Total DNA Methylation Changes Reflect Random Oxidative DNA Damage in Gliomas

DNA modifications can be used to monitor pathological processes. We have previously shown that estimating the amount of the main DNA epigenetic mark, 5-methylcytosine (m5C), is an efficient and reliable way to diagnose brain tumors, hypertension, and other diseases. Abnormal increases of reactive oxygen species (ROS) are a driving factor for mutations that lead to changes in m5C levels and cancer evolution. 8-oxo-deoxyguanosine (8-oxo-dG) is a specific marker of ROS-driven DNA-damage, and its accumulation makes m5C a hotspot for mutations. It is unknown how m5C and 8-oxo-dG correlate with the malignancy of gliomas. We analyzed the total contents of m5C and 8-oxo-dG in DNA from tumor tissue and peripheral blood samples from brain glioma patients. We found an opposite relationship in the amounts of m5C and 8-oxo-dG, which correlated with glioma grade in the way that low level of m5C and high level of 8-oxo-dG indicated increased glioma malignancy grade. Our results could be directly applied to patient monitoring and treatment protocols for gliomas, as well as bolster previous findings, suggesting that spontaneously generated ROS react with m5C. Because of the similar mechanisms of m5C and guanosine oxidation, we concluded that 8-oxo-dG could also predict glioma malignancy grade and global DNA demethylation in cancer cells.


Introduction
Gliomas are the largest and most diverse group of primary brain tumors [1]. They are subdivided according to their histological and molecular features [2]. Diffuse gliomas, the most abundant subgroup, show broad infiltration in the surrounding central nervous system (CNS) parenchyma. They present the malignancy range from WHO grade II (low grade) to IV (glioblastoma, high grade). Other glial tumors, like pilocytic astrocytomas (WHO grade I), have a more circumscribed growth pattern. Glioblastoma (GBM) is the most frequent diffuse glioma in adults and is the deadliest form with a median patient survival of about one year. GBM cells have the predilection for aggressive invasion into the normal brain tissue and infiltration along the vascular tracts, which limits the possibility of complete tumor resection and reduces the effect of localized radiotherapy. Defects in multiple intracellular signaling pathways are involved in the gliomas' pathogenesis and lead to the evasion of cell death, which constitutes a hallmark of cancer [3].
Cellular damage accelerates the advancement of degenerative processes, aging, apoptosis, as well as carcinogenesis [4,5]. Reactive oxygen species (ROS) are the main damaging agent of various cell components, including DNA [4]. Under physiological conditions, there is a balance between ROS

Analysis of m 5 C Contents in DNA
DNA hydrolysis, labeling, and TLC chromatography were performed as previously described [19][20][21][22][23]. The radioactive intensity of spots corresponding to m 5 C, C (cytosine), and T (thymine) was used for calculations. The data obtained with the post labeling and TLC methods were compared to the ELISA assay [26]. . Analysis of dG for reference was performed at 260 nm. 8-oxo-dG was determined with the following electrochemical detection settings: guard cell +400 mV, detector 1: +130 mV (screening electrode), detector 2: +350 mV (measuring electrode set on the 100 nA sensitivity).

Statistical Analysis
Unless otherwise specified, the results were expressed as the mean ± SD (Standard Deviation) in the error bars, and the data were collected from three independent analyses. Statistical significance was determined using the STATISTICA software, 1998 edition (StatSoft Poland, 1995-2008, as in our previous studies [28], by ANOVA test, and a p-value <0.05 was considered as significant.

Patients' Characteristics
The analyzed cohort consisted of 36 individuals diagnosed with brain glioma, aged from 19 to 81 years (Table S1 and S2). The median patient's age at the time of tumor surgery was 53.1 ± 13.3 years. There were 21 (58.3%) males and 15 (41.7%) females. The tumors were mainly primary (i.e., the samples were taken before any other oncological treatment), but the cohort included some cases of recurrent gliomas (those undergone previously first surgery, radiotherapy, and chemotherapy) (Table S1 and S2). The control group consisted of 34 generally healthy persons, aged from 18 to 66 years (mean 43.1 ± 14.5 years), with 15 (44.1%) males and 19 (55.9%) females (Table S3).

