The DNA methylation is one of the typical epigenetic changes, which occurs at cysteine bases, converting which to 5′-methylcytosine. Cytosine methylation occurs in cytosine–phosphate–guanine (CpG) dinucleotides within the 5′-regulatory regions of the genes. This process is catalyzed by the DNA methyltransferases (DNMTs) family, including DNMT-1, DNMT-3A, and DNMT-3B (Levenson et al., 2006; Chestnut et al., 2011). These subgroups show the obvious difference in their expression distribution and functions. The DNMT-3A/B can be thought as the de novo methyltransferases, which transfer the methyl to the unmodified DNA, while the DNMT-1 maintains the DNA methylation status (Mateen et al., 2017). The DNA methylation takes part in the regulation of chromatin structure and the transcription rate of genes. Generally, it is considered to turn the chromatin into the closed, highly compacted state restrictive to the transcription, resulting in the gene silencing. Interestingly, when exploring the DNA methylation changes in the experience-induced events including neuronal activity and learning and memory process, the researchers find these changes are the gene-specific and bidirectional, with some genes exhibiting increased DNA methylation while others decreasing (Miller and Sweatt, 2007; Guo et al., 2011). It is considered that the balance between the DNA methylation of memory activators and memory suppressors is more important (Zovkic et al., 2013).

During the past decade, some studies demonstrated the role of DNA methylation in the neurotoxicity induced by the neonatal GA (Ju et al., 2016; Wu et al., 2016; Fan et al., 2021). It has been reported that the neonatal sevoflurane exposure increased the mRNA and protein levels of DNMT-3A and DNMT-3B, which upregulated the methylation level of brain-derived neurotrophic factor (BDNF) and Reelin promoter regions. These methylation changes induced the decrease of the plasticity-related proteins, BDNF, and Reelin, which led to the synaptic plasticity deficiency and cognitive impairments as a result (Ju et al., 2016). Another recent study focused on the DNA methylation in the isoflurane induced neurotoxicity (Wu et al., 2016). On postnatal 65 days, it is said that it is the increased expression of DNMT-1 and its occupancy on the promoter region of Bdnf exon IV in the hippocampal CA1 region of the rats that increase the methylation of 5′-cytosine in the promoter region of Bdnf exon IV, so the mRNA and protein levels of BDNF witnessed an obvious decrease in the rats with neonatal isoflurane exposure. To compare these two with each other, interestingly, the different anesthetics can induce the DNA hypermethylation of different genes through different DNMT subgroups at different time. Ju et al. (2016) used the DNMTs inhibitor 5-aza-2-deoxycytidine (5-AZA) to successfully reverse this sevoflurane induced abnormalities and Wu et al. (2016) chose the environmental enrichment to reverse, both of which reveal that this kind of epigenetic regulation can be chose as a target to prevent or treat the anesthesia induced neurotoxicity. To sum up, the neonatal GA exposure can lead to DNA hypermethylation of some of genes through the change of DNMT and then induce the neurotoxicity.

Histone acetylation is the most-widely explored and best-understood epigenetic modification, which is the addition of acetyl groups to the lysine residues on N-terminal histone tails catalyzed by histone acetyltransferases (HATs), which keep balance with histone deacetylases (HDACs). The HDACs in mammals are divided into four subfamilies, classes I, II, III, and IV. Classes I, IIa, IIb, and IV belong to the zinc-dependent HDACs and class III belongs to the NAD⁺-dependent HDACs named sirtuins (Berndsen and Denu, 2008). With the addition of negative charge on the acetyl group, histone acetylation decreases the affinity of histones to DNA and turns the chromatin into the open state amenable to translation. Through the regulation of target genes such as Bdnf, many recent studies demonstrated that histone acetylation takes part in the formation of long-term memory and conversely, the memory loss (Levenson et al., 2004).

