, and the average response to blank condition was used for spontaneous neuronal activity.

Half of the complex phase of ∑R(θ)*e^2i*θ/∑R(θ) was calculated (70), which is essentially a weighted mean of θ across a 180° cycle. This value was further converted to the preferred orientation (pref_O) by subtracting 90° because θ was expressed in stimulus direction. The difference in preferred orientation between the two eyes was calculated by subtracting ipsilateral pref_O from contralateral pref_O along the 180° cycle (−90° to 90°). The absolute values of these differences (ΔO) were used in all quantifications (WT: n = 52; MeCP2^−/y: n = 54).

OSI was calculated as the ratio of (R′_pref – R_orth)/(R′_pref + R_orth), where R′_pref was the mean response of R_pref at θ_pref and θ_pref+π as the two angles have the same orientation, and R_orth was the mean response of the two directions orthogonal to the preferred direction. Orientation tuning width was determined by obtaining the half-width at half-maximum response after fitting the tuning curve as the sum of two Gaussians (42, 70). Note that only cells that were well fitted were included in the analysis of tuning width (WT: n = 32; MeCP2^−/y: n = 33). All values were presented as mean ± SEM. Differences between groups were tested for significance using the KS test. Statistical analyses and graphs were done with Prism (GraphPad Software) and Matlab (Mathworks). In Fig. 7, ***P < 0.001.

Optical Imaging of Ocular Dominance Plasticity.

Monocular deprivation of the right eye was performed in WT and MeCP2^−/y mice at P15 or P26 under isofluorane anesthesia (1.5–2% in O₂) following published procedures (73, 74). Mice whose eyelids were not completely sealed close or those with any indication of corneal damage or cataracts were excluded from the study before imaging. The ocular dominance of these mice was determined 5–7 d later by optical imaging of intrinsic signal (73, 75) under isofluorane anesthesia (0.5–1.5% in O₂). A craniotomy was performed to expose the left visual cortex, and the cortex was then covered with agarose and coverslip to form an imaging window. The visual stimulus was a thin bar (2° in height and 20° in width) drifting continuously and periodically upward or downward. It was shown between −5° to 15° azimuth (the vertical meridian defined as 0° with negative values for ipsilateral visual field) and full-screen elevations. The spatial frequency of the drifting bar was one cycle/100°, and temporal frequency one cycle/8 s. Optical images were acquired at 610 nm using a 1M30 CCD camera (Dalsa) and the Fourier component of the reflectance changes was extracted at the temporal frequency of the stimulus as described previously (75).

The response magnitude map evoked by ipsilateral eye stimulation was first smoothed by a uniform kernel of 5 × 5 filter and then thresholded at 30% of the peak response amplitude, to determine the binocular zone for ocular dominance analysis. The ocular dominance index was calculated as the mean of (C – I)/(C + I) of all pixels within the selected region, where C and I represent the response magnitude to the contralateral and ipsilateral eyes, respectively (73). The ODI ranges from +1 to −1, where a positive value indicates a contralateral bias and negative ODI, an ipsilateral bias. The Mann–Whitney test was used to test for statistical significance between two groups in optical imaging experiments. Statistical analyses and graphs were done with Prism (GraphPad Software). In Figs. 6 and 8, *P < 0.05 and **P < 0.01.