5-Hydroxymethylcytosine Is Increased in Autism Cerebellum and within the EN-2 Gene: Epigenetic Implications

Background:  5-hydroxymethylcytosine (5-hmC) is created by the oxidation of 5-methylcytosine (5-mC) by TET enzymes and is most abundant in the brain, with highest concentration in Purkinje neurons.  5-hmC is now known to mediate the process of DNA demethylation and is often associated with activated neuronal genes.  It’s role as a stable epigenetic mark involved in the regulation of gene expression and it’s association with MeCP2 binding and oxidative stress is less well defined.  We recently reported DNA hypermethylation in the autism cerebellum and in the promoter region of the EN-2 gene, a HOX transcription factor required for normal cerebellar development. Unexpectedly, promoter hypermethylation was associated with an increase in EN-2 protein and gene expression.  Of note, the 5-mC binding protein MeCP2 does not bind strongly to intragenic 5-hmC.  Reduced  binding of the MeCP2 repressive complex could lead to local chromatin conformation changes and gene overexpression.  It is important to note that current methods for detection of DNA methylation are not able to distinguish between 5-mC and 5-hmC; therefore, the independent contribution of each requires clarification for accurate interpretation.

Objectives: To simultaneously measure 5-mC and 5-hmC content in cerebellar tissue and within the EN-2 gene and explore the association of 5-hmC with MeCP2 binding and with 8-oxo-dG, a marker of oxidative stress and DNA damage.  

Methods: In 15 case and 15 control post mortem cerebellar samples, 5-mC, 5-hmC and 8-oxo-dG were quantified by LC/MS/MS after enzymatic digestion to DNA bases and expressed per μg DNA.  The EpiMark™ 5-hmC and 5-mC Analysis Kit was used to quantify 5-hmC and 5-mC within the EN-2 gene. MeCP2 binding to DNA was determined using standard chromatin immunoprecipitation methods.  TET enzyme gene expression was measured using quantitative reverse transcription real-time PCR.

Results: 5-hmC and 8-oxo-dG were significantly increased relative to control cerebellar samples and demonstrated a positive correlation whereas tissue glutathione levels were significantly decreased.  Unexpectedly, 5-mC levels were also significantly increased in case compared to control cerebellum and were positively correlated with 5-hmC levels.  These data suggest that elevated levels of 5-hmC were stable and independent of TET 5-mC demethylation activity.  Consistent with this possibility, TET2 gene expression was not significantly different between case and control cerebellum samples.  Within the EN-2 promoter region, 5-hmC and 5-mC levels were also significantly increased and positively correlated.  MeCP2 binding to the EN-2 promoter was significantly decreased relative to control samples. 

Conclusions: Previous reports of altered DNA methylation in autism brain need to be redefined in terms of the proportion of cytosine methylation due to 5-hmC and to 5-mC.  The presence of 5-hmC in cerebellum appears to be a stable epigenetic mark that is independent of TET enzyme 5-mC demethylation.  Increased levels of 5-hmC  may decrease gene-specific binding of MeCP2 and lead to gene overexpression.  Finally, the positive correlation between 5-hmC and 8-oxo-dG suggests the possibility that oxidative stress may be a non-enzymatic mechanism to oxidize 5-mC to 5-hmC.