siRNA-mediated Cell-Autonomous MeCP2 Knockdown Increases the Proportion of Immature “Tortuous” Dendritic Spines in CA1 Pyramidal Neurons of Hippocampal Slices

CHAPLEAU CA, CALFA G, RUTHERFORD JM, POZZO-MILLER L

Chicago

Congreso; 8th Annual Rett Syndrome Symposium, International Rett Syndrome Foundation (IRSF); 2007

IRSF

It has been long recognized that neurological disorders associated with mental retardation, including Rett Syndrome (RTT), are accompanied with alterations in the number and form of dendritic spines in cortical pyramidal neurons. Dendritic spines are the small processes extending from the surface of dendrites where the majority of excitatory glutamatergic synapses are formed, thus playing a fundamental role in synaptic transmission as well as in synaptic plasticity models of learning and memory. Mutations in the gene encoding the transcriptional repressor methyl-CpG DNA binding protein 2 (MeCP2) are associated with the majority of RTT cases (Amir et al. 1999). Although mouse models carrying a null deletion of the MECP2 gene (Chen et al. 2001; Guy et al. 2001), or expressing a truncated, non-functional form of the MeCP2 protein (MeCP2308; Shahbazian et al. 2002) recapitulate several cellular, physiological and behavioral features of human RTT, the analyses of their dendritic and synaptic structure have produced varied results. In a study of symptomatic null MECP2 mice, pyramidal neurons in layer II/III of the neocortex were found to be smaller and having less complex dendritic organization; however, no differences in dendritic spine density were observed (Kishi and Macklis 2004). On the other hand, dendritic complexity was not affected in pyramidal neurons of the frontal cortex of MeCP2308 mice, while no change in spine synapse number was observed in CA1 pyramidal neurons of the hippocampus (Moretti et al. 2006). Considering these apparent discrepancies with earlier observations in RTT brains (Armstrong et al. 1995; Belichenko et al. 1994) and the mosaic nature of mutant MECP2 expression in RTT girls, we chose to use a biolistic gene-transfer approach, which yields only a few transfected neurons surrounded by wild-type tissue. In those experiments, hippocampal pyramidal neurons expressing RTT-associated MECP2 mutations showed reduced dendritic complexity and spine density, while those expressing wild-type MECP2were not different to control neurons transfected with eYFP (for neuronal morphology). To determine whether these effects resulted from a gain- or loss-offunction of MeCP2, we next investigated the consequences of siRNA-mediated cell-autonomous MeCP2 knockdown on dendritic spine density and form. The effectiveness of a cDNA plasmid expressing a smallhairpin interference RNA designed to knockdown MeCP2 was confirmed in PC12 cells, where it significantly reduced MeCP2 expression levels. Next, this MeCP2 siRNA plasmid was transfected into hippocampal slice cultures using a gene-gun, followed 96hrs later by quantitative analyses of dendritic spine density and form by confocal microscopy of eYFP, which was co-transfected to reveal neuronal morphology. In these studies, CA1 pyramidal neurons transfected with MeCP2 siRNA had spine densities in their apical dendrites similar to control neurons expressing only eYFP (Ctl 8.7±0.7 spines/10mm vs. MeCP2 siRNA 6.9±1.0 spines/10mm; n=10, p=0.17). Together with the significant reduction in spine density in neurons expressing RTT-associated MECP2 mutations, and the observations of unaltered spine density in MECP2 null and MeCP2308 mice, these results suggest that mutant MeCP2 expression leads to spine loss by a gain-of-function mechanism. However, MeCP2 knockdown caused a significant increase in the proportion of thin spines (Ctl 0.12±0.01 vs. MeCP2 siRNA 0.22±0.03; p<0.05), without affecting the proportion of the other major spine types, i.e. stubby (Ctl 0.59±0.04 vs. MeCP2 siRNA 0.48±0.05; p>0.05) and mushroom (Ctl 0.28±0.03 vs. MeCP2 siRNA 0.30±0.03; p>0.05). Since long and thin spines are thought to represent immature spines, and such tortuous spines predominate in several mental retardation-associated disorders, the effect of MeCP2 knockdown on spine morphology seems to recapitulate a cellular RTT phenotype through a loss-of-function mechanism. It remains to be determined whether these tortuous immature spines also predominate in the hippocampus of  MeCP2 animal models and RTT patients.