Background: Active demethylation of 5-methyl-2′-deoxycytidine (5-mdC) in DNA occurs by oxidation to 5- (hydroxymethyl)-2′-deoxycytidine (5-hmdC) and further oxidation to 5-formyl-2′-deoxycytidine (5-fdC) and 5- carboxy-2′-deoxycytidine (5-cadC), and is carried out by enzymes of the ten-eleven translocation family (TETs 1, 2, 3). Decreased level of epigenetic DNA modifications in cancer tissue may be a consequence of reduced activity/expression of TET proteins. To determine the role of epigenetic DNA modifications in colon cancer development, we analyzed their levels in normal colon and various colonic pathologies. Moreover, we determined the expressions of TETs at mRNA and protein level. The study included material from patients with inflammatory bowel disease (IBD), benign polyps (AD), and colorectal cancer (CRC). The levels of epigenetic DNA modifications and 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) in examined tissues were determined by means of isotope-dilution automated online two-dimensional ultraperformance liquid chromatography with tandem mass spectrometry (2D-UPLC-MS/MS). The expressions of TET mRNA were measured with RT-qPCR, and the expressions of TET proteins were determined immunohistochemically. Results: IBD was characterized by the highest level of 8-oxodG among all analyzed tissues, as well as by a decrease in 5-hmdC and 5-mdC levels (at a midrange between normal colon and CRC). AD had the lowest levels of 5-hmdC and 5-mdC of all examined tissues and showed an increase in 8-oxodG and 5-(hydroxymethyl)-2′- deoxyuridine (5-hmdU) levels. CRC was characterized by lower levels of 5-hmdC and 5-mdC, the lowest level of 5-fdC among all analyzed tissues, and relatively high content of 5-cadC. The expression of TET1 mRNA in CRC and AD was significantly weaker than in IBD and normal colon. Furthermore, CRC and AD showed significantly lower levels of TET2 and AID mRNA than normal colonic tissue. (Continued on next page) * Correspondence: firstname.lastname@example.org; email@example.com Department of Clinical Biochemistry, Faculty of Pharmacy, Collegium Medicum in Bydgoszcz, Nicolaus Copernicus University in Torun, Torun, Poland Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 2 of 11 (Continued from previous page) Conclusions: Our findings suggest that a complex relationship between aberrant pattern of DNA epigenetic modification and cancer development does not depend solely on the transcriptional status of TET proteins, but also on the characteristics of premalignant/malignant cells. This study showed for the first time that the examined colonic pathologies had their unique epigenetic marks, distinguishing them from each other, as well as from normal colonic tissue. A decrease in 5-fdC level may be a characteristic feature of largely undifferentiated cancer cells. Keywords: DNA epigenetic modification, Ten-eleven translocation protein, Colon cancer, Inflammatory bowel disease, Adenoma, Demethylation, Background dynamic molecular process, being a form of response to During the last decade, one of the hot topics in oncogen- inflammation-related environmental/metabolic changes. esis was the so-called cancer epigenome, having implica- The downstream steps of active demethylation process tions for cancer promotion and progression. This, in turn, may be, at least partially, responsible for the loss of is linked with a plethora of abnormalities based on somatic 5-mdC. Furthermore, 5-fdC and 5-cadC were shown to heritable modifications that are not caused by alterations be recognized by a larger number of proteins than in primary sequence of DNA. 5-hmdC, despite markedly higher level of the latter [11, Methylation of cytosine, usually in CpG dinucleotides, is 12]. a key epigenetic modification exerting a profound impact In the vast majority of previous studies, 5-hmdC, on gene repression, cellular identity, and organismal fate 5-fdC, and 5-cadC were determined semi-quantitatively, . However, equally important is an opposite reaction, by means of immunohistochemistry; consequently, the DNA demethylation, resulting in activation of previously results of these might be biased due to ultra-low content silenced genes. Although a large body of evidence suggests of these modifications in genomic DNA of the tumor that active demethylation may occur in mammalian cells, . It should be also stressed that the accuracy of immu- its molecular background is still unclear (for review, see nohistochemical studies depends largely on the sensitiv- ). The most plausible mechanism behind the active de- ity/specificity of antibodies against a given modification. methylation of 5-methyl-2′-deoxycytidine (5-mdC) moiety In our present study, instead of using a semiquantitive in DNA involves ten-eleven translocation (TET) proteins, method with anti-5-hmdC antibodies, we determined which catalyze oxidation of 5-mdC to form 5-(hydroxy- 5-mdC, 5-hmdC, 5-fdC, 5-cadC, and 5-hmU with a methyl)-2′-deoxycytidine (5-hmdC) and further oxidation highly specific and highly sensitive method developed reactions that generate 5-formyl-2′-deoxycytidine (5-fdC) recently in our laboratory: isotope-dilution automated and 5-carboxy-2′-deoxycytidine (5-cadC) [2, 3]. Evidence online two-dimensional ultra-performance liquid from experimental studies supports the hypothesis that chromatography with tandem mass spectrometry TET enzymes may be also involved in the synthesis of (2D-UPLC-MS/MS) . 