A number of studies reported that the neonatal GA results in the decrease of histone acetylation, inducing the cognitive dysfunction as a result, which can put the blame on the imbalance of HDACs and HATs, the increase of HDACs or the decrease of HATs, either the quantity or the activity. Wu et al. (2016) reported a significant increase in the expression of HDAC2 along with an increased occupancy of HDAC2 on the promoter region of Bdnf exon IV in hippocampus of rats neonatally exposed to isoflurane, which as a result, revealed a decreased histone H3 acetylation in the promoter region of Bdnf exon IV as well. Zhong et al. (2015) reported that the neurocognitive dysfunction induced by repeated exposure to isoflurane in neonatal mice is related to the dysregulation of hippocampal histone acetylation which only affected the specific residue H4K12. To compare with these two studies, we can speculate that the histone acetylation modification has species difference. Jia et al. (2016) explored the neurotoxicity induced by sevoflurane exposure in neonatal rats. In this study, researchers reported that the elevated levels of HDAC3 and HDAC8 but no significant change in the levels of HDAC1 and HDAC2 in the hippocampus of rats neonatally exposed to sevoflurane, and they also found the decreased levels of acetyl-H3K9, acetyl-H3K14, acetyl-H4, acetyl-H4K5, and acetyl-H4K12 in the hippocampal CA1 region of the sevoflurane-exposed rats. The data upon were all tested on the postnatal 33 days, 1 h after the fear conditioning test. When explored the GA effects on the maternal-offspring module, the researchers indicated that infusion with propofol in maternal rats on gestational day-7 leads to impairment of learning and memory in off-spring, increased the protein levels of HDAC2, decreased the levels of acetylated H3K14 and H4K12 (Lin et al., 2018b). The different results of the different anesthetics suggest that the histone acetylation may have the drug specificity.

Apart from the histone hypoacetylation regulated by the increase of HDACs, some studies also reported the other way which decreases the acetylation of histones that plays an important role in the neonatal anesthesia induced cognitive dysfunction. The cAMP response element-binding (CREB) protein is a transcription factor that regulates the expression of genes related with learning and memory process. Phosphorylation of CREB is recognized as active state which is a marker of memory processing in the hippocampus for spatial learning (Mizuno et al., 2002). It was reported that propofol anesthesia during pregnancy resulted in decrease in phosphor-CREB protein in offspring rats’ hippocampus (Lin et al., 2018b). In the hippocampus of rats 24 h after exposing to a sedative dose of midazolam followed by combined nitrous oxide and isoflurane anesthesia for 6 h at postnatal 6 days, the researchers noticed the fragmentation of the CREB-binding protein (CBP), which has the HAT activity that transfers the acetyl group on the lysine residues of histone tails (Dalla Massara et al., 2016). The fragmentation of CBP led to the significant decrease of its HAT enzyme activity, the obvious reduction of the expression of acetylated H3 at c-Fos and Bdnf promoter and CREB transcription start sites and further, the decrease in the mRNA and proteins of c-Fos and BDNF in the hippocampus. This post-translational regulation points out a novel way to explore the mechanism of GA induced histone acetylation changes. Also, they show us a new method to prevent or treat the neonatal GA-induced neurotoxicity. In some of the studies, the researchers have successfully attenuated the GA-induced dysregulation of histone acetylation in the hippocampus and the neurocognitive impairment through the treatment of the HDACs inhibitor trichostatin A (TSA) and sodium butyrate (Zhong et al., 2015; Jia et al., 2016). However, no chemical interventions have been used to increase the HAT enzyme activity so to reverse the GA-induced hypoacetylation of histones or ameliorate the cognitive dysfunction, and maybe this will be explored further and testified in the near future.

Histone methylation is another covalent post-translational modification but have a different positive or negative effects on gene transcription depending on the residue modified and the number of methyl groups that added (Berger, 2007). Generally, dimethylation of H3K9 correlates with transcriptional silencing and trimethylation of H3K4 is linked to active transcription. Recent studies focused on the function of histone methylation in the process of long-term memory storage. The researchers reported both trimethylation of H3K4 and dimethylation of H3K9 in hippocampus are in the process of memory formation, suggesting both increases in gene expression and gene repression exist during this process (Gupta et al., 2010). Nowadays, the role of histone methylation in the neonatal GA exposure-induced neurotoxicity is largely unknown. However, a recent study explored its association with impaired cognition in middle-aged mice experiencing anesthesia and surgery. Wu et al. (2019) indicated that the expression of H3K9me3 and its binding to the Bdnf exon IV promoter increased significantly in mice 24 h after the surgery under GA, and also the expression of BDNF was down-regulated. Although the module of this study was post-operation middle-aged mice, considering the bidirectional regulation of this modification and the lack of deep investigation, this mechanism gives us a novel insight into the neonatal anesthesia neurotoxicity as well.

One class of the ncRNA that has been widely studied in recent years is miRNAs which are endogenous approximately 22-nucleotide-long RNAs that play as post-transcriptional gene regulatory molecules in multicellular organisms (Bartel, 2004). Recent advanced studies uncovered numerous molecular mechanisms of miRNA, among which, translation inhibition or degradation of target messenger RNAs (mRNAs) is the most widely and mainly explored mechanism (Correia de Sousa et al., 2019). Considering the regulatory function of miRNAs, it has been clearly demonstrated that miRNAs have a crucial impact on cognitive function and take part in the development of several neuropsychiatric disorders, including schizophrenia, intellectual disability, and autism spectrum disorders (Xu et al., 2012).