Total Contents of m 5 C and 8-oxo-dG in DNA from Tumor Tissue Samples
For 28 patients, the total level of m 5 C and 8-oxo-dG in genomic DNA from brain tumor tissue was analyzed (Table S1). There were variations in the contents of both compounds in the tumor histological subgroups, most clearly seen for glioblastoma group, reflecting the heterogeneity of those tumors (Table S1, Figure 2).

Statistical Analysis
Unless otherwise specified, the results were expressed as the mean ± SD (Standard Deviation) in the error bars, and the data were collected from three independent analyses. Statistical significance was determined using the STATISTICA software, 1998 edition (StatSoft Poland, 1995-2008, as in our previous studies [28], by ANOVA test, and a p-value <0.05 was considered as significant.

Patients' Characteristics
The analyzed cohort consisted of 36 individuals diagnosed with brain glioma, aged from 19 to 81 years (Table S1 and S2). The median patient's age at the time of tumor surgery was 53.1 ± 13.3 years. There were 21 (58.3%) males and 15 (41.7%) females. The tumors were mainly primary (i.e., the samples were taken before any other oncological treatment), but the cohort included some cases of recurrent gliomas (those undergone previously first surgery, radiotherapy, and chemotherapy) (Table S1 and S2). The control group consisted of 34 generally healthy persons, aged from 18 to 66 years (mean 43.1 ± 14.5 years), with 15 (44.1%) males and 19 (55.9%) females (Table S3).

Total Contents of m 5 C and 8-oxo-dG in DNA from Tumor Tissue Samples
For 28 patients, the total level of m 5 C and 8-oxo-dG in genomic DNA from brain tumor tissue was analyzed (Table S1). There were variations in the contents of both compounds in the tumor histological subgroups, most clearly seen for glioblastoma group, reflecting the heterogeneity of those tumors (Table S1, Figure 2).
We observed a significant stepwise increase of 8-oxo-dG contents in DNA from brain glioma tissue with increasing tumor malignancy. Glioblastoma (WHO IV) had 12-121 times higher amount of 8-oxo-dG than pilocytic astrocytoma (WHO I) (Table S1, Figure 2A). Low levels of total DNA methylation were concomitant with high levels of 8-oxo-dG and in higher grade gliomas. The m 5 C amount in pilocytic astrocytoma (WHO I) was 2-7 times higher than glioblastoma (WHO IV) (Table S1, Figure 2B). The number of m 5 C residues in relation to the number of 8-oxo-dG residues was ca. 1000:1 in low grade (WHO I-II) gliomas, and dropped significantly to ca. 100:1 or less in high grade (WHO III-IV) tumors, indicating higher DNA damage (Table S2). We observed a significant stepwise increase of 8-oxo-dG contents in DNA from brain glioma tissue with increasing tumor malignancy. Glioblastoma (WHO IV) had 12-121 times higher amount of 8-oxo-dG than pilocytic astrocytoma (WHO I) (Table S1, Figure 2A). Low levels of total DNA methylation w The mean amounts of 8-oxo-dG were increasing with increasing tumor grade, while total DNA methylation was decreasing ( Figure 3). The differences between the subgroups comprising tumors of the same grade were statistically significant. The ANOVA test for 8-oxo-dG showed F= 4.873 and p = 0.009; only for WHO groups III and IV, F = 10.346, p = 0.004; whereas for m 5 C, F = 107.194 and p < 0.005. The mean amounts of 8-oxo-dG were increasing with increasing tumor grade, while total DNA methylation was decreasing ( Figure 3). The differences between the subgroups comprising tumors of the same grade were statistically significant. The ANOVA test for 8-oxo-dG showed F= 4.873 and p = 0.009; only for WHO groups III and IV, F = 10.346, p = 0.004; whereas for m 5 C, F = 107.194 and p < 0.005.  The application of an exponential one-phase decay function for m 5 C in relation to 8-oxo-dG showed an extremely high correlation coefficient of 0.9804 ( Figure 4). A clear trend was observed that a small increase in 8-oxo-dG contents led to great depletion in m 5 C amount. One phase decay curve of m 5 C in relation to 8-oxo-dG contents in DNA from human brain tumors tissues. 8-oxo-dG allows subdividing gliomas within a similar total DNA methylation level. The application of an exponential one-phase decay function for m 5 C in relation to 8-oxo-dG showed an extremely high correlation coefficient of 0.9804 ( Figure 4). A clear trend was observed that a small increase in 8-oxo-dG contents led to great depletion in m 5 C amount.  The application of an exponential one-phase decay function for m 5 C in relation to 8-oxo-dG showed an extremely high correlation coefficient of 0.9804 ( Figure 4). A clear trend was observed that a small increase in 8-oxo-dG contents led to great depletion in m 5 C amount.