5-(hydroxymethyl)-2′-deoxyuridine (5-hmdU), a molecule To provide a better insight in the relationship between with epigenetic function . epigenetic DNA modifications and factors which may in- Several recent studies showed that the level of fluence formation thereof and to determine their role in 5-hmdC in many various types of human malignancies, CRC development, we analyzed their levels in normal including CRC, is profoundly reduced [5–7] and the de- colon and various colonic pathologies, which predispose gree of the reduction is proportional to tumor stage . to CRC development. Moreover, we determined the ex- Either the mechanism or the reason behind the de- pressions of TETs and AID at mRNA and protein level. crease in 5-hmdC level in cancer tissues is still not fully The study included samples from patients with CRC understood. Perhaps, this phenomenon reflects a de- (n = 97, both from the tumor and from normal colonic crease in the activity/expression of TET proteins . tissue), colon adenomas (AD, n = 39), and IBD (n = 49). However, it also cannot be excluded that the regulatory Since both CRC and chronic inflammation are associ- mechanisms of active DNA demethylation are deter- ated with oxidative stress, aside from the epigenetic mined by external conditions (e.g., chronic inflamma- DNA modifications, we also determined an established tion, oxidative stress, nutritional status), which results in marker of oxidatively modified DNA, 8-oxodG. The a release of different products. rationale of the study was to fill the gap in existing Chronic inflammation being a direct consequence of knowledge, explaining how conditions which predispose inflammatory bowel disease (IBD) is considered the most to CRC development can influence the synthesis of important etiological factor of sporadic colorectal malig- TET-mediated DNA modifications and oxidatively modi- nancies . Epigenetic modification of DNA is a fied DNA. Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 3 of 11 Methods model tissue section selected according to The Human Study group Protein Atlas (http://www.proteinatlas.org) and The study material originated from three groups of pa- antibody specification (Additional file 1: Table S1) and tients with (1) IBD (n = 49, median age 35 years, 53% of the antibody datasheet. Negative controls were prepared women), (2) AD, i.e., histologically confirmed adenoma from the examined tissues treated with 1% solution of tubulare (90%) or adenoma tubulovillosum (10%) (n =39, bovine serum albumin (BSA) in phosphate buffered sa- median age 65 years, 46% of women), and (3) CRC, i.e., line (PBS), instead of the primary antibody. Paraffin histologically confirmed stage A (8%), stage B (45%), stage TMA blocks and archived FFPE tissue sections were cut C (29%), or stage D (9%) adenocarcinoma, or malignant with a manual rotary microtome (AccuCut, Sakura, polyps (9%) (n = 97, median age 65 years, 46% of women). Torrance, USA) to obtain 4-μm slices, which were then None of the study subjects were related with one another, processed routinely and mounted on extra adhesive and all of them were Caucasians. All participants of the slides (SuperFrostPlus, MenzelGlasser, Braunschweig, study were recruited in a hospital setting (Collegium Germany). Medicum, Nicolaus Copernicus University, Bydgoszcz, The deparaffinization, rehydration, and antigen retrieval Poland) and subjected to colonoscopy. At the enrollment, were carried out in PT-Link system (Dako, Agilent all subjects completed a questionnaire containing infor- Technologies, USA). The slides were heated for 20 min in mation about their demographics, smoking, diet, and Epitope Retrieval Solution high-pH (95–98 °C; Dako, medical history. The study groups were matched for Agilent Technologies). Then, the activity of endogenous eating habits, age, body weight, and smoking status. No peroxidase was blocked by a 15-min incubation with 3% significant intergroup differences were found in terms of H O solution, and non-specific binding was eliminated by 2 2 body weight and body stature of male and female subjects. a 15-min incubation with 5% BSA solution; both reactions To make the study groups even more homogenous, the were carried out at room temperature. Subsequently, the subjects who reported overeating or use of dietary supple- slides were incubated with primary antibodies against ments during a month preceding the study were not TET1, TET2, and TET3 (specified in Additional file 1: included in the analysis. The questionnaire survey was Table S1). The antibody complexes were detected with En- conducted by the team physician (Dr. Banaszkiewicz, Dr. Vision Flex Anti-Mouse/Rabbit HRP-Labeled Polymer Klopocka). (Dako, Agilent Technologies) and localized using 3–3′di- aminobenzidine (DAB) as a chromogen. Finally, the slides Preparation of tissue microarrays (TMA) for were counterstained with hematoxylin, subsequently dehy- immunochemical analysis drated, cleared in series of xylenes, and coverslipped using Immunohistochemical studies were performed using ar- mounting medium (Dako, Agilent Technologies). chived formaldehyde-fixed paraffin-embedded (FFPE) tissue sections derived in the Department of Clinical Evaluation of protein expression based on Pathomorphology, Collegium Medicum in Bydgoszcz, immunohistochemical staining Nicolaus Copernicus University in Torun. Each slide was examined under ECLIPSE E400 light Hematoxylin and eosin (H&E)-stained microscopic microscope (Nikon Instruments Europe, Amsterdam, slides of archived FFPE tissue sections (donor blocks) Netherlands) with the low-power (20×) objective. The were used to identify representative tumor areas with at result of immunohistochemical staining was expressed least 80% tumor cells. Then, two such regions, each according to Immunoreactive Remmele-Stegner (IRS) 2 mm in diameter, were transferred from the donor score , described in detail in our previous papers blocks to a recipient TMA block using an automated tis- [14–16, 19]. Total IRC score (from 0 to 12) was obtained sue arrayer (TMA Master3D HISTECH, Budapest, by multiplying the staining intensity score (0—negative, Hungary). The same procedure was repeated for normal 1—weak staining, 2—moderate staining, 3—strong stain- tissue located at least 2 cm from the tumor resection ing) by the relative proportion of immunolabeled speci- margin. Then, another set of H&E-stained slides was men area (0—none; 1—less than 10%; 2—10 to 50%; 3— prepared to verify the accuracy of the TMA blocks. 