Loss of 5-hydroxymethylcytosine is linked to gene body hypermethylation in kidney cancer

Both 5-methylcytosine (5mC) and its oxidized form 5-hydroxymethylcytosine (5hmC) have been proposed to be involved in tumorigenesis. Because the readout of the broadly used 5mC mapping method, bisulfite sequencing (BS-seq), is the sum of 5mC and 5hmC levels, the 5mC/5hmC patterns and relationship of these two modifications remain poorly understood. By profiling real 5mC (BS-seq corrected by Tet-assisted BS-seq, TAB-seq) and 5hmC (TAB-seq) levels simultaneously at single-nucleotide resolution, we here demonstrate that there is no global loss of 5mC in kidney tumors compared with matched normal tissues. Conversely, 5hmC was globally lost in virtually all kidney tumor tissues. The 5hmC level in tumor tissues is an independent prognostic marker for kidney cancer, with lower levels of 5hmC associated with shorter overall survival. Furthermore, we demonstrated that loss of 5hmC is linked to hypermethylation in tumors compared with matched normal tissues, particularly in gene body regions. Strikingly, gene body hypermethylation was significantly associated with silencing of the tumor-related genes. Downregulation of IDH1 was identified as a mechanism underlying 5hmC loss in kidney cancer. Restoring 5hmC levels attenuated the invasion capacity of tumor cells and suppressed tumor growth in a xenograft model. Collectively, our results demonstrate that loss of 5hmC is both a prognostic marker and an oncogenic event in kidney cancer by remodeling the DNA methylation pattern.     

Decreased 5-hydroxymethylcytosine is associated with neural progenitor phenotype in normal brain and shorter survival in malignant glioma

Epigenetic modification of DNA by cytosine methylation to produce 5-methylcytosine (5mC) has become well-recognized as an important epigenetic process in human health and disease. Recently, further modification of 5mC by the ten eleven translocated (TET) family of enzymes to produce 5-hydroxymethylcytosine (5hmC) has been described. In the present study, we used immunohistochemistry to evaluate the distribution of 5hmC in human brain during different periods of development and in a large series of gliomas (n=225). We found that during development, 5hmC levels are high in more differentiated compartments like the fetal cortex, but low in the periventricular progenitor cell regions. In adults, we found 5hmC levels to be highest in the cortex, but present in all intrinsic cell types in the brain including stromal elements. In brain tumors, 5hmC levels were high in low grade tumors and reduced in malignant glioma, but did not exhibit any correlation with IDH1 mutation status. Additionally, we identified a significant relationship between low levels of 5hmC and reduced survival in malignant glioma. This observation was further supported by in silico analysis showing differential expression of genes involved in 5hmC homeostasis in aggressive subsets of glioblastoma. Finally, we show that several genes involved in regulating the levels of 5hmC are also prognostic in malignant glioma. These findings suggest that 5hmC regulation in malignant glioma may represent an important determinant of tumor differentiation and aggressive behavior, as well as a potential therapeutic target.     

Mouse olfactory bulb methylome and hydroxymethylome maps reveal noncanonical active turnover of DNA methylation

Hydroxylation of 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC) by TET enzymes presents a particular regulatory mechanism in the mammalian brain. However, although methylation and hydroxymethylation of cytosines in non-CpG contexts have been reported, these mechanisms remain poorly understood. Here, we applied TAB-seq and oxBS-seq selectively to detect 5hmC and 5mC at base resolution in olfactory bulb derived from female mice. We found that active turnover of 5mC to 5hmC occurred in both CpG and non-CpG contexts. Strikingly, we identified a different sequence preference for 5mC and 5hmC in a CH context, in which H = A, C, or T, TNCA/TC for 5mC and NNCA/T/CN for 5hmC. More importantly, we found that genes showing 5mC to 5hmC turnover showed only limited overlap in CpG and CH contexts, and that olfactory receptor genes were marked with higher turnover of 5mC to 5hmC in non-CpG context. Collectively, we identified an unexpected sequence preference for non-CpG hydroxymethylation and distinct target genes regulated by the turnover of 5mC to 5hmC in CpG and CH contexts.     