In the recent years, many studies have focused on the roles of this post-transcriptional modification in the neonatal GA-induced prolong cognitive dysfunction (Figure 1). Lin et al. (2018a) carried out a comprehensive unbiased profile assay that revealed the effects of neonatal sevoflurane exposure on miRNA expression of the brain. The data shows that the neonatal sevoflurane had significant impact on the expression of specific miRNAs, and this GA-induced changes have region and time specificity. The hippocampus miRNAs have a region-specific expression pattern compared to the whole brain. On the same day of sevoflurane exposure, the differentially expressed miRNAs in the whole brain were miR-136, miR-145, miR-19a, miR-29a, miR-300, and miR-361, in the hippocampus were miR-129-3p, miR-15a, miR-20a + 20b, miR-382, miR-532-5p, and miR-99a, which showed a transient change after anesthesia. However, when analyzed the miRNA expression at adult to explore the long-lasting impact, some difference occurred. At adult, among the tissue from the whole brain, the expression of miR-145, miR-29a, and miR-300 showed the significant difference, and in the hippocampus, the changes were observed in miR-20a, miR-20b, miR-532, and miR-99a.

Some other articles have reported single change of certain miRNAs and changes of their target gene transcription induced by neonatal GA exposure. It has been demonstrated that rats inhaled sevoflurane at postnatal-seven-day exhibited notable increase expression of miR-96 which targets and negatively regulates insulin-like growth factor-1 receptor (IGF-1R), a receptor tyrosine kinase (Xu et al., 2019). The IGF-1 is a kind of neuroprotective hormone and plays an important role in learning and memory process (Deak and Sonntag, 2012). The downregulation of IGF-1 contributes cognitive dysfunction and memory impairment in aged-mice after sevoflurane anesthesia (Jiang et al., 2017). So, under the consideration of IGF1, Xu et al. (2019) also reported the neuronal apoptosis and cognitive dysfunction through this mechanism. Lv et al. (2017) found the increased expression levels of miR-27a-3p, the downregulation of PPAR-γ, a target of miR-27a-3p, and also the neuron injury in vitro after treatment of 4% sevoflurane. When inhibiting miR-27a-3p in vivo, the sevoflurane-induced neuronal apoptosis and cognitive dysfunction were reversed, which demonstrated the important role of miR-27a-3p.

Apart from the inhaled anesthetics, some scientists explored the neonatal neurotoxicity of intravenous anesthetics. Zhang et al. (2017) found that the repeated propofol anesthesia resulted in a downregulation or upregulation of miR-132 levels at different time points and the changes of mRNA and protein level of p250GAP, a prominent target for miR-132. This change of miRNA-132 resulted in a decrease in the number of dendritic spines in the hippocampus leading to learning and memory impairment of rats. In the study of Xu et al. (2015), they revealed the in vitro results that miR-124 was upregulated by ketamine in hippocampal organotypic culture. To investigate the function of the change of miR-124, they did the in vivo experiments by using the lentiviral vectors containing miR-124 antisense oligonucleotide to inhibit miR-124. The results showed that miR-124 inhibition strengthened glutamate receptors and upregulated PKC pathway after ketamine-exposure at postnatal 14 days, which reversed the hippocampal neurodegeneration and improved memory loss. Generally, many miRNAs have been demonstrated to take part in the GA-induced neurotoxicity through whether profile analysis or the focus on some certain miRNAs, and this modification can also be interfered by virus.

The lncRNAs are another class of ncRNAs, recent studies have identified many paradigms that illustrate how lncRNAs function, including targeting proteins to specific genomic location to affect transcription patterns, modulating the activity of protein-binding partners and serving as precursors for small RNAs (Wilusz et al., 2009). A recent study in 2018 uncovered the effects of lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) on sevoflurane-induced neurotoxicity in developing rats (Hu et al., 2019). It demonstrated that rats anesthetized with sevoflurane show the high expression of MALAT1, the downregulation of BDNF and NGF mRNA and protein levels, the increase of neuron apoptosis and pathological structural changes in hippocampus and the impairment of spatial learning and memory ability. However, the downregulation of MALAT1 through MALAT1 small interfering RNA alleviated these changes. Although the relative studies are not so numerous, this kind of regulation has shown the relationship with apoptosis, deformation and memory dysfunction which are known as possible mechanism of neonatal neurotoxicity, the prospect of ncRNAs in the field of GA-induced neurotoxicity is wide and broaden.