50 to 80%; 4—at least 80%). TMA blocks were verified and double-checked by two independent pathologists. Extraction of DNA from tissues and its hydrolysis to deoxynucleosides Immunohistochemistry DNA from examined fresh frozen tissues was isolated as Immunohistochemical staining was carried out as described elsewhere , with some modifications. Iso- described elsewhere [14–16], and the results were stan- lated DNA was dissolved in 100 mM ammonium acetate dardized against a series of positive and negative (Sigma-Aldrich) containing 0.1 mM ZnCl (pH 4.3). The controls. Positive control staining was performed on a dissolved DNA samples (50 μl) were mixed with 1 U of Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 4 of 11 nuclease P1 (Sigma-Aldrich) and tetrahydrouridine by relative quantitative RT-PCR (RT-qPCR) with rele- (Calbiochem) (as cytidine deaminase inhibitor, 10 μg vant primers and short hydrolysis probes substituted per sample) and incubated at 37 °C for 1 h. Subse- with Locked Nucleic Acids from the Universal Probe quently, 12 μl5%(v/v)NH OH (JT Baker) and 1.3 U Library (UPL, Roche) (see: Additional file 3: Table S3). of alkaline phosphatase (Sigma-Aldrich) were added The probes were labeled with fluorescein (FAM) at the to each sample following 1-h incubation at 37 °C. 5′-end and with a dark quencher dye at the 3′-end. Ex- Finally, all DNA hydrolysates were acidified with pressions of target genes were normalized for two CH COOH (Sigma-Aldrich) (to final v/v concentra- selected reference genes, HMBS (GeneID: 3145) and tion of 2%) and ultrafiltered prior to injection. TBP (GeneID: 6908), using UPL Ready Assay #100092149 and #100092158, respectively. Real-time Isolation of DNA and determination of epigenetic PCR mixes (in 20 μl volumes) were prepared from modifications and 8-oxodG in DNA isolates cDNA following the standard procedures for LightCy- The methodology used to determine 5-methyl-2′-deoxy- cler480 Probes Master (Roche), provided with the re- cytidine (5-mdC), 5-hydroxymethyl-2′-deoxycytidine agent set. The reactions were carried out on 96-well (5-hmdC), 5-formyl-2′-deoxycytidine (5-fdC), 5-carboxy plates. Aside from the proper samples, each plate in- -2’deoxycytidine (5-cadC), 5-(hydroxymethyl)-2′-deox- cluded also no-template control and no-RT control. yuridine (5-hmdU), and 8-oxodG levels by means of Quantitative real-time PCR was carried out with Light- 2D-UPLC-MS/MS has been described elsewhere . Cycler 480 II, using the following cycling parameters: Transition patterns and specific detector settings for all 10 s at 95 °C, followed by 45 repeats 10 s each at 95 °C; analyzed compounds are presented in the Additional file 2: 30 s at 58 °C; and finally, 1 s at 72 °C with acquisition Table S2. mode (parameters of wavelength excitation and detec- tion equal 465 and 510 nm, respectively). The reaction Gene expression analysis for each gene was standardized against a standard curve, Isolated leukocytes were stored at − 80 °C until the ana- to estimate amplification efficiency. Standardization pro- lysis. RNA was isolated with MagNA Pure 2.0 (Roche) cedure included preparation of 10-fold serial dilutions following the standard procedures. Concentration and with controlled relative amount of targeted template. purity of RNA aliquots were verified spectrophotomet- The efficiency of amplification was assessed based on a rically with NanoDrop 2000 (Thermo Scientific). A / slope of the standard curve. Standard dilutions were A ratio was used as an indicator of protein contamin- amplified in separate wells, but within the same run. ation and A /A ratio as a measure of contamination Then, the samples were subjected to qPCR with meas- 260 230 with polysaccharides, phenol, and/or chaotropic salts. urement of C , and amplification efficiencies were auto- Quality and integrity of total RNA were assessed by matically calculated and displayed on the analysis visualization of 28S/18S/5.8S rRNA band pattern in a window of LightCycler 480 software, version 22.214.171.124 1.2% agarose gel. Non-denaturing electrophoresis was (Roche). The same software was also used for sample carried out at 95 V for 20 min in TBE buffer (Tris – setup, real-time PCR analysis, and calculation of relative Boric Acid – EDTA). The gel was stained with ethidium C values referred to as “ratios.” bromide or SimplySafe and visualized using GBox EF Gel Documentation System (SynGene). Purified RNA Statistical methods was stored at − 80 °C. The samples with RNA concentra- The results are presented as medians, interquartile ranges, tions greater than 50 ng/μl were qualified for further and non-outlier ranges. Normal distribution of the study analysis. 0.5 microgram of total RNA from each sample variables was verified with Kolmogorov-Smirnov test with (in 20-μl volume) was used for cDNA synthesis by Lilliefors correction and based on visual inspection of reverse transcription with High-Capacity cDNA Reverse plotted histograms. Variables with non-normal distribu- Transcription Kit (Applied Biosystems, catalog no. tions (5-hmdC, 5-fdC, 5-cadC, 5-hmdU, 8-oxodG concen- 43-688-14), according to the manufacturer’s instruction. tration and TETs, AID mRNA expression) were subjected The reaction was carried out with Mastercycler Nexus to Box-Cox transformation prior to statistical analyses Gradient thermocycler (Eppendorf). To exclude contam- with parametric tests. Normalized data were subjected to ination with genomic DNA, reverse transcriptase reac- one-way analysis of variance (ANOVA) followed by LSD tion included also a negative control. cDNA was either and Tukey post hoc tests. Associations between pairs of used for qPCR setup immediately after obtaining or variables were assessed based on Pearson correlation coef- stored at − 20 °C. The RT-qPCR complies with the Mini- ficients for raw or normalized data, where applicable. All mum Information for Publication of Quantitative statistical transformations and analyses were carried out Real-time PCR Experiments (MIQE) guidelines. Three with STATISTICA 13.1 PL [Dell Inc. (2016). Dell Statis- gene transcripts, TET1, TET2, and TET3, were analyzed tica (data analysis software system), version 13. Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 5 of 11 software.dell.com.]