Accurate Measurement of 5-Methylcytosine and 5-Hydroxymethylcytosine in Human Cerebellum DNA by Oxidative Bisulfite on an Array (OxBS-Array)

The Infinium 450K Methylation array is an established tool for measuring methylation. However, the bisulfite (BS) reaction commonly used with the 450K array cannot distinguish between 5-methylcytosine (5mC) and 5-hydroxymethylcytosine (5hmC). The oxidative-bisulfite assay disambiguates 5mC and 5hmC. We describe the use of oxBS in conjunction with the 450K array (oxBS-array) to analyse 5hmC/5mC in cerebellum DNA. The “methylation” level derived by the BS reaction is the combined level of 5mC and 5hmC at a given base, while the oxBS reaction gives the level of 5mC alone. The level of 5hmC is derived by subtracting the oxBS level from the BS level. Here we present an analysis method that distinguishes genuine positive levels of 5hmC at levels as low as 3%. We performed four replicates of the same sample of cerebellum and found a high level of reproducibility (average r for BS = 98.3, and average r for oxBS = 96.8). In total, 114,734 probes showed a significant positive measurement for 5hmC. The range at which we were able to distinguish 5hmC occupancy was between 3% and 42%. In order to investigate the effects of multiple replicates on 5hmC detection we also simulated fewer replicates and found that decreasing the number of replicates to two reduced the number of positive probes identified by > 50%. We validated our results using qPCR in conjunction with glucosylation of 5hmC sites followed by MspI digestion and we found good concordance with the array estimates (r = 0.94). This experiment provides a map of 5hmC in the cerebellum and a robust dataset for use as a standard in future 5hmC analyses. We also provide a novel method for validating the presence of 5hmC at low levels, and highlight some of the pitfalls associated with measuring 5hmC and 5mC.     

The role of 5-hydroxymethylcytosine in human cancer

The patterns of DNA methylation in human cancer cells are highly abnormal and often involve the acquisition of DNA hypermethylation at hundreds or thousands of CpG islands that are usually unmethylated in normal tissues. The recent discovery of 5-hydroxymethylcytosine (5hmC) as an enzymatic oxidation product of 5-methylcytosine (5mC) has led to models and experimental data in which the hypermethylation and 5mC oxidation pathways seem to be connected. Key discoveries in this setting include the findings that several genes coding for proteins involved in the 5mC oxidation reaction are mutated in human tumors, and that a broad loss of 5hmC occurs across many types of cancer. In this review, we will summarize current knowledge and discuss models of the potential roles of 5hmC in human cancer biology.     

Age-related epigenome-wide DNA methylation and hydroxymethylation in longitudinal mouse blood

DNA methylation at cytosine-phosphate-guanine (CpG) dinucleotides changes as a function of age in humans and animal models, a process that may contribute to chronic disease development. Recent studies have investigated the role of an oxidized form of DNA methylation – 5-hydroxymethylcytosine (5hmC) – in the epigenome, but its contribution to age-related DNA methylation remains unclear. We tested the hypothesis that 5hmC changes with age, but in a direction opposite to 5-methylcytosine (5mC), potentially playing a distinct role in aging. To characterize epigenetic aging, genome-wide 5mC and 5hmC were measured in longitudinal blood samples (2, 4, and 10 months of age) from isogenic mice using two sequencing methods – enhanced reduced representation bisulfite sequencing and hydroxymethylated DNA immunoprecipitation sequencing. Examining the epigenome by age, we identified 28,196 unique differentially methylated CpGs (DMCs) and 8,613 differentially hydroxymethylated regions (DHMRs). Mouse blood showed a general pattern of epigenome-wide hypermethylation and hypo-hydroxymethylation with age. Comparing age-related DMCs and DHMRs, 1,854 annotated genes showed both differential 5mC and 5hmC, including one gene – Nfic – at five CpGs in the same 250 bp chromosomal region. At this region, 5mC and 5hmC levels both decreased with age. Reflecting these age-related epigenetic changes, Nfic RNA expression in blood decreased with age, suggesting that age-related regulation of this gene may be driven by 5hmC, not canonical DNA methylation. Combined, our genome-wide results show age-related differential 5mC and 5hmC, as well as some evidence that changes in 5hmC may drive age-related DNA methylation and gene expression.     

