Restoration of Mecp2 expression in GABAergic neurons is sufficient to rescue multiple disease features in a mouse model of Rett syndrome

Abstract

The postnatal neurodevelopmental disorder Rett syndrome, caused by mutations in MECP2, produces a diverse array of symptoms, including loss of language, motor, and social skills and the development of hand stereotypies, anxiety, tremor, ataxia, respiratory dysrhythmias, and seizures. Surprisingly, despite the diversity of these features, we have found that deleting Mecp2 only from GABAergic inhibitory neurons in mice replicates most of this phenotype. Here we show that genetically restoring Mecp2 expression only in GABAergic neurons of male Mecp2 null mice enhanced inhibitory signaling, extended lifespan, and rescued ataxia, apraxia, and social abnormalities but did not rescue tremor or anxiety. Female Mecp2(+/-) mice showed a less dramatic but still substantial rescue. These findings highlight the critical regulatory role of GABAergic neurons in certain behaviors and suggest that modulating the excitatory/inhibitory balance through GABAergic neurons could prove a viable therapeutic option in Rett syndrome.

Keywords: GABAergic neurons; Rett syndrome; inhibition; mouse; neuroscience.

Figure 1.. Conditional restoration of Mecp2 expression solely in GABAergic neurons normalized body weight and appearance and significantly extended lifespan.

(A) Schematic of breeding and mice generated (left column). Photos of each genotype show male Viaat-Cre; Mecp2^lox-Stop/Y mice were indistinguishable from wildtype mice at 4 months of age (second column). Cell images show expression of MeCP2 (green) was limited to GAD67-positive GABAergic neurons (red; third column, arrowheads) and was not found in CamKII-expressing excitatory neurons (red, right column, arrowheads). (B) Male rescue mice maintained similar body weights to wildtype males throughout life. (C) Male rescue mouse lifespan was significantly extended, with ~40% surviving past 1 year of age. n = 33–47 per genotype. Scale bar is 20 μm. DOI: http://dx.doi.org/10.7554/eLife.14198.003

Figure 1—figure supplement 1.. Viaat-Cre was not expressed in glia, microglia, or bone marrow.

(A–B)Viaat-Cre; ROSA-YFP did not colocalize with glial marker GFAP (A) or activated microglia marker Iba1 (B) in the cortex. Scale bars = 40 μm. (C) No YFP+ cells were detected in the bone marrow of Viaat-Cre;ROSA-YFP mice. DOI: http://dx.doi.org/10.7554/eLife.14198.004

Figure 1—figure supplement 2.. Male rescue mice maintained more of their body weight up to the point of death and showed improved EEG activity.

(A) Male rescue mice lose less body weight from their peak body weight until death than Mecp2^lox-STOP/Y littermates. (B) Abnormal EEG activity has been previously reported in Mecp2^-/Y mice (Chao et al., 2010) and was observed in two of the three recorded Mecp2^lox-STOP/Y male mice (third trace). Five rescue mice (fifth trace) had similar EEG traces to wildtype and Viaat-Cre controls (first and second trace, respectively), while one rescue animal displayed abnormal activity (fourth trace). Wildtype = 5, Viaat-Cre = 2, Mecp2^lox-STOP/Y = 3, Rescue = 6. DOI: http://dx.doi.org/10.7554/eLife.14198.005

Figure 2.. Male rescue mice exhibited significant behavioral rescue.

(A–B) Male rescue mice were indistinguishable from wildtype and Viaat-Cre controls at 6 weeks of age on the rotarod (n = 23–28 per genotype) (A) and grip strength (B) assays (n = 20–26 per genotype). (C) Rescue mice trended toward improved locomotor activity in the open field assay at 9 weeks of age (n = 8–15 per genotype). (D–E) Apraxia at 9 weeks of age was reversed in rescue mice as indicated by marble burying (n = 9–15) (D) or, at 8 weeks, by nest building (E, n = 23–27 per genotype). (F) Rescue mice had similar sociability to wildtype mice in the partition test at 8 weeks of age. * represents significance of genotype (by color) compared to Mecp2^lox-Stop/Y. (G) Acoustic startle response was unchanged in rescue mice at 8 weeks of age (n = 16–21). (H−I) Rescue mice were similar in anxiety behavior to Mecp2^lox-Stop/Y in the elevated plus maze test (G) but were similar to wildtype in the light dark box assay (I). Error bars show SEM. *p<0.05 **p<0.01 ***p<0.001 DOI: http://dx.doi.org/10.7554/eLife.14198.006

Figure 3.. Rescue persisted in male mice at 30 weeks of age.