. The results were considered statisti- possible association between level of the epigenetic cally significant at P values lower than 0.05. modifications and tumor progression reflected in tumor stage from A to D. Significant decrease of Results 5-mdC, 5-hmdC, and 5-fdC was characteristic for Levels of epigenetic modifications and 8-oxodGuo in DNA early stage of CRC development (stage A), and no from tissue specimens further changes were observed along the disease The highest levels of 5-mdC and 5-hmdC were found progression. in normal colonic tissue, followed by IBD, AD, and The highest levels of 5-hmdU and 8-oxodG were CRC specimens (Fig. 1a, b); the level of 5-mdC in observed in IBD and AD and the lowest in CRC and AD turned out to be significantly lower than in other normal colonic tissue; also, these intergroup differences tissues. In turn, CRC specimens were characterized by were statistically significant (Fig. 1e, f). significantly lower levels of 5-fdC than other samples Furthermore, significant correlations were found in (Fig. 1c). The level of 5-cadC in AD was significantly the levels of 5-mdC, 5-cadC, 8-oxodG, and 5-hmdU (2- to 2.5-fold) lower than in other tissues; in turn, between CRC and normal colon (Fig. 2). the highest level of this modification was found in We have analyzed relationship/correlation between normal colonic tissue (Fig. 1d). We also analyzed age and 5-hmCyt (and other modifications). However in Fig. 1 Levels of DNA epigenetic modifications—5-mdC (a), 8-oxodG (b), 5-hmdC (c), 5-fdC (d), 5-cadC (e), and 5-hmdU (f) in normal colonic tissue (n = 90); inflammatory lesions, IBD (n = 49); polyps, AD (n = 39); and cancer tissue, CRC (n = 97). Marker in the center of the box represents median value. The length of each box (IQR, interquartile range) represents the range of values for 50% of the most typical observations, and its edges correspond to the first and third quartile. Whiskers represent variance outside the upper and lower quartile. P value was determined with Mann-Whitney U test Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 6 of 11 Fig. 2 Correlations between the levels of DNA epigenetic modifications, 5-mdC (a), 5-cadC (b), 8-oxodG (c), and 5-hmdU (d), in normal colonic tissue and cancer tissue CRC patients as well as in other groups, no such correl- understood, a major contributor to CRC development ation was found. seems to be aberrant methylation of DNA and oxida- tively damaged DNA. A growing body of evidence sug- Expression of TET and AID mRNA gests that decreased levels of 5-mdC (for review, see: The expression of TET1 in AD and CRC was significantly ) and 5-hmdC  may be found not only in human weaker than in normal colonic tissue and IBD (Fig. 3a). malignancies but also in their precursor lesions, such as The expression of TET2 in normal colonic tissue was sig- adenomas. This implies that the level of this modifica- nificantly weaker than in AD andCRC; moreover, a signifi- tion may decrease gradually throughout carcinogenesis. cant difference was found in TET2 expressions in IBD and However, it is still unclear if hypomethylation is a late or AD (Fig. 3b). The examined tissues did not differ signifi- early event in cancer development, and whether this cantly in terms of TET3 expressions (Fig. 3c). Irrespective process is directly involved in carcinogenesis (for review, of the examined tissue, the levels of AID mRNA were very see: ). Therefore, the aim of this study was to verify low or below the detection threshold. Nevertheless, the if CRC and its precursor lesions differ from normal levels of AID mRNA in CRC turned out to be significantly colonic tissue in the levels of DNA epigenetic lowerthaninother tissues(Fig. 3d). No statistically signifi- modifications. cant correlations were found between TETs expression and Similar to previous studies, we demonstrated that epigenetic DNA modifications (data not shown). 5-hmdC level in CRC was several times lower than in normal colonic tissue. However, the level of this modifi- Immunohistochemical analysis of protein expression cation in cancer precursor lesions still raises some con- Immunoreactivity to anti-TET1 antibodies in CRC troversies. In some studies, the levels of 5-hmdC in turned out to be significantly lower than in normal co- benign lesions were shown to be lower than in normal lonic tissue (Fig. 4). In the case of TET2, immunohisto- tissues [7, 22]. However, in a recent study, the results of chemical analysis was on the borderline of statistical which were published in Cell , melanocytes forming significance with p = 0.06. benign nevi showed relatively high levels of 5-hmdC, whereas a significant decrease or complete loss of this Discussion epigenetic mark was observed in melanoma cells. It Although a molecular link between adenomas, chronic should be stressed that in all these studies, 5-hmdC was inflammation, and carcinogenesis is still not completely determined with a less accurate semiquantitative Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 7 of 11 Fig. 3 Expressions of TET1 (a), TET2 (b), TET3 (c), and AID (d) mRNA in normal colonic tissue, cancer tissue (CRC, n = 49), inflammatory lesions (IBD, n =7), and polyps (AD, n = 14). Marker in the center of the box represents median value. The lengthofeachbox (IQR,interquartile range) represents the range of values for 50% of the most typical observations, and its edges correspond to the first and third quartile. Whiskers represent variance outside the upper and lower quartile. P value was determined with Mann-Whitney U test method. Our present study, involving highly accurate Previous semiquantitative studies with specific anti- quantitative technique for 5-hmdC determination, dem- bodies demonstrated elevated levels of 5-cadC in human onstrated that the level of this modification in AD and breast cancer and gliomas . CRC was essentially the same, approximately four times Recent evidence suggests that oncogenic transcription lower than in normal colonic tissue. These findings are factors, Myc and Max, and perhaps also an array of consistent with the results published by Uribe-Lewis et regulatory proteins, can specifically recognize 5-cadC, al. , who also showed that 5-hmdC levels in CRC having lesser affinity for 5-fdC and showing only a trace and adenoma were substantially lower than in normal of affinity towards 5-mdC and 5-hmdC. It should be colon. However, to