Cytosine 5-Hydroxymethylation of the LZTS1 Gene Is Reduced in Breast Cancer

Change of DNA cytosine methylation (5mC) is an early event in the development of cancer, and the recent discovery of a 5-hydroxymethylated form (5hmC) of cytosine suggests a regulatory epigenetic role that might be different from 5-methylcytosine. Here, we aimed at elucidating the role of 5hmC in breast cancer. To interrogate the 5hmC levels of the leucine zipper, putative tumor suppressor 1 (LZTS1) gene in detail, we analyzed 75 primary breast cancer tissue samples from initial diagnosis and 12 normal breast tissue samples derived from healthy persons. Samples were subjected to 5hmC glucosyltransferase treatment followed by restriction digestion and segment-specific amplification of 11 polymerase chain reaction products. Nine of the 11 5′LZTS1 fragments showed significantly lower (fold change of 1.61–6.01, P < .05) 5hmC content in primary breast cancer tissue compared to normal breast tissue samples. No significant differences were observed for 5mC DNA methylation. Furthermore, both LZTS1 and TET1 mRNA expressions were significantly reduced in tumor samples (n = 75, P < .001, Student''s t test), which correlated significantly with 5hmC levels in samples. 5hmC levels in breast cancer tissues were associated with unfavorable histopathologic parameters such as lymph node involvement (P < .05, Student''s t test). A decrease of 5hmC levels of LZTS1, a classic tumor suppressor gene known to influence metastasis in breast cancer progression, is correlated to down-regulation of LZTS1 mRNA expression in breast cancer and might epigenetically enhance carcinogenesis. The study provides support for the novel hypothesis that suggests a strong influence of 5hmC on mRNA expression. Finally, one may also consider 5hmC as a new biomarker.     

Genome-wide regulation of 5hmC, 5mC, and gene expression by Tet1 hydroxylase in mouse embryonic stem cells

DNA methylation at the 5 position of cytosine (5mC) in the mammalian genome is a key epigenetic event critical for various cellular processes. The ten-eleven translocation (Tet) family of 5mC-hydroxylases, which convert 5mC to 5-hydroxymethylcytosine (5hmC), offers a way for dynamic regulation of DNA methylation. Here we report that Tet1 binds to unmodified C or 5mC- or 5hmC-modified CpG-rich DNA through its CXXC domain. Genome-wide mapping of Tet1 and 5hmC reveals mechanisms by which Tet1 controls 5hmC and 5mC levels in mouse embryonic stem cells (mESCs). We also uncover a comprehensive gene network influenced by Tet1. Collectively, our data suggest that Tet1 controls DNA methylation both by binding to CpG-rich regions to prevent unwanted DNA methyltransferase activity, and by converting 5mC to 5hmC through hydroxylase activity. This Tet1-mediated antagonism of CpG methylation imparts differential maintenance of DNA methylation status at Tet1 targets, ultimately contributing to mESC differentiation and the onset of embryonic development.     

Examination of the specificity of DNA methylation profiling techniques towards 5-methylcytosine and 5-hydroxymethylcytosine

DNA cytosine-5 methylation is a well-studied epigenetic pathway implicated in gene expression control and disease pathogenesis. Different technologies have been developed to examine the distribution of 5-methylcytosine (5mC) in specific sequences of the genome. Recently, substantial amounts of 5-hydroxymethylcytosine (5hmC), most likely derived from enzymatic oxidation of 5mC by TET1, have been detected in certain mammalian tissues. Here, we have examined the ability of several commonly used DNA methylation profiling methods to distinguish between 5mC and 5hmC. We show that techniques based on sodium bisulfite treatment of DNA are incapable of distinguishing between the two modified bases. In contrast, techniques based on immunoprecipitation with anti-5mC antibody (methylated DNA immunoprecipitation, MeDIP) or those based on proteins that bind to methylated CpG sequences (e.g. methylated-CpG island recovery assay, MIRA) do not detect 5hmC and are specific for 5mC unless both modified bases occur in the same DNA fragment. We also report that several methyl-CpG binding proteins including MBD1, MBD2 and MBD4 do not bind to sequences containing 5hmC. Selective mapping of 5hmC will require the development of unique tools for the detection of this modified base.     

