Loss and Gain of MeCP2 Cause Similar Hippocampal Circuit Dysfunction that Is Rescued by Deep Brain Stimulation in a Rett Syndrome Mouse Model

Abstract

Loss- and gain-of-function mutations in methyl-CpG-binding protein 2 (MECP2) underlie two distinct neurological syndromes with strikingly similar features, but the synaptic and circuit-level changes mediating these shared features are undefined. Here we report three novel signs of neural circuit dysfunction in three mouse models of MECP2 disorders (constitutive Mecp2 null, mosaic Mecp2(+/-), and MECP2 duplication): abnormally elevated synchrony in the firing activity of hippocampal CA1 pyramidal neurons, an impaired homeostatic response to perturbations of excitatory-inhibitory balance, and decreased excitatory synaptic response in inhibitory neurons. Conditional mutagenesis studies revealed that MeCP2 dysfunction in excitatory neurons mediated elevated synchrony at baseline, while MeCP2 dysfunction in inhibitory neurons increased susceptibility to hypersynchronization in response to perturbations. Chronic forniceal deep brain stimulation (DBS), recently shown to rescue hippocampus-dependent learning and memory in Mecp2(+/-) (Rett) mice, also rescued all three features of hippocampal circuit dysfunction in these mice.

Figure 1. Loss of the asynchronous network state in hippocampal CA1 of MeCP2 mutant mice

(A) Schematic of ex vivo 2-photon imaging. (B) Left, 2-photon image of hippocampal slice from a thy1-GCaMP3 mouse. Box: pyramidal layer of hippocampal CA1. Middle, High-magnification image of GCaMP3-labeled CA1 pyramidal neurons. Right, Sample dF/F (% change in fluorescence intensity, see Methods) calcium traces from four simultaneously imaged neurons labeled with arrows in the middle image. (C) Left: representative heat map plots of calcium transients. Each row is a neuron; each column is an imaging frame (~167 ms). Arrows point to synchronous events. Right: graph visualization of all neurons recorded simultaneously in this experiment. The location of each neuron is denoted by a triangle. Each neuron’s event rate is represented by the triangle’s brightness (linear scale, range .003–.3 Hz). Significant positive correlations between neurons (2 s.d. above the shuffled distribution mean; see Methods) are plotted as graph edges, where the thickness/brightness of the edge signifies the strength of the positive correlation (linear scale, range 0–0.6). (D) Cumulative histogram of inter-neuronal correlation coefficients per cell pair in Null (blue, left), Tg1 (orange, middle) and Rett (purple, right) mice vs. wild-type (WT) littermates (black). Dotted lines represent shuffled correlation distributions calculated from calcium traces circularly permuted in time (see Methods). *** p < 0.001. Mann-Whitney U test or Kolmogorov-Smirnov test. Insets: Mean Pearson correlation per cell pair as a function of distance between cells in Null (left) and Tg1 (right). ** p < 0.01, ANOVA. (E) Example brightfield image of coronal hippocampal slice from a tetrode-implanted animal. Arrow indicates electrolytic lesion produced by the tip of the tetrode, confirming localization in stratum pyramidale of CA1. (F) Heat map depiction of simultaneously recorded tetrode single-unit firing rates in hippocampal CA1 of Rett mice (bottom) and WT littermates (top) during quiescent wakefulness. Arrows point to synchronous events. Black traces show the simultaneously recorded hippocampal local field potential, high-pass filtered to show ripple events. (G) Cumulative histogram of correlation coefficients measured from Rett mice (purple, n=739 pairs, 123 neurons, 6 animals) and WT mice (black, n=569 pairs, 100 neurons, 5 animals). *** p < 0.0001, Mann-Whitney U test or Kolmogorov-Smirnov test. Inset, correlation coefficient at each firing rate bin size for Rett and WT mice, in the awake quiescent (dotted line) and awake moving (solid line) state. Note that pairwise neuronal correlations are consistently elevated in Rett mice across firing rate bin sizes in both behavioral states. Arrow points to the bin size (167 ms) of the data sets used in panel F and the cumulative histogram. Error bars represent mean ± SE.