the best of our knowledge, our remembered that dysregulation of MYC-MAX transcrip- present study was the first one to show that the level of tional network is a common mechanism driving progres- 5-hmdC in IBD was significantly lower than in normal sion of human malignancies . This may at least colonic tissue, at a midrange between the values found partially explain higher levels of 5-cadC found in cancer in this material and in CRC (Fig. 1b). tissue. Moreover, Xiong et al.  showed recently that Interestingly, a significant decrease in 5-fdC content Sall4, an oncogenic protein being overexpressed in colon was observed solely in CRC, and the level of this modifi- cancer , may cooperate with TET2, catalyzing oxida- cation in both types of precursor lesions, IBD and AD, tion of 5-hmdC and contributing to formation of was essentially the same as in normal colonic tissue 5-cadC. Another study demonstrated that TET3 may (Fig. 1c). Furthermore, both IBD and AD were charac- specifically bind to 5-cadC, initiating BER pathway and terized by significantly higher levels of 8-oxodG (the thus activating the process of demethylation . marker of oxidative stress) than CRC and normal colon Intriguingly, the analyses of associations between over- (Fig. 1f). It should be remembered that the induction of all survival and the levels of epigenetic modifications in oxidative stress in a cell culture was previously shown to CRC patients demonstrated that the only correlation of contribute to a decrease in 5-hmdC level . longer survival was low level of 5-cadC in marginal tis- The level of another higher-order oxidative modifica- sue (Fig. 5). Hence, an important question arises why tion of 5-mdC, i.e., 5-cadC, was significantly higher in this parameter is a meaningful predictor of longer sur- CRC than in AD or IBD (Fig. 1a, d). Furthermore, we vival in cancer patients? Although histopathological found a significant correlation between the levels of this examination did not demonstrate presence of cancer modification in CRC and normal colonic tissue (Fig. 2). cells in marginal/normal tissue, the molecular assays Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 8 of 11 in CRC and normal colonic tissue and a relatively high content of this modification in CRC, one can expect that marginal tissue with lower level of 5-cadC is less likely to contain cancer cells. Both CRC and its precursor lesions, especially AD, showed significantly lower levels of 5-mdC than normal colonic tissue (Fig. 1). A dramatic decrease in 5-mdC and 5-hmdC levels in AD may contribute to genomic instability and thus represent a decisive step in CRC de- velopment. Interestingly, a substantial decrease in the levels of these modifications in AD (observed also in IBD specimens) co-existed with an increase in 5-hmU, 8-oxodG, and 5-fdC content. This suggests that the decrease in 5-mdC level observed in cancer precursor le- sions may be associated with the recently proposed phenomenon of processive DNA demethylation. Plaus- ibly, 5-fdC, 5-hmd, and perhaps also 8-oxodG initiate processive demethylation of DNA, as proposed by Fran- chini et al. [31, 32]. In line with this hypothesis, an alternative pathway, the so-called processive DNA de- methylation, exists aside from the active process in- volved in local and specific demethylation of DNA. According to the authors of this hypothesis, a single ini- tiating event (such as a certain mismatch, e.g., 5-hmUra-G) may trigger processive demethylation of numerous 5-mdCs (and perhaps also 5-hmdCs) on the same locus via long-path BER, DNA mismatch repair (MMR), or nucleotide excision repair (NER) pathway. Recent experiments with cell-free extracts and circular heteroduplex DNA substrate demonstrated that 5-hmU may trigger the removal of distant epigenetic modifica- tions (5-mdC and 5-hmdC) on MMR- and long-path BER-dependent pathway . Our present study showed that the expression of TET1 mRNA in CRC and AD was significantly weaker than in IBD and normal colon (Fig. 3). Furthermore, CRC and AD showed significantly lower levels of TET2 and AID mRNA than normal colonic tissue. However, at a protein level, the only significant difference between the exam- ined tissues was found in the case of TET1, significantly Fig. 4 Expressions of TET1 (a), TET2 (b), and TET3 (c) protein in normal more abundant in normal colon than in CRC. colonic tissue and cancer tissue (CRC, n = 19). Marker in the center of the A main factor contributing to a decrease in the activity box represents median value. The length of each box (IQR, interquartile of TET proteins are mutations in catalytic domains of range) represents the range of values for 50% of the most typical these enzymes . Another reason behind the reduced observations, and its edges correspond to the first and third quartile. Whiskers represent variance outside the upper and lower quartile. P activity of TETs may be an inhibitory effect of accumu- value was determined with Mann-Whitney U test lated onco-metabolites, such as 2-OH-glutarate [35, 36], resulting primarily from the presence of IDH1/2 muta- detected the cells being clonally related to the tumor tions. However, these mutations were observed mainly (field cancerization) . Moreover, in the case of CRC in hematopoietic malignances and are rare or completely colon marginal tissue, field cancerization may involve up absent in solid tumors, such as CRC [37–39]. This im- to 10-cm patches . Therefore, the relatively large plies that a decrease in 5-hmdC level may be caused by specimen of marginal tissue used for DNA isolation other factors than TET/IDH mutations, for example oxi- likely corresponded to the area of field cancerization. dative stress. Indeed, recent evidence suggests that oxi- Considering a strong correlation between 5-cadC levels dative stress may contribute to post-translational Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 9 of 11 Fig. 5 Relation between the level of 5-cadC in marginal colon tissue and survival of CRC patients following surgery modulation of TET2 . Recently, it was demonstrated Conclusion that the inhibition of TET proteins may be a direct Our hereby