Epigenetics and Cell Death: DNA Hypermethylation in Programmed Retinal Cell Death

Background

Vertebrate genomes undergo epigenetic reprogramming during development and disease. Emerging evidence suggests that DNA methylation plays a key role in cell fate determination in the retina. Despite extensive studies of the programmed cell death that occurs during retinal development and degeneration, little is known about how DNA methylation might regulate neuronal cell death in the retina.

Methods

The developing chicken retina and the rd1 and rhodopsin-GFP mouse models of retinal degeneration were used to investigate programmed cell death during retinal development and degeneration. Changes in DNA methylation were determined by immunohistochemistry using antibodies against 5-methylcytosine (5mC) and 5-hydroxymethylcytosine (5hmC).

Results

Punctate patterns of hypermethylation paralleled patterns of caspase3-dependent apoptotic cell death previously reported to occur during development in the chicken retina. Degenerating rd1 mouse retinas, at time points corresponding to the peak of rod cell death, showed elevated signals for 5mC and 5hmC in photoreceptors throughout the retina, with the most intense staining observed in the peripheral retina. Hypermethylation of photoreceptors in rd1 mice was associated with TUNEL and PAR staining and appeared to be cCaspase3-independent. After peak rod degeneration, during the period of cone death, occasional hypermethylation was observed in the outer nuclear layer.

Conclusion

The finding that cell-specific increases of 5mC and 5hmC immunostaining are associated with the death of retinal neurons during both development and degeneration suggests that changes in DNA methylation may play a role in modulating gene expression during the process of retinal degeneration. During retinal development, hypermethylation of retinal neurons associates with classical caspase-dependent apoptosis as well as caspase-3 independent cell death, while hypermethylation in the rd1 mouse photoreceptors is primarily associated with caspase-3 independent programmed cell death. These findings suggest a previously unrecognized role for epigenetic mechanisms in the onset and/or progression of programed cell death in the retina.     

Tissue distribution of 5-hydroxymethylcytosine and search for active demethylation intermediates

5-Hydroxymethylcytosine (hmC) was recently detected as the sixth base in mammalian tissue at so far controversial levels. The function of the modified base is currently unknown, but it is certain that the base is generated from 5-methylcytosine (mC). This fuels the hypothesis that it represents an intermediate of an active demethylation process, which could involve further oxidation of the hydroxymethyl group to a formyl or carboxyl group followed by either deformylation or decarboxylation. Here, we use an ultra-sensitive and accurate isotope based LC-MS method to precisely determine the levels of hmC in various mouse tissues and we searched for 5-formylcytosine (fC), 5-carboxylcytosine (caC), and 5-hydroxymethyluracil (hmU) as putative active demethylation intermediates. Our data suggest that an active oxidative mC demethylation pathway is unlikely to occur. Additionally, we show using HPLC-MS analysis and immunohistochemistry that hmC is present in all tissues and cell types with highest concentrations in neuronal cells of the CNS.     

Altering TET dioxygenase levels within physiological range affects DNA methylation dynamics of HEK293 cells

The TET family of dioxygenases (TET1/2/3) can convert 5-methylcytosine (5mC) into 5-hydroxymethylcytosine (5hmC) and has been shown to be involved in active and passive DNA demethylation. Here, we demonstrate that altering TET dioxygenase levels within physiological range can affect DNA methylation dynamics of HEK293 cells. Overexpression of TET1 increased global 5hmC levels and was accompanied by mild DNA demethylation of promoters, gene bodies and CpG islands. Conversely, the simultaneous knockdown of TET1, TET2, and TET3 led to decreased global 5hmC levels and mild DNA hypermethylation of above-mentioned regions. The methylation changes observed in the overexpression and knockdown studies were mostly non-reciprocal and occurred with different preference depending on endogenous methylation and gene expression levels. Single-nucleotide 5hmC profiling performed on a genome-wide scale revealed that TET1 overexpression induced 5mC oxidation without a distribution bias among genetic elements and structures. Detailed analysis showed that this oxidation was related to endogenous 5hmC levels. In addition, our results support the notion that the effects of TET1 overexpression on gene expression are generally unrelated to its catalytic activity.