BioProject

The realization in the last decade that dysregulated microglia are intimately involved in Alzheimer’s disease (AD) pathogenesis has been a major advance. More...

The realization in the last decade that dysregulated microglia are intimately involved in Alzheimer’s disease (AD) pathogenesis has been a major advance. The precise mechanisms controlling pathogenic programs of microglia gene expression, however, remain poorly understood. The transcription factor (TF) c/EBPß is highly expressed in microglia and is known to regulate expression of pro-inflammatory genes. Notably, c/EBPß is upregulated in AD. Despite mounting evidence that the levels of this pivotal pro-inflammatory TF are tightly regulated, how this is achieved is unclear as alterations in its amounts are not reflected in transcript levels. Remarkably, we find that this TF is primarily regulated post-translationally. Here we show that the ubiquitin ligase Cop1 functions as a “brake” on microglial activation by targeting c/EBPß for ubiquitination and subsequent proteasomal mediated degradation. In the absence of Cop1, c/EBPß protein rapidly and dramatically accumulates leading to engagement of a potent pro-inflammatory and ApoE gene-expression program, evidenced by increased neurotoxicity in microglia-neuronal co-cultures. Antibody blocking studies revealed that the neurotoxicity was almost entirely attributable to complement. Unexpectedly, loss of a single allele of c/EBPß, rescued the pro-inflammatory phenotype underscoring a significant gene dosage effect. We also found that Cop1 deletion accelerated disease progression in a mouse model of tau-mediated neurodegeneration where elevated ApoE plays a deleterious role. Taken together these results point to c/EBPß as a potential therapeutic target for inflammation-driven neurodegeneration as the heterozygote animal is otherwise normal. Overall design: ChIP-seq of CEBPb and 2 histone modifications (H3K4me3, H3K27ac) in microglia was performed on the following genotypes in two replicates: Cop1 wildtype and Cop1 knock out. We had one pooled input sample as control. Less...