presented findings suggest that a complex rela- consequence of hypoxia . Hypoxia is a common tionship between aberrant pattern of DNA epigenetic phenomenon in solid tumors, which may at least modification and cancer development does not depend partially explain a decrease in 5-hmdC and 5-fdC levels solely on the transcriptional status of TET proteins, but observed in CRC. Interestingly, hypoxia increases overall also on the characteristics of premalignant/malignant cells. oxidative stress and can change redox status of the This in turn implies that epigenetic modification of DNA is cell . linked to oxidative stress. However, the exact character of Since the shape of TET co-substrates (2-ketoglutarate, this complex relationship is still poorly understood. Fe ) depends on the redox state of the cell, the change Our findings are consistent with the results of previ- in the activity of these enzymes may reflect the severity ous studies, showing that aberrant methylation of DNA of oxidative stress. Furthermore, it cannot be excluded occurs at very early stages of CRC development. More- −2 that also superoxide (O ), an anion radical of dioxygen over, the hereby presented data add to existing evidence, and the precursor of free radicals, plays an important showing that a decrease in the level of epigenetic marks role in TET-mediated active DNA demethylation [43, is characteristic for early stages of CRC development, 44]. Thus, changes in the activity of TET proteins may and further progression of the tumor is not associated result from a persistent increase in the severity of oxida- with any additional changes in these parameters. To the tive stress, which promotes aberrant generation of DNA best of our knowledge, this study was the first one to epigenetic modifications during iterative oxidation of show that CRC, AD, and IBD had their unique epigen- 5-mdC. In this study, we found elevated levels of etic marks, distinguishing them from each other as well 8-oxodG in IBD and AD. Of note, level of 8-oxodG in as from normal colonic tissue: (i) IBD was characterized DNA may directly inform about oxidative stress in nu- by the highest level of 8-oxodG among all analyzed tis- clei of cells, where epigenetic processes take place. Re- sues, as well as by a decrease in 5-hmdC and 5-mdC cently, it was also demonstrated that 8-oxodG may serve levels (at a midrange between normal colon and CRC); as a demethylation signal : binding to 8-oxodG; (ii) AD had the lowest levels of 5-hmdC and 5-mdC of OGG1 glycosylase may recruit TET1 which in turn may all examined tissues and showed an increase in 8-oxodG be involved in specific DNA demethylation in response and 5-hmdU levels; (iii) CRC was characterized by lower to oxidative stress/oxidatively damaged DNA. This may levels of 5-hmdC and 5-mdC, the lowest level of 5-fdC at least partially explain a decrease in 5-mdC level ob- among all analyzed tissues, and relatively high content of served in AD and IBD, i.e., in precursor lesions charac- 5-cadC. This implies that a decrease in 5-hmdC level is terized by elevated levels of 8-oxodG. not a unique feature of CRC (as previously reported) Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 10 of 11 and can be also found in its precursor lesions, in par- Competing interests The authors declare that they have no competing interests. ticular in AD. The mechanism behind a substantial de- crease in 5-fdC generation at advanced stages of carcinogenesis is still unclear, and the same refers to Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in the consequences of this phenomenon. A recent ob- published maps and institutional affiliations. servation that 5-fdC is rich in active enhancers in- volved in tissue development/differentiation  Author details Department of Clinical Biochemistry, Faculty of Pharmacy, Collegium sheds a new light on these relationships, implying Medicum in Bydgoszcz, Nicolaus Copernicus University in Torun, Torun, that reduced level of this modification may be a char- Poland. Department of Clinical Pathomorphology, Faculty of Medicine, acteristic feature of largely undifferentiated cancer Collegium Medicum in Bydgoszcz, Nicolaus Copernicus University in Torun, Torun, Poland. Department of Surgery, Faculty of Medicine, Collegium cells. Medicum in Bydgoszcz, Nicolaus Copernicus University in Torun, Torun, It cannot be excluded that the analysis of a larger Poland. Department of Vascular Diseases and Internal Medicine, Faculty of spectrum of DNA epigenetic modifications, rather than Health Sciences, Collegium Medicum in Bydgoszcz, Nicolaus Copernicus University in Torun, Torun, Poland. Department of Oncologic Pathology and solely 5-hmdC, supported by a transcriptional information, Prophylaxis, Poznan University of Medical Sciences and Greater Poland might provide a better insight in carcinogenesis, risk factors Cancer Center, Poznan, Poland. Department of Otolaryngology and of CRC, and perhaps also therapy of this malignancy. Laryngeal Oncology, K. Marcinkowski University of Medical Sciences, Poznan, Poland. Department of Clinical Biochemistry, Collegium Medicum in Bydgoszcz, Nicolaus Copernicus University, Karlowicza 24, 85-095 Bydgoszcz, Additional files Poland. Received: 4 December 2017 Accepted: 17 May 2018 Additional file 1: Table S1. Characteristics of antibodies used in immunohistochemical staining. (PDF 499 kb) Additional file 2: Table S2. Transition patterns specific detector settings References and sources of standards for analyzed compounds (relative response ratio 1. Feng S, Jacobsen SE, Reik W. Epigenetic reprogramming in plant and = area under the peak of qualifierion/area under the peak of quantifier animal development. Science. 2010;330:622–7. ion). (PDF 82 kb) 2. Bhutani N, Burns DM, Blau HM. DNA demethylation dynamics. Cell. 2011; Additional file 3: Table S3. Primers and short hydrolysis probes used 146:866–72. for TETs and AID mRNA expression analysis. (PDF 442 kb) 3. Tahiliani M, Koh KP, Shen Y, Pastor WA, Bandukwala H, Brudno Y, Agarwal S, Iyer LM, Liu DR, Aravind L, Rao A. Conversion of 5-methylcytosine to 5- hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science. Abbreviations 2009;324:930–5. 