Despite the significant increase in Bdnf mRNA levels
Despite the significant increase in Bdnf1 mRNA levels after DNMT inhibition, the expression of Bdnf4 and Bdnf9 was unaltered. It is particularly surprising that Bdnf4 is not affected by the DNMT inhibitor, as the promoter of the Bdnf4 exon is known to be susceptible to epigenetic changes (Kotera et al., 2004) and has an eminent role in the processes of learning and memory in rodents (Lubin, 2011). A possible explanation for the above observation could be related to the differential involvement of BDNF transcripts in distinct memory processes of different cognitive tasks. For example, in a study in which the contextual fear conditioning paradigm was used, it was shown that exposure to a novel context leads to upregulation of Bdnf1 in the hippocampus within 2 h, while consolidation of associative memory was accompanied by elevation in expression of Bdnf4 (Lubin, Roth, & Sweatt, 2008). Interestingly, a different study in which they used the object recognition memory task showed that short-term recognition memory is positively correlated with increased methylation of Bdnf1 in the hippocampus of the rats, though BDNF protein levels in the hippocampus were decreased. Of note, an opposite effect was observed for BDNF in the perirhinal cortex (Muñoz, Aspé, Contreras, & Palacios, 2010). Therefore, the memory process and task specific requirement of different BDNF transcripts in brain structures could explain the differential results regarding expression of the BDNF splice variants after treatment. Although in our study we did not measure protein levels after drug administration, it would be interesting to check if our increases in Bdnf1 mRNA levels could subsequently result in decreased or increased BDNF synthesis in different brain areas. Similar to Bdnf4, acute administration of RG108 did not alter gene expression of Gria1 and Hdac2, despite their relevance for the task. Gria1 belongs to the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic libraries receptor (AMPAR) family that is known to be crucial for LTP and the strengthening of synapses that is important for learning and memory (Lisman, Yasuda, & Raghavachari, 2012). For that, rapid trafficking of GluA1- containing AMPARs at the postsynaptic membrane is required. The trafficking may involve delivery of already existing AMPARs or synthesis of new AMPARs (Ju et al., 2004, Nayak et al., 1998, Oh et al., 2006, Penn et al., 2017). Nevertheless, the timing of these dynamic changes is difficult to be predicted. Therefore, it is possible that 1 h after treatment represents a time-point that does not involve transcription of GluA1-AMPARs, implying it to be too early or too late in this respect. Regarding Hdac2, its expression levels were determined because it works in close concert with DNMTs. Opposed to methylation, acetylation normally leads to an increase in gene-transcription, by opening the chromatin and making it more accessible to transcription factors. The strong interplay between DNMT and HDAC2 makes it interesting to assess whether DNMT inhibition also induces Hdac2 expression changes (Feng et al., 2007, Tsankova et al., 2007). In contrast to Hdac1 that is mainly expressed in astrocytes, Hdac2 is mainly found in neurons and HDAC2 has a crucial role in normal brain development and cell survival (Hagelkruys et al., 2014). HDAC2 however is only linked to one of the two mechanism of methylation-induced gene silencing via binding in the MBDs and could therefore remain unchanged even if methylation changes (Feng et al., 2007). In order to extend our Bdnf1 findings, we performed DNA methylation analysis in 14 CpG sites in the promoter I of Bdnf and observed an increase in the methylation status of 3 of them. Specifically, for the treatment that had an intermediate effect (0.1 mg/kg) in OPS performance, there was an increase at the CpG 2, 3 and 6 in promoter I, while for the most effective treatment (0.3 mg/kg) the increased methylation was restricted in CpG2. The methylation status of the remaining CpG islands tested was unaltered between the different experimental groups. Considering that the animals were treated with a DNMT inhibitor, the observed increased methylation was not anticipated. A paradoxical effect in the action of DNMTs was also observed in a study showing that stress induced an increase in DNA methylation levels in the hippocampus and a decrease in methylation levels in the prefrontal cortex. Administration of RG108 was able to compensate for these changes, suggesting that both increase and decrease in stress-induced DNA methylation could be regulated by DNMT activity (Sales & Joca, 2016).