The RNA methylation, as the most prevalent modification of RNA, has been widely explored in the past decade. The RNAs exist in some modified bases that can fine-tune the translation of target genes. In the past, it was thought that tRNA and rRNA have these bases and intensive researches have poured into them, while recently it was discovered that also mRNAs have modified bases, including m⁶A, m⁵C, and m¹A, where m6A is the most-widely explored modification of bases (Squires et al., 2012; Fu et al., 2014; Dominissini et al., 2016). The m⁶A RNA methylation has been described as a dynamic and reverse modification which executes its function through a series of enzymes, including methyltransferases that help the addition of methyl groups, called writers, demethylases that help removal, called erasers and m⁶A binding proteins that help recognition and function, called readers. Considering the characteristic upon, it is also seen as a member of post-transcriptional epigenetic modifications. The m⁶A has been demonstrated to take part in the regulation of mRNAs, including splicing, localization and translation (Wang et al., 2014, 2015; Zhao et al., 2014). Considering this function, m⁶A also regulates many cellular processes and affects the memory formation and cortical neurogenesis (Yoon et al., 2017; Leonetti et al., 2020).

Nowadays, some researchers tried to focus on the role of RNA methylation modification in neonatal GA induced neurotoxicity. Zhang et al. (2021) reported that multiple sevoflurane anesthesia exposures may induce the downregulation of YTHDF1 in the prefrontal cortex of mice which is a critical reader protein of m⁶A RNA methylation reported to promote synthesis of target transcripts in the hippocampus of adult mice and thereby facilitate learning and memory (Shi et al., 2018). As a result, this change of YTHDF1 led to the decrease of m⁶A RNA methylation and protein level of the presynaptic protein synaptophysin and cognitive deficits (Zhang et al., 2021). After this experiment, to further investigate the possible function and effect of m⁶A modification on this sevoflurane-related neurotoxicity, these researchers published an article this year that revealed the changes in the m⁶A methylation profile of the prefrontal cortex of infant rhesus macaques (Chen et al., 2022). Compared with the control group, a dramatic increase in the number of m⁶A marks was observed in the sevoflurane group, and the hypermethylated m⁶A sites were enriched in the regulation of synaptic plasticity which may have impact on cortical development and learning and memory ability. Although the research that investigate the role of m⁶A methylation in neonatal GA-induced neurotoxicity is not that many, the articles before have offered some clues for the future exploration. The transient or prolonged changes of certain genes in the anesthesia process may be deeply assessed and can also be chosen as the targets for prevention or treatment of the neurotoxicity. Apart from the side effects of anesthetics for neonates, He et al. reported the role of m⁶A RNA methylation in POCD. They demonstrated that the weakened METTL3 (the methyltransferase works as “writer”) affinity contributed to the disorder of m⁶A RNA methylation in the hippocampus of POCD mice (He and Wang, 2021). Unlike regulations occurred on the “reader” level, the changes in POCD were showed to be as a result of “writer”, maybe the difference can attribute to the different model with different age point or some other deep reasons needed to be explored further.

Unlike two sides of the same coin, the interactions of these three epigenetic modifications are close and comprehensive that still need long time to understand (Mateen et al., 2017). When the methyl groups were added on the region of DNA promoters, they could recruit histone-modifying enzymes, induce deacetylation of histone protein tails and condense the chromatin. This series of events reveal that the methylation of DNA results in the deacetylation of histones which works as a compounded inhibitory effect on transcription (Siegfried and Simon, 2010; Baylin and Jones, 2011).

In contrast, the altered histone methylation can change DNA methylation. When researchers explored the function of epigenetic modification in the process of memory formation and consolidation, they reported that the increased H3K4 trimethylation can be associated with either increased DNA methylation (Zif268) (Gupta et al., 2010) or decreased DNA methylation (Bdnf) (Lubin et al., 2008) at genes that are transcriptionally activated in memory process.

This kind of bidirectional interactions were also reported exist in the GA-induced neurotoxicity. Wu et al. (2016) have indicated the bridge molecular MeCP2 that binds to the methylated CpG sites in the promoter region and form the MeCP2-DNMT1 complex to maintain DNA methylation, at the meanwhile, MeCP2 recruits HDAC2 to remove the acetyl group from the histones (Kimura and Shiota, 2003), thereby leading to the more repressive transcription of target gene Bdnf, and this mechanism is involved in the isoflurane induced cognitive dysfunction. Also, some of miRNAs regulate DNA methylation through downregulating the expression of DNMT (Chen et al., 2014) and some can regulate the histone modifications through the transcription factor CREB. Gao et al. (2010) illustrated that the increase expression of miR-134 following SIRT1 (belongs to HDAC III family) deficiency results in the downregulation of CREB and BDNF, thereby impairing synaptic plasticity. It has been reported that the neonatal GA can lead to the decreased expression of SIRT1 and then resulted in the cognitive dysfunction (Tang et al., 2020, 2021), so whether the downstream of SIRT1, miRNAs take part in this GA induced change is worthy to be explored further.