2D-UPLC-MS/MS: Two-dimensional ultra-performance liquid chromatography 4. Pfaffeneder T, Spada F, Wagner M, Brandmayr C, Laube SK, Eisen D, Truss M, with tandem mass spectrometry; 5-cadC: 5-Carboxy-2’deoxycytidine; 5-fdC: 5- Steinbacher J, Hackner B, Kotljarova O, Schuermann D, Michalakis S, Formyl-2′-deoxycytidine; 5-hmdC: 5-Hydroxymethyl-2′-deoxycytidine; 5- Kosmatchev O, Schiesser S, Steigenberger B, Raddaoui N, Kashiwazaki G, hmdU: 5-(Hydroxymethyl)-2′-deoxyuridine; 5-mdC: 5-Methyl-2′-deoxycytidine; Muller U, Spruijt CG, Vermeulen M, Leonhardt H, Schar P, Muller M, Carell T. 8-oxodG: 8-Oxo-2′-deoxyguanosine; AD: Adenoma; AID: Activation-induced Tet oxidizes thymine to 5-hydroxymethyluracil in mouse embryonic stem cytidine deaminase protein; CRC: Colorectal cancer; FFPE: Formaldehyde- cell DNA. NatChemBiol. 2014;10:574–81. fixed paraffin-embedded; IBD: Inflammatory bowel disease; TET: Ten-eleven 5. Jin SG, Jiang Y, Qiu R, Rauch TA, Wang Y, Schackert G, Krex D, Lu Q, translocation protein; TMA: Tissue microarrays Pfeifer GP. 5-Hydroxymethylcytosine is strongly depleted in human cancers but its levels do not correlate with IDH1 mutations. Cancer Res. Funding 2011;71:7360–5. This work was supported by the Polish National Science Center [grant no. 6. Lian CG, Xu Y, Ceol C, Wu F, Larson A, Dresser K, Xu W, Tan L, Hu Y, Zhan Q, 2013/09/B/NZ5/00767]. Lee CW, Hu D, Lian BQ, Kleffel S, Yang Y, Neiswender J, Khorasani AJ, Fang R, Lezcano C, Duncan LM, Scolyer RA, Thompson JF, Kakavand H, Houvras Y, Zon LI, Mihm MC Jr, Kaiser UB, Schatton T, Woda BA, Murphy GF, Shi YG. Availability of data and materials Loss of 5-hydroxymethylcytosine is an epigenetic hallmark of melanoma. The datasets used and/or analyzed during the current study are available Cell. 2012;150:1135–46. from the corresponding author on reasonable request. 7. Yang H, Liu Y, Bai F, Zhang JY, Ma SH, Liu J, Xu ZD, Zhu HG, Ling ZQ, Ye D, Guan KL, Xiong Y. Tumor development is associated with decrease of TET Authors’ contributions gene expression and 5-methylcytosine hydroxylation. Oncogene. 2013;32: RO, AM, MF, ZB, MK, DG, and TD designed the research. ZB, MK, and ALi recruited 663–9. the participants of the study. AM and MB performed the immunochemical 8. Chen ML, Shen F, Huang W, Qi JH, Wang Y, Feng YQ, Liu SM, Yuan BF. analysis. TD, MS, EZ, MM, AS, JS, MG, KL, and ALa performed the isolation of DNA Quantification of 5-methylcytosine and 5-hydroxymethylcytosine in and its hydrolysis to deoxynucleosides. DG and MS performed determination of genomic DNA from hepatocellular carcinoma tissues by capillary the epigenetic modifications in DNA. TD, JG, and KL performed isolation of RNA hydrophilic-interaction liquid chromatography/quadrupole TOF mass and gene expression analysis. TD, DG, and MF performed statistical analysis. TD, spectrometry. Clin Chem. 2013;59:824–32. MF, and RO supervised the research. TD and RO wrote the manuscript. RO, AM, 9. Cimmino L, Abdel-Wahab O, Levine RL, Aifantis I. TET family proteins and MF, ZB, MK, DG, and TD revised and edited the manuscript. All authors read and their role in stem cell differentiation and transformation. Cell Stem Cell. approved the final manuscript. 2011;9:193–204. 10. Kraus S, Arber N. Inflammation and colorectal cancer. CurrOpinPharmacol. Ethics approval and consent to participate 2009;9:405–10. The study was conducted in accordance with the Declaration of Helsinki, its 11. Iurlaro M, Ficz G, Oxley D, Raiber EA, Bachman M, Booth MJ, Andrews S, protocol was approved by the Local Bioethics Committee at Collegium Balasubramanian S, Reik W. A screen for hydroxymethylcytosine and Medicum, Nicolaus Copernicus University in Bydgoszcz (Poland), and written formylcytosine binding proteins suggests functions in transcription and informed consent was sought from all the subjects. chromatin regulation. Genome Biol. 2013;14:R119. Dziaman et al. Clinical Epigenetics (2018) 10:72 Page 11 of 11 12. Spruijt CG, Gnerlich F, Smits AH, Pfaffeneder T, Jansen PW, Bauer C, 30. Hawthorn L, Lan L, Mojica W. Evidence for field effect cancerization in Munzel M, Wagner M, Muller M,Khan F,Eberl HC,MensingaA, colorectal cancer. Genomics. 2014;103:211–21. Brinkman AB, Lephikov K, Muller U, Walter J, Boelens R, van IH, 31. Franchini DM, Chan CF, Morgan H, Incorvaia E, Rangam G, Dean W, Santos Leonhardt H, Carell T, Vermeulen M. Dynamic readers for 5- F, Reik W, Petersen-Mahrt SK. Processive DNA demethylation via DNA (hydroxy)methylcytosine and its oxidized derivatives. Cell. 2013;152: deaminase-induced lesion resolution. PLoSOne. 2014;9:e97754. 1146–59. 32. Olinski R, Starczak M, Gackowski D. Enigmatic 5-hydroxymethyluracil: 13. Gackowski D, Starczak M, Zarakowska E, Modrzejewska M, Szpila A, oxidatively modified base, epigenetic mark or both? MutatResRevMutatRes. Banaszkiewicz Z, Olinski R. Accurate, direct, and high-throughput analyses of 2016;767:59–66. a broad spectrum of endogenously generated DNA base modifications with 33. Grin I, Ishchenko AA. An interplay of the base excision repair and mismatch isotope-dilution two-dimensional ultraperformance liquid chromatography repair pathways in active DNA demethylation. Nucleic Acids Res. 2016;44: with tandem mass spectrometry: possible clinical implication. AnalChem. 3713–27. 2016;88:12128–36. 34. Kohli RM, Zhang Y. TET enzymes, TDG and the dynamics of DNA demethylation. Nature. 2013;502:472–9. 14. Bodnar M, Szylberg L, Kazmierczak W, Marszalek A. Tumor progression 35. Xu W, Yang H, Liu Y, Yang Y, Wang P, Kim SH, Ito S, Yang C, Wang P, Xiao driven by pathways activating matrix metalloproteinases and their MT, Liu LX, Jiang WQ, Liu J, Zhang JY, Wang B, Frye S, Zhang Y, Xu YH, Lei inhibitors. JOral PatholMed. 2015;44:437–43. QY, Guan KL, Zhao SM, Xiong Y. Oncometabolite 2-hydroxyglutarate is a 15. Bodnar M, Luczak M, Bednarek K, Szylberg L, Marszalek A, Grenman R, competitive inhibitor of alpha-ketoglutarate-dependent dioxygenases. Szyfter K, Jarmuz-Szymczak M, Giefing M. Proteomic profiling identifies the Cancer Cell. 2011;19:17–30. inorganic pyrophosphatase (PPA1) protein as a potential biomarker of 36. Figueroa ME, Bdel-Wahab O, Lu C, Ward PS, Patel J, Shih A, Li Y, Bhagwat N, metastasis in laryngeal squamous cell carcinoma. AminoAcids. 