Total DNA Methylation and 8-oxo-dG Contents in Peripheral Blood Samples
For eight glioma patients, we analyzed the total level of m 5 C and 8-oxo-dG in genomic DNA from peripheral blood samples (Table S2). The group included seven glioblastoma (WHO IV) patients along with one anaplastic astrocytoma (WHO III) for comparison. The total DNA methylation levels were almost equal in that group (Table S2, Figure 5). For eight glioma patients, we analyzed the total level of m 5 C and 8-oxo-dG in genomic DNA from peripheral blood samples (Table S2). The group included seven glioblastoma (WHO IV) patients along with one anaplastic astrocytoma (WHO III) for comparison. The total DNA methylation levels were almost equal in that group (Table S2, Figure 5). The relationship of m 5 C and 8-oxo-dG contents observed for peripheral blood samples was similar to those from glioma tissues. The amount of m 5 C was 1.7-2.3 times lower and 8-oxo-dG 11.4-16.9 times higher in WHO IV samples when compared to the WHO III sample.
The differences between m 5 C and 8-oxo-dG contents for blood samples in glioma patients versus the control group were statistically significant. In the control group of generally healthy individuals, the mean m 5 C contents in DNA was 314 ± 31 × 10 5 (Figure 6), so significantly higher than in the brain glioma WHO III-IV group (Table S2, Figure 5). It has been already reported that the amount of 8-oxo-dG in normal human cells' DNA is of 1 base per 10 7 guanosine residues [29]. The relation of the 8-oxo-dG contents with subjects' older age showed clear increase:, 4.63 ± 1.11 (for mean age 13), 3.59 ± 1.60 (for mean age 31), 4.82 ± 1.62 (for mean age 50), and 5.79 ± 2.22 (for mean age 67) per 10 6 guanosine residues [30]. We calculated 8-oxo-dG amounts per whole genome (DNA). The 8-oxo-dG contents in peripheral blood DNA of glioma patients were substantially higher. The one-way ANOVA test values were F = 309.347 and p < 0.005 for total DNA methylation in glioma and control group, and F = 49.966 and p < 0.005 for 8-oxo-dG contents between glioma group and the literature data [29]. The relationship of m 5 C and 8-oxo-dG contents observed for peripheral blood samples was similar to those from glioma tissues. The amount of m 5 C was 1.7-2.3 times lower and 8-oxo-dG 11.4-16.9 times higher in WHO IV samples when compared to the WHO III sample.
The differences between m 5 C and 8-oxo-dG contents for blood samples in glioma patients versus the control group were statistically significant. In the control group of generally healthy individuals, the mean m 5 C contents in DNA was 314 ± 31 × 10 5 (Figure 6), so significantly higher than in the brain glioma WHO III-IV group (Table S2, Figure 5). It has been already reported that the amount of 8-oxo-dG in normal human cells' DNA is of 1 base per 10 7 guanosine residues [29]. The relation of the 8-oxo-dG contents with subjects' older age showed clear increase: 4.63 ± 1.11 (for mean age 13), 3.59 ± 1.60 (for mean age 31), 4.82 ± 1.62 (for mean age 50), and 5.79 ± 2.22 (for mean age 67) per 10 6 guanosine residues [30]. We calculated 8-oxo-dG amounts per whole genome (DNA). The 8-oxo-dG contents in peripheral blood DNA of glioma patients were substantially higher. The one-way ANOVA test values were F = 309.347 and p < 0.005 for total DNA methylation in glioma and control group, and F = 49.966 and p < 0.005 for 8-oxo-dG contents between glioma group and the literature data [29].

