Bazzocco et al. Clin Epigenet (2021) 13:88 https://doi.org/10.1186/s13148-021-01070-0 RESEARCH Open Access Identification of ZBTB18 as a novel colorectal tumor suppressor gene through genome‑wide promoter hypermethylation analysis Sarah Bazzocco1†, Higinio Dopeso1† , Águeda Martínez‑Barriocanal1,8† , Estefanía Anguita1 , Rocío Nieto1 , Jing Li1 , Elia García‑Vidal1 , Valentina Maggio1 , Paulo Rodrigues1 , Priscila Guimarães de Marcondes1 , Simo Schwartz Jr2,3 , Lauri A. Aaltonen4 , Alex Sánchez5 , John M. Mariadason6,7 and Diego Arango1,8* Abstract Background: Cancer initiation and progression are driven by genetic and epigenetic changes. Although genome/ exome sequencing has significantly contributed to the characterization of the genetic driver alterations, further inves‑ tigation is required to systematically identify cancer driver genes regulated by promoter hypermethylation. Results: Using genome-wide analysis of promoter methylation in 45 colorectal cancer cell lines, we found that higher overall methylation levels were associated with microsatellite instability (MSI), faster proliferation and absence of APC mutations. Because epigenetically silenced genes could represent important oncogenic drivers, we used mRNA expression profiling of colorectal cancer cell lines and primary tumors to identify a subset of 382 (3.9%) genes for which promoter methylation was negatively associated with gene expression. Remarkably, a significant enrich‑ ment in zinc finger proteins was observed, including the transcriptional repressor ZBTB18. Re-introduction of ZBTB18 in colon cancer cells significantly reduced proliferation in vitro and in a subcutaneous xenograft mouse model. Moreover, immunohistochemical analysis revealed that ZBTB18 is frequently lost or reduced in colorectal tumors, and reduced ZBTB18 expression was found to be associated with lymph node metastasis and shorter survival of patients with locally advanced colorectal cancer. Conclusions: We identified a set of 382 genes putatively silenced by promoter methylation in colorectal cancer that could significantly contribute to the oncogenic process. Moreover, as a proof of concept, we demonstrate that the epigenetically silenced gene ZBTB18 has tumor suppressor activity and is a novel prognostic marker for patients with locally advanced colorectal cancer. Keywords: Colorectal cancer, Zinc finger, ZBTB18, Methylation, Tumor suppressor Background Colorectal cancer is the third most common can- cer worldwide, and although patient prognosis largely depends on tumor stage at the time of diagnosis [1], *Correspondence: darango@irblleida.cat † Sarah Bazzocco, Higinio Dopeso, Águeda Martínez-Barriocanal have colorectal cancer represents the second cause of cancer- contributed equally to this work related death [2]. Progression of colorectal cancer is asso- 1 Group of Biomedical Research in Digestive Tract Tumors, ciated with the accumulation of genetic and epigenetic CIBBIM‑Nanomedicine, Vall d’Hebron University Hospital, Research Institute (VHIR), Universitat Autònoma de Barcelona, Passeig Vall defects that drive the multistep process from normal d’Hebron, 119‑129, 08035 Barcelona, Spain intestinal epithelial cells to a metastatic colorectal tumor. Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Bazzocco et al. Clin Epigenet (2021) 13:88 Page 2 of 13 The genetic defects driving the tumorigenic process in chromatin that can be recognized by transcriptional colorectal cancer have been studied for several decades complexes to activate or repress gene expression. Histone and include mutations in APC, KRAS and TP53 among methylation and demethylation are catalyzed by histone other genes [3]. Epigenetic changes, such as DNA methyl- methyltransferases (HMTs) and histone demethylases ation or histone modification, modulate gene expression (HDMs), respectively [4]. without altering the primary DNA nucleotide sequence. While the genetic events that drive the tumorigenic DNA methylation is one of the most ubiquitous epige- process are relatively well understood for colorectal netic modifications, and the best-characterized DNA cancer, the epigenetic events and their impact on the methylation process, involving the addition of a methyl transcriptional reprogramming observed in colorectal group at the C5 position of the cytosine ring by DNA tumors have not been as extensively characterized. Spe- methyltransferases [4]. Cytosine methylation in the con- cifically for DNA methylation, although genome-wide text of CpG dinucleotides is common across the genome studies have analyzed the genomic distribution of hyper- of normal human cells, but global hypomethylation has methylated CpGs in colorectal tumors [3, 7], a detailed been observed in cancer cells [5]. However, regions rich analysis of the subset of these events that are important in CpG dinucleotides, known as CpG islands, are often for gene expression