2016;48: Vasanthakumar A, Fernandez HF, Tallman MS, Sun Z, Wolniak K, Peeters JK, 1469–76. Liu W, Choe SE, Fantin VR, Paietta E, Lowenberg B, Licht JD, Godley LA, 16. Bodnar M, Burduk P, Antosik P, Jarmuz-Szymczak M, Wierzbicka M, Marszalek Delwel R, Valk PJ, Thompson CB, Levine RL, Melnick A. Leukemic IDH1 and A. Assessment of BRAF V600E (VE1) protein expression and BRAF gene IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 mutation status in codon 600 in benign and malignant salivary gland function, and impair hematopoietic differentiation. Cancer Cell. 2010;18:553– neoplasms. JOral PatholMed. 2017;46:340–5. 17. Uhlen M, Oksvold P, Fagerberg L, Lundberg E, Jonasson K, Forsberg M, 37. Wu X, Zhang Y. TET-mediated active DNA demethylation: mechanism, Zwahlen M, Kampf C, Wester K, Hober S, Wernerus H, Bjorling L, Ponten F. function and beyond. NatRevGenet. 2017;18:517–34. Towards a knowledge-based human protein atlas. NatBiotechnol. 2010;28: 38. Kan Z, Jaiswal BS, Stinson J, Janakiraman V, Bhatt D, Stern HM, Yue P, 1248–50. Haverty PM, Bourgon R, Zheng J, Moorhead M, Chaudhuri S, Tomsho LP, 18. Remmele W, Stegner HE. Recommendation for uniform definition of an Peters BA, Pujara K, Cordes S, Davis DP, Carlton VE, Yuan W, Li L, Wang W, immunoreactive score (IRS) for immunohistochemical estrogen receptor Eigenbrot C, Kaminker JS, Eberhard DA, Waring P, Schuster SC, Modrusan Z, detection (ER-ICA) in breast cancer tissue. Pathologe. 1987;8:138–40. Zhang Z, Stokoe D, de Sauvage FJ, Faham M, Seshagiri S. Diverse somatic 19. Burduk PK, Bodnar M, Sawicki P, Szylberg L, Wisniewska E, Kazmierczak W, mutation patterns and pathway alterations in human cancers. Nature. 2010; Martynska M, Marszalek A. Expression of metalloproteinases 2 and 9 and 466:869–73. tissue inhibitors 1 and 2 as predictors of lymph node metastases in 39. Seshagiri S, Stawiski EW, Durinck S, Modrusan Z, Storm EE, Conboy CB, oropharyngeal squamous cell carcinoma. Head Neck. 2015;37:418–22. Chaudhuri S, Guan Y, Janakiraman V, Jaiswal BS, Guillory J, Ha C, Dijkgraaf 20. Guz J, Foksinski M, Siomek A, Gackowski D, Rozalski R, Dziaman T, Szpila A, GJ, Stinson J, Gnad F, Huntley MA, Degenhardt JD, Haverty PM, Bourgon R, Olinski R. The relationship between 8-oxo-7,8-dihydro-2′-deoxyguanosine Wang W, Koeppen H, Gentleman R, Starr TK, Zhang Z, Largaespada DA, Wu level and extent of cytosine methylation in leukocytes DNA of healthy TD, de Sauvage FJ. Recurrent R-spondin fusions in colon cancer. Nature. subjects and in patients with colon adenomas and carcinomas. MutatRes. 2012;488:660–4. 2008;640:170–3. 40. Zhang YW, Wang Z, Xie W, Cai Y, Xia L, Easwaran H, Luo J, Yen RC, Li Y, 21. Herceg Z, Vaissiere T. Epigenetic mechanisms and cancer: an interface Baylin SB. Acetylation enhances TET2 function in protecting against between the environment and the genome. Epigenetics. 2011;6:804–19. abnormal DNA methylation during oxidative stress. MolCell. 2017;65:323–35. 22. Uribe-Lewis S, Stark R, Carroll T, Dunning MJ, Bachman M, Ito Y, Stojic L, 41. Thienpont B, Steinbacher J, Zhao H, D'Anna F, Kuchnio A, Ploumakis A, Halim S, Vowler SL, Lynch AG, Delatte B, de Bony EJ, Colin L, Defrance M, Ghesquiere B, Van DL, Boeckx B, Schoonjans L, Hermans E, Amant F, Krueger F, Silva AL, Ten HR, Ibrahim AE, Fuks F, Murrell A. 5- Kristensen VN, Koh KP, Mazzone M, Coleman ML, Carell T, Carmeliet P, hydroxymethylcytosine marks promoters in colon that resist DNA Lambrechts D. Tumour hypoxia causes DNA hypermethylation by reducing hypermethylation in cancer. Genome Biol. 2015;16:69. TET activity. Nature. 2016;537:63–8. 23. Delatte B, Jeschke J, Defrance M, Bachman M, Creppe C, Calonne E, Bizet M, 42. Debevec T, Millet GP, Pialoux V. Hypoxia-induced oxidative stress Deplus R, Marroqui L, Libin M, Ravichandran M, Mascart F, Eizirik DL, Murrell modulation with physical activity. Front Physiol. 2017;8:84. A, Jurkowski TP, Fuks F. Genome-wide hydroxymethylcytosine pattern 43. Afanas’ev I. Mechanisms of superoxide signaling in epigenetic processes: changes in response to oxidative stress. SciRep. 2015;5:12714. relation to aging and cancer. Aging Dis. 2015;6:216–27. 24. Eleftheriou M, Pascual AJ, Wheldon LM, Perry C, Abakir A, Arora A, Johnson 44. Cyr AR, Domann FE. The redox basis of epigenetic modifications: from AD, Auer DT, Ellis IO, Madhusudan S, Ruzov A. 5-Carboxylcytosine levels are mechanisms to functional consequences. AntioxidRedoxSignal. 2011;15:551–89. elevated in human breast cancers and gliomas. ClinEpigenetics. 2015;7:88. 45. Zhou X, Zhuang Z, Wang W, He L, Wu H, Cao Y, Pan F, Zhao J, Hu Z, Sekhar 25. Wang D, Hashimoto H, Zhang X, Barwick BG, Lonial S, Boise LH, Vertino PM, C, Guo Z. OGG1 is essential in oxidative stress induced DNA demethylation. Cheng X. MAX is an epigenetic sensor of 5-carboxylcytosine and is altered Cell Signal. 2016;28:1163–71. in multiple myeloma. Nucleic Acids Res. 2017;45:2396–407. 46. Iurlaro M, McInroy GR, Burgess HE, Dean W, Raiber EA, Bachman M, Beraldi 26. Xiong J, Zhang Z, Chen J, Huang H, Xu Y, Ding X, Zheng Y, Nishinakamura D, Balasubramanian S, Reik W. In vivo genome-wide profiling reveals a R, Xu GL, Wang H, Chen S, Gao S, Zhu B. Cooperative action between tissue-specific role for 5-formylcytosine. Genome Biol. 2016;17:141. SALL4A and TET proteins in stepwise oxidation of 5-methylcytosine. Mol Cell. 2016;64:913–25. 27. Cheng J, Deng R, Zhang P, Wu C, Wu K, Shi L, Liu X, Bai J, Deng M, Shuai X, Gao J, Wang G, Tao K. miR-219-5p plays a tumor suppressive role in colon cancer by targeting oncogene Sall4. Oncol Rep. 2015;34:1923–32. 28. Jin SG, Zhang ZM, Dunwell TL, Harter MR, Wu X, Johnson J, Li Z, Liu J, Szabo PE, Lu Q, Xu GL, Song J, Pfeifer GP. Tet3 reads 5-carboxylcytosine through its CXXC domain and is a potential guardian against neurodegeneration. Cell Rep. 2016;14:493–505. 29. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brakenhoff RH. A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications. Cancer Res. 2003;63:1727–30.
– Springer Journals
Published: May 30, 2018