Discussion
Currently, there is a need for better complex disease characterization on the molecular level that can assist clinicians in identification and optimal treatment of brain gliomas. Molecular markers turned out to be a powerful aid for clinical diagnostics, precise treatment implementation, and estimating clinical outcomes [31]. DNA hypomethylation is an exceptionally frequent change linked to numerous neoplasms [32,33]. Carcinogenesis is also linked to the overproduction of reactive oxygen species, including superoxide radical, hydroxyl radical and hydrogen peroxide, as well as a shortage of antioxidants [34]. ROS levels in cancer cells are typically higher than physiologic because of increased metabolic activity, cellular signaling, oncogene activity, and mitochondrial dysfunction [35,36].
The brain is especially susceptible to the damaging effects of ROS because of its high metabolic activity, oxygen consumption, and low capacity for cellular regeneration [5]. The impact of ROS, as well as reactive nitrogen species (RNS), on gliomagenesis have been described, mainly in signal transduction and metabolomics aspects [37,38].
We previously showed that there's a clear correlation between total DNA methylation and brain tumor malignancy, in tumor tissues as well as in peripheral blood samples [19,21,23]. We also The amounts of m 5 C were significantly higher than in glioma patients.

Discussion
Currently, there is a need for better complex disease characterization on the molecular level that can assist clinicians in identification and optimal treatment of brain gliomas. Molecular markers turned out to be a powerful aid for clinical diagnostics, precise treatment implementation, and estimating clinical outcomes [31]. DNA hypomethylation is an exceptionally frequent change linked to numerous neoplasms [32,33]. Carcinogenesis is also linked to the overproduction of reactive oxygen species, including superoxide radical, hydroxyl radical and hydrogen peroxide, as well as a shortage of antioxidants [34]. ROS levels in cancer cells are typically higher than physiologic because of increased metabolic activity, cellular signaling, oncogene activity, and mitochondrial dysfunction [35,36]. The brain is especially susceptible to the damaging effects of ROS because of its high metabolic activity, oxygen consumption, and low capacity for cellular regeneration [5]. The impact of ROS, as well as reactive nitrogen species (RNS), on gliomagenesis have been described, mainly in signal transduction and metabolomics aspects [37,38].
We previously showed that there's a clear correlation between total DNA methylation and brain tumor malignancy, in tumor tissues as well as in peripheral blood samples [19,21,23]. We also identified single papers that estimated 8-oxo-dG in glioma tissues, noticing the relation of higher oxidative stress with tumor malignancy [39][40][41]. In the present study, we focused on the comparison between m 5 C, a DNA epigenetic marker, and 8-oxo-dG, which is an established well-known oxidative damage marker, in glioma tissues and peripheral blood samples. Both nucleosides frequently occur in m 5 CpG dinucleotides [42]. One previous study on genome-wide methylation profiling in pediatric glioblastoma showed a possible association of ROS with pediatric GBM. Gene Ontology analysis performed there showed that the pediatric GBM-specific methylome identified the "superoxide metabolic process" and the "oxygen and ROS metabolic process", indicating ROS action as a driving phenomenon in those tumors [43]. What is unique in our present findings is that for the first time we've shown a direct correlation between the oxidative DNA damage (monitored by 8-oxo-dG) and epigenetic regulation (through total m 5 C), and their relation with tumor malignancy. Moreover, 8-oxo-dG allowed subdividing gliomas within a similar total DNA methylation level (Figure 4). That is a step further than nowadays existing classification systems [1], and holds a potential for better patient stratification in terms of treatment planning and prognosis.
The •OH radical diffusion rate properties and ability to react with DNA is confined to short distances, limited to 2 nm in cells and tissues. Therefore, it acts seemingly at generation sites [44]. The Fenton reaction is the most common source of •OH in the cell [45]. It involves reduced redox-active metal ions, as ferrous and cuprous, that react with metabolically produced hydrogen peroxide (H 2 O 2 ). There is usually the site-specific output of •OH from Fenton-like reactions, for example, including metal ions close to or bound to DNA [44].
Guanine (G) is the most susceptible to oxidation (lowest redox potential, 1.29 mV) out of the four DNA bases [42]. The 8-oxo-dG redox potential is even lower, 0.74 mV, which results in its further oxidation [46]. That property is explored for electrochemical analyses of this compound at a picomol level [47]. To determine the extent of oxidative stress to which a cell has been exposed, the cellular level of 8-oxo-dG in the genome can be estimated and regarded as a status of oxidative stress leading to mutations [48]. The frequency for 8-oxo-dG to cause de novo spontaneous and heritable G→T transversion mutations in mice was estimated at the range of~1% [49]. There is a steady-state level of~30,000 8-oxo-dGs (mouse embryonic stem cell) that theoretically generates~300 mutation events during replication [50]. m 5 C deamination can also lead to C→T transition [51]. In mouse embryonic stem cells there are~50 million m 5 C sites that could result in~1000 mutations [50]. Therefore, it can be estimated that m 5 C leads to~3-fold more mutations than 8-oxo-dG. That corresponds to deep genome sequencing studies that identified more C→T than G→T mutations [52]. Moreover, the presence of 8-oxo-dG negatively affects adjacent sites of DNA methylation and suppresses methylation of nascent DNA strand one or two base pairs away from the damaged guanine [53,54]. On the other hand, DNA methylation may influence the mutagenic effect of ROS [55].
The demarcation between genetic (mutations) and epigenetic DNA modifications is blurred. The oxidative damage to m 5 C is extremely important for the formation of epimutations [56], as well as for its function as an epigenetic regulator [15,16,32]. Therefore, it should lead to a reevaluation of modifications that are considered mutagenic.
A parallel analysis of m 5 C (epigenetic mark) and 8-oxo-dG (oxidation marker) provides a challenge of the biological significance of an enzymatic m 5 C hydroxylation mechanism leading to DNA demethylation in brain tumors. Data are showing that Ten-eleven translocation (TET) proteins are responsible for m 5 C oxidation and, in fact, demethylation [57]. They cause hydroxylation of the m 5 C methyl group forming 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC). TET-mediated m 5 C oxidation is considered as an active DNA demethylation mechanism for genome stability. DNA demethylation sites are concomitant with base excision repair. 8-Oxoguanine glycosylase (OGG1) binds and recruits TET1 to the 8-oxo-G adjacent m 5 C sites to prevent mutations [58]. Tet2 is responsible for marking sites of DNA-damage with 5hmC [59]. Cancer genomes are characterized by focal DNA hypermethylation, co-existing with widespread hypomethylation. However, recent data show, that even in the case of TET loss of function, general loss of DNA methylation occurs [60]. That raises the question about the exact mechanism of oxidative DNA modifications.
It is generally accepted that 8-oxo-dG formation is the result of the attack of •OH on carbon 8 (C8) of guanosine, which is the most reactive site easily available to •OH in the major groove of DNA (Figure 7). function, general loss of DNA methylation occurs [60]. That raises the question about the exact mechanism of oxidative DNA modifications. It is generally accepted that 8-oxo-dG formation is the result of the attack of •OH on carbon 8 (C8) of guanosine, which is the most reactive site easily available to •OH in the major groove of DNA (Figure 7). The question arises whether m 5 C, a neighbor nucleoside in m 5 CpG dinucleotide, is reacting with •OH (Figure 7). The hydrophobic methyl group of m 5 C protrudes into the major groove of DNA double helix and is prone to •OH attack. A direct oxidation product of m 5 C is 5hmC [62]. Moreover, oxidation of m 5 C induces final loss of a methyl group in the form of formaldehyde or formic acid [63]. Further support for that observation is a relatively low ratio (3)(4)(5) of the methyl group modification in comparison to guanosine modification. It is a result of a small efficiency of hydrogen abstraction from a methyl group and formation of C-CH2•. Our data provide strong support for ROS-induced demethylation of DNA in glioma patients' samples. Furthermore, it was demonstrated that ROS-mediated DNA oxidation that yields 8-oxo-dG in gene promoters is a signal for gene activation. The presence of 8-oxo-dG in a G-quadruplex-forming promoter sequence resulted in ca. 300% in gene expression increase [64]. A comparison of 8-oxo-dG yield and m 5 C loss in human brain tumors of different malignancy strongly suggests that both nucleosides are modified by ROS, particularly by •OH, during oxidative stress conditions.
There is no doubt that elevated ROS rates promote tumor development and progression [36]. However, the relations of DNA methylation changes with cancer are less obvious. Cancer-linked DNA hypermethylation includes promoters of tumor suppressor genes resulting in repression of gene expression, thereby facilitating cancer formation. Global DNA hypomethylation during the neoplastic process doesn't have such straightforward relation but is observed in many tumors, and results in global gene expression deregulation [65]. It is also observed during cellular senescence [20]. It is generally accepted that the antineoplastic effect of chemotherapeutic drugs is due to the induction of oxidative stress and ROS-mediated cell injury [66]. In the same way, DNA demethylation agents are used [67]. But is the way of