The Role of the Histone Demethylase KDM5A in Cellular Differentiation

The retinoblastoma tumor suppressor pRB is required for differentiation of several tissues. pRB is a transcriptional regulator. It functions as both transcriptional activator and repressor by interacting with a multitude of co-regulators. It has been shown that the lysine-specific histone demethylase KDM5A is a critical co-regulator in RB-mediated differentiation. In RB1 null cells, the additional loss of the KDM5A gene can restore features of the differentiated phenotype indicating an antagonistic relationship with pRB. However it is also known that KDM5A can cooperate with members of the RB family to repress cell cycle genes during differentiation. The mechanism by which the loss of KDM5A rescues differentiation has not been determined, and the details of KDM5A cooperation with the RB family to repress cell cycle genes during differentiation have not been fully explored. The current model based on previous works proposes that KDM5A negatively regulates a set of genes by demethylation of trimethylated histone 3 on lysine 4 (H3K4me3) and that these genes must be upregulated to allow differentiation to proceed. pRB promotes differentiation at least in part by relieving the repressive effect of KDM5A on this gene set. Simultaneously KDM5A binds to a set of cell cycle genes as differentiation progresses to drive the requisite cell cycle withdrawal in cooperation with the RB family proteins. The work presented here further explores this model to determine the mechanism of regulation of differentiation, and the requirement for KDM5A demethylase function during differentiation and cell cycle exit. I have used the myogenic model of differentiation with mouse embryonic fibroblasts from WT, Rb1-/-, Kdm5a-/-, and double knockout mice (DKO) to identify the critical genes during differentiation to be a large set that codes for components of the mitochondria. These genes are direct targets of both pRB and KDM5A. They are deregulated by the loss of Rb1-/- and rescued by the additional loss of Kdm5a. Accordingly the mitochondrial phenotype mirrors the expression pattern of these genes. Mitochondrial biogenesis and function are defective in Rb1-/- MEFs induced to differentiate. Additional loss of Kdm5a in the DKOs rescues these defects. Additionally I show that a functional catalytic domain is required for KDM5A function in cell cycle withdrawal and differentiation.