Figure 2. Loss of neural circuit homeostasis in models of MeCP2 disorders

(A) Heat map depiction of example traces from WT, Null, Tg1 and female heterozygous Rett CA1 pyramidal neurons before (left) and after (right) bath application of small amounts (0.5 μM) of the GABA_A receptor blocker Gabazine. (B) Cumulative histogram of inter-neuronal correlation coefficients per imaging site obtained with low-dose Gabazine in Null (blue, left), Tg1 (orange, middle) and Rett (purple, right) mice vs. WT littermates (black). Dotted lines represent shuffled correlation distributions (see Methods). * p < 0.05, ** p < 0.01, Mann-Whitney U test or Kolmogorov-Smirnov test. (C) Normalized mean CA1 pyramidal neuron correlation (corr.) coefficients of Null, Tg1, Rett mice and WT littermates at baseline (BL) and with low-dose Gabazine (GA). Correlation coefficient of the mutants was normalized to WT at baseline and with low-dose Gabazine, respectively. Note that correlation coefficient strength in the mutants was disproportionately increased after low dose Gabazine compared to baseline. ** p <0.01, *** p < 0.0001, Mann-Whitney U test, analyzed per pair. See Table 1 for n numbers. (D) Normalized mean correlation (corr.) coefficients of CA1 pyramidal neuron in excitatory-specific Mecp2 knockout mice (E-cKO, teal), somatostatin neuron-specific KO (SST-cKO, magenta), parvalbumin-neuron-specific KO (PV-cKO, maroon), and excitatory-specific Mecp2 overexpression (E-cOE, light orange) mice and control littermates at baseline and with low-dose Gabazine. Correlation coefficient of the mutants was normalized to control (combination of WT and Cre) at baseline and with low-dose Gabazine, respectively. ** p < 0.01, *** p < 0.0001, Mann-Whitney U test, analyzed per pair. See Table 1 for n numbers.

Figure 3. Shared circuit dysfunction at the synapse level in MeCP2 disorders

(A) Bright field image of hippocampal CA1 acquired with 40X objective lens of Zeiss Axio Examiner D1 upright microscope. Excitatory pyramidal neurons were identified by their localization in the stratum pyramidale and their adapting response to current injection (left). Inhibitory interneurons were identified by their somatic localization in the stratum oriens and their characteristic fast-spiking, non-adapting response to current injection (right). (B) Spontaneous EPSCs (sEPSC) measured from CA1 oriens-layer interneurons (I-sEPSC) in Null (blue n=11), Tg1 (orange n=19), and WT littermates (black, n=9 and 19, respectively). (C) Spontaneous EPSCs measured from CA1 pyramidal neurons (E-sEPSC) in Null (blue n=14), Tg1 (orange n=17), and WT littermates (black n=13, n= 15). From left to right: Experiment schematic, example sEPSC traces, mean sEPSC frequency, mean sEPSC amplitude. * p < 0.05, ** p < 0.01, t test.

Figure 4. Forniceal DBS rescued the abnormal network activity and synaptic transmission in hippocampus of Rett mice

(A) Brightfield picture of coronal sections illustrating the placement of the DBS electrode in the fornix (left) and the recording electrode in CA1 (middle). Note that an anodal current was passed through the electrodes at the end of the experiments for identification of electrode placement. cc: corpus callosum; LV: lateral ventricle; D3V: dorsal 3rd ventricle; DG: dentate gyrus. Right, Representative evoked potential trace of the fimbria-fornix pathway recorded in CA1. (B, C) Summary of averaged inter-neuronal correlation coefficients per imaging site obtained before (B) and after (C) bath-application of 0.5 μM Gabazine in DBS- or sham-treated Rett mice (purple) vs. WT littermates (black). * p < 0.001, ANOVA. (D) Comparison of averaged inter-neuronal correlation coefficients per imaging site obtained before and after bath-application of 0.5 μM Gabazine in DBS- or sham-treated Rett mice (purple) vs. WT littermates (black). Refer to B and C for statistics as same data set is used. (E) Sample Images showing Immunostaining for Dapi, biocytin, MeCP2 and Somatostatin (SST) in a slice from Rett mice with two oriens-layer interneurons recorded for sEPSCs via class pipettes loaded with 0.2% biocytin. White arrows indicate the soma of interneurons recorded. Both interneurons are MeCP2-positive and only one of them is SST-positive. The merged image of Dapi and MeCP2 staining showed mosaic expression of MeCP2. Scale: 50 μm. Insert: zoom-in image of soma area. (F) Experiment schematic of recordings from oriens-layer interneuron. (G) Mean sEPSC frequency (left) and amplitude (right) of sEPSC measured from CA1 oriens-layer interneurons in DBS- or sham-treated Rett mice (purple) vs. WT littermates (black). Numbers in parentheses represent the number of neurons recorded for each condition. Frequency (left) and amplitude (right) of sEPSC were calculated from the same set of neurons. * p < 0.05 to all other groups, t test.