deregulation during colorectal tumo- hypermethylated in tumors [6, 7]. These CpG islands are rigenesis is currently lacking. To bridge this gap, in this present in the promoter regions of over 50% of human study we used genome-wide CpG methylation analysis of genes, and changes in the levels of methylation constitute a panel of 45 colorectal cancer cell lines and 223 primary an important mechanism of transcriptional deregulation colorectal tumors available at The Cancer Genome Atlas during tumor progression [6]. (TCGA) to investigate the distribution of DNA meth- Development of colorectal cancer is driven by three ylation changes in colorectal cancer [3]. In addition, by major classical mechanisms of genomic instability. Chro- integrating methylation data with mRNA expression data mosomal instability, resulting in large chromosomal gains we identified a set of 382 genes, enriched in zinc finger and losses, is a common feature in approximately 80% proteins, for which high levels of promoter methylation of colorectal tumors and is an important driver of the was associated with reduced gene expression. As a proof oncogenic process. An alternative tumorigenic pathway of concept, we demonstrate that ZBTB18, a zinc finger known as ‘CpG island methylator phenotype’ (CIMP) is transcriptional repressor that is frequently methylated in observed in 10–20% of the tumors, and it is character- colorectal tumors, has tumor suppressor activity and is a ized by the widespread hypermethylation of CpG islands novel prognostic factor in colorectal cancer. throughout the genome [8]. Finally, defects in mismatch repair mechanisms result in the accumulation of muta- Results tions mostly in nucleotide repeats and the microsatel- Genome‑wide analysis of CpG methylation in colorectal lite instable (MSI) phenotype observed in 10–20% of the cancer cell lines tumors in the colon and rectum [9–11]. Notably, about To investigate genome-wide promoter methylation levels 70–80% of MSI colorectal tumors can be attributed to and possible associations with frequent genetic changes hypermethylation of the MLH1 promoter in CIMP cases observed during colorectal tumorigenesis, the meth- [12, 13], although CIMP also occurs without MSI [14]. ylation of 27,578 CpG dinucleotides in 14,495 genes was Interestingly, MSI + and CIMP + tumors are associated assessed in a panel of 45 colorectal cancer cell lines using with proximal tumor location, poor differentiation and an Infinium HumanMethylation27 BeadChip platform BRAF mutations [12] and may arise from serrated adeno- (Illumina). A quantitative measurement of the meth- mas or polyps [13, 15]. ylation levels (beta values) in each of the 27,578 CpGs Histone modifications also play an important role in was obtained. Global methylation values ranged from 0 the epigenetic regulation of colorectal tumorigenesis. The (unmethylated) to 1 (fully methylated). As expected, low most common and best-studied histone modifications levels of methylation were observed in an unmethylated altering gene expression are acetylation and methylation. control (mean methylation 0.05) and an in vitro methyl- Acetylation is catalyzed by histone acetyl transferases ated sample showed the highest methylation levels (mean (HATs) and counteracts the positive charge on the tails of methylation 0.86; Fig. 1a). Moreover, independent deter- histones, weakening the electrostatic interaction between minations of genome-wide methylation in SW48 cells DNA and histones and consequently, reducing the com- demonstrated excellent correlation (inset in Fig. 1a), and paction of chromatin and enhancing gene expression. independent bisulfite sequencing analysis of selected pro- Contrary, deacetylation is associated with gene silenc- moters in a subset of cell lines was in perfect agreement ing and is controlled by histone deacetylases (HDACs). with the HumanMethylation27 BeadChip determinations In turn, histone methylation creates docking sites in the obtained (Additional file 1: Fig. S1). Bazzocco et al. Clin Epigenet (2021) 13:88 Page 3 of 13 a b 1.0 50 Average CpG methylation 1 .0 P e a r s o n 's R = 0 .9 9 p < 0 .0 0 0 1 40 CpGs interrogated (%) 30 0 .8 S W 4 8 r e p lic a t e 1 0.8 0 .6 15 0 .4 0.6 0 .2 10 0.4 0 .0 0 .0 0 .2 0 .4 0 .6 0 .8 1 .0 S W 4 8 r e p lic a t e 2 0.2 5 0.0 0 LS174T RW7213 FET IS3 LS1034 RKO IS2 SW403 SKCO1 SW948 RW2982 SW1116 Colo320 V9P LOVO LIM1215 LS513 Colo201 SW620 Caco2 Vaco5 IS1 Control + Control - Colo205 SW837 HT29 LIM2405 KM12 Ala TC71 SW48 Co115 GP5D T84 HDC87 HDC114 HDC9 HDC75 HCA7 HCT15 DLD1 HCC2998 HDC111 HDC15 HDC108 HCT116 <0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 0.5-0.6 0.6-0.7 0.7-0.8 0.8-0.9 >0.9 Range of methylation Fig. 1 CpG methylation in colorectal cancer cell lines. a The average (± SEM) methylation (beta-value) of all the 27,578 CpGs interrogated is shown in a panel of 45 colorectal cancer cell lines. Control − : