Abstract
Purpose:
DNA methylation is an important epigenetic mechanism, which is orchestrated in mammalian genome by three DNA methyltransferases (Dnmts): Dnmt1, Dnmt3a and Dnmt3b. These enzymes catalyze the conversion of cytosine to 5’-methyl cytosine. Heavily methylated DNA, a mark for gene silencing, is typical for transcriptionally inactive, closed heterochromatin. Unmethylated or hypomethylated DNA is present in transcriptionally active (open) euchromation, typical for stem/progenitor cells. Dnmt1 is the "maintenance" DNA methyltransferase that replicates the methylation pattern on the daughter DNA strand during DNA replication. Dnmt3a and Dnmt3b were reported to be involved in “de novo” DNA methylation and their primary role was linked to establishing tissue-specific DNA methylation patterns. The level of DNA methylation in cells has good correlationwith stemness (developmental plasticity) and the ability of cells to re-enter cell division. The purpose of this study is to elucidate the role of Dnmts in the development of mammalian retina, which is an excellent model to study (i) differentiation mechanisms in the Central Nervous System (CNS) and (ii) CNS regeneration mechanisms.v
Methods:
We earlier generated conditional knockdown of Dnmt1 with Rx-Cre driver done at embryonic (e) day 10 and reported a profound impact on photoreceptor [PR] development. In this study, we developed conditional models of retinal inactivation of Dnmts using Rx-Cre, Lhx2-Cre and Pax6-alpha-Cre for Dnmt1, Dnmt1+Dnmt3a, Dnmt1+Dnmt3b, and all three Dnmt genes in neural retina/RPE, and separately, in neural retina alone (Pax6-alpha-Cre). Cre-loxP recombination was evaluated by doing excision tests on genomic DNA isolated from the neural retina and from RPE. Retinal development was evaluated in mutant and control eyes by histology, immunofluorescence, flatmounts and electroretinogram. Expression analysis was done by quantitative (q) RT-PCR.
Results:
Retina from Dnmt1, Dnmt3a, Dnmt3b flox/flox Rx-cre mice had a robust loss of PR outer segments at postnatal day (P)15.5, retinal thinning and very hypoplastic and thin RPE layer (P15.5), which by histological analysis displayed somewhat more severe phenotype, compared to that in Dnmt1 flox/flox, RxCre mice.
Conclusions:
Our initial data point to the likely contribution of Dnmt3a/3b to neural retina and RPE terminal differentiation.