(A) Male rescue mice at 30 weeks of age continued to appear indistinguishable from wildtype and Viaat-Cre littermates. (B) Locomotion in the open field assay decreased when compared to wildtype but not Viaat-Cre. (C–D) Rotarod performance was maintained at wildtype levels, as was grip strength (D). n = 20–25 per genotype. Error bars show SEM. *p<0.05 DOI: http://dx.doi.org/10.7554/eLife.14198.007

Figure 4.. Expression of Mecp2 in GABAergic neurons normalizes levels of GABA and other MeCP2 responsive genes.

(A) GABA concentrations were restored to wildtype levels in striatum from rescue mice (n = 4–5 per genotype). (B) RNA expression of genes related to GABAergic neuronal function, particularly gad1 and gad2, were normalized in rescue mouse cerebella. (C) Genes known to be downregulated or upregulated in Mecp2-null cerebellum were normalized in the rescue mice. Error bars show SEM. *p< 0.05 **p<0.01 ***p<0.001 ****p<0.0001. n = 6–7 per genotype. DOI: http://dx.doi.org/10.7554/eLife.14198.008

Figure 5.. Inhibitory signaling in rescue mice showed some improvement while excitatory signaling was unchanged.

(A–B) Sample traces of mIPSC (A, Wildtype = 11, Viaat-Cre = 17, Mecp2^lox-STOP/Y = 14, Rescue = 11) and mEPSC (B, Wildtype = 17, Viaat-Cre = 16, Mecp2^lox-STOP/Y = 18, Rescue = 17) from pyramidal cells of the somatosensory cortex of wildtype, Viaat-Cre, Mecp2^lox-STOP/Y, and Rescue male mice, including cumulative distributions of amplitude and interval, grand average minis, and summaries of frequency, amplitude, decay, and average charge. Error bars show SEM. *p<0.05 DOI: http://dx.doi.org/10.7554/eLife.14198.009

Figure 5—figure supplement 1.. Inhibitory synapse numbers and spontaneous action potential firing are normalized in rescue mice.

(A) VGAT+ inhibitory synapses are decreased in the CA1 of Mecp2^lox-STOP/Y mice but are normalized in the rescue; Vglut1+ excitatory synapses are not significantly changed in any genotype. (B) Left: Representative traces of spontaneous action potential firing from layer V cortical pyramidal neurons. Right: Summary of spontaneous firing rates across genotypes (wildtype n = 15; Viaat-Cre = 12; Mecp2^-/Y = 20; Rescue = 17). (C) Left: Representative traces of intrinsic neuronal excitability. Right: Summary of intrinsic neuronal excitability across genotypes. Error bars show SEM. *p<0.05 DOI: http://dx.doi.org/10.7554/eLife.14198.010

Figure 6.. Female rescue mice exhibited a partial but sustained functional recovery.

(A) Female rescue mice appeared grossly normal at 23 weeks of age. (B) Female rescue mice body weight was partially rescued (n = 27–37 per genotype). (C) Female rescue mice performed better at 9 weeks of age on the accelerating rotarod than Mecp2^lox-STOP/+ females, but not to wildtype levels (n = 14–16 per genotype). (D) Acoustic startle response at 9 weeks of age was not rescued (n = 14–17 per genotype). (E–G) At 30 weeks of age, female rescue mice showed partial rescue of grip strength (E) and complete rescue of hypersociability in the partition test (F) and nesting ability (n = 13–18 per genotype) (G). (H–K) Rescue females exhibited a partial but sustained rescue of locomotion in the open field (H), rotarod (I), footslip (J), and grip strength (K) assays (n = 20–35 per genotype). Error bars show SEM. *p<0.05 **p<0.01 ***p<0.001 ****p<0.0001. DOI: http://dx.doi.org/10.7554/eLife.14198.011

Figure 6—figure supplement 1.. Female rescue mice show partial rescue particularly at late ages.

(A–C) At 9 weeks of age, no differences in locomotion were noted (A, n = 11–17 per genotype). (B) 18-week-old female mice of all genotypes showed similar levels of motor activity in the open field assay. (C) Mecp2^lox-Stop/+ females exhibited more footslips, whereas rescue females were more coordinated (n = 8–18 per genotype). (D) Even at 30 weeks of age, females of all genotypes exhibited similar levels of motor activity (though less than younger mice; n = 14–17 per genotype). (E–F) Mecp2^lox-Stop/+ females were unable to remain balanced on a dowel at 30 weeks (E, n = 14–18 per genotype) or 13 months (F, n = 25–35 per genotype) of age, but rescue mice performed significantly better at both time points. Error bars show SEM. *p<0.05 **p<0.01 ***p<0.001. DOI: http://dx.doi.org/10.7554/eLife.14198.012