complete oxidative damage of the cell the right way of oncological treatment? The cells that survive such therapy are highly resistant to any treatment and produce aggressive recurrences. Moreover, significant harm is done to normal organism's cells [66]. Based on the results of our study, it may be presumed that the opposite actions should be promoted and DNA hypermethylating strategies should be concerned. Antioxidants' intake is a well-known prophylactic and treatment supporting method. They generally increase the response rate to chemotherapy, help restore proper redox balance of the organism, and some even impact patients' survival [68]. The examples are also known for brain gliomas [69]. Besides adding them to the therapeutic regimen, one can explore the ROS-scavenging properties of already used The question arises whether m 5 C, a neighbor nucleoside in m 5 CpG dinucleotide, is reacting with •OH (Figure 7). The hydrophobic methyl group of m 5 C protrudes into the major groove of DNA double helix and is prone to •OH attack. A direct oxidation product of m 5 C is 5hmC [62]. Moreover, oxidation of m 5 C induces final loss of a methyl group in the form of formaldehyde or formic acid [63]. Further support for that observation is a relatively low ratio (3)(4)(5) of the methyl group modification in comparison to guanosine modification. It is a result of a small efficiency of hydrogen abstraction from a methyl group and formation of C-CH 2 •. Our data provide strong support for ROS-induced demethylation of DNA in glioma patients' samples. Furthermore, it was demonstrated that ROS-mediated DNA oxidation that yields 8-oxo-dG in gene promoters is a signal for gene activation. The presence of 8-oxo-dG in a G-quadruplex-forming promoter sequence resulted in ca. 300% in gene expression increase [64]. A comparison of 8-oxo-dG yield and m 5 C loss in human brain tumors of different malignancy strongly suggests that both nucleosides are modified by ROS, particularly by •OH, during oxidative stress conditions.
There is no doubt that elevated ROS rates promote tumor development and progression [36]. However, the relations of DNA methylation changes with cancer are less obvious. Cancer-linked DNA hypermethylation includes promoters of tumor suppressor genes resulting in repression of gene expression, thereby facilitating cancer formation. Global DNA hypomethylation during the neoplastic process doesn't have such straightforward relation but is observed in many tumors, and results in global gene expression deregulation [65]. It is also observed during cellular senescence [20]. It is generally accepted that the antineoplastic effect of chemotherapeutic drugs is due to the induction of oxidative stress and ROS-mediated cell injury [66]. In the same way, DNA demethylation agents are used [67]. But is the way of complete oxidative damage of the cell the right way of oncological treatment? The cells that survive such therapy are highly resistant to any treatment and produce aggressive recurrences. Moreover, significant harm is done to normal organism's cells [66]. Based on the results of our study, it may be presumed that the opposite actions should be promoted and DNA hypermethylating strategies should be concerned. Antioxidants' intake is a well-known prophylactic and treatment supporting method. They generally increase the response rate to chemotherapy, help restore proper redox balance of the organism, and some even impact patients' survival [68]. The examples are also known for brain gliomas [69]. Besides adding them to the therapeutic regimen, one can explore the ROS-scavenging properties of already used drugs. In the case of brain gliomas, mannitol, the widely used medication against brain edema, presents such effects [70]. That detail can impact the extension of its use in therapy.

Conclusions
We showed that brain gliomas were characterized by strong oxidative stress, which resulted in global DNA damage. Two different types of DNA modifications, mainly occurring randomly, were characterized. 8-oxo-dG levels were indicative for increased oxidative stress, DNA damage, and therefore higher tumor malignancy. In parallel, DNA demethylation was observed, resulting in pathological DNA gene expression.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4409/8/9/1065/s1. Table S1. The characteristics of 28 patients for whom the contents of m 5 C (post labeling method) and 8-oxo-dG (electrochemical detector) in DNA from tumor tissue were analyzed, as well as patients' clinical characteristics. Table S2. The characteristics of 8 patients with WHO III and IV brain gliomas for whom the contents of m 5 C (post labeling method) and 8-oxo-dG (electrochemical detector) in DNA from peripheral blood samples were analyzed, as well as patients' clinical characteristics. Table S3. The characteristics of 34 healthy individuals comprising the control group for whom the contents of m 5 C (postlabeling method) in DNA from peripheral blood samples was analysed.

Conflicts of Interest:
The authors declare no conflict of interest.