Rett syndrome (RTT) is an X chromosome-linked neurodevelopmental disorder associated with the characteristic neuropathology of dendritic spines common in diseases presenting with mental retardation (MR). Here, we present the first quantitative analyses of dendritic spine density in postmortem brain tissue from female RTT individuals, which revealed that hippocampal CA1 pyramidal neurons have lower spine density than age-matched non-MR female control individuals. The majority of RTT individuals carry mutations in MECP2, the gene coding for a methylated DNA-binding transcriptional regulator. While altered synaptic transmission and plasticity has been demonstrated in Mecp2-deficient mouse models of RTT, observations regarding dendritic spine density and morphology have produced varied results. We investigated the consequences of MeCP2 dysfunction on dendritic spine structure by overexpressing (∼twofold) MeCP2-GFP constructs encoding either the wildtype (WT) protein, or missense mutations commonly found in RTT individuals. Pyramidal neurons within hippocampal slice cultures transfected with either WT or mutant MECP2 (either R106W or T158M) showed a significant reduction in total spine density after 48hrs of expression. Interestingly, spine density in neurons expressing WT MECP2 for 96hrs was comparable to that in control neurons, while neurons expressing mutant MECP2 continued to have lower spines density than controls after 96hrs of expression. Knockdown of endogenous Mecp2 with a specific small hairpin interference RNA (shRNA) also reduced dendritic spine density, but only after 96hrs of expression. On the other hand, the consequences of manipulating MeCP2 levels for dendritic complexity in CA3 pyramidal neurons were only minor. Together, these results demonstrate reduced dendritic spine density in hippocampal pyramidal neurons from RTT patients, a distinct dendritic phenotype also found in neurons expressing RTT-associated MECP2 mutations or after shRNA-mediated endogenous Mecp2 knockdown, suggesting that this phenotype represent a cell-autonomous consequence of MeCP2 dysfunction.

The Na+-dependent glutamate transporter, GLT-1 (EAAT2), shows selective expression in astrocytes, and neurons induce expression of GLT-1 in astrocytes. In unpublished analyses of GLT-1 promoter reporter mice, we identified an evolutionarily conserved domain of 467 nucleotides ~8 kb upstream of the GLT-1 translation start site that is required for astrocytic expression. Using in silico approaches, we identified Pax6 as a transcription factor that could contribute to the control of GLT-1 expression by binding within this region. We demonstrated expression of Pax6 protein in astrocytes in vivo. Lentiviral transduction of astrocytes with exogenous Pax6 increased expression of enhanced green fluorescent protein (eGFP) in astrocytes prepared from transgenic mice that use a bacterial artificial chromosome (BAC) containing a large genomic region surrounding the GLT-1 gene to control expression of eGFP. It also increased GLT-1 protein, and GLT-1-mediated activity, while there was no effect on the levels of astroglial glutamate transporter, GLAST. Transduction of astrocytes with an shRNA directed against Pax6 reduced neuron-dependent induction of GLT-1 or eGFP. Finally, we confirmed Pax6 interaction with the predicted DNA binding site in electrophoretic mobility assays (EMSA) and chromatin immunoprecipitation (ChIP). Together, these studies show that Pax6 contributes to regulation of GLT-1 through an interaction with these distal elements and identify a novel role of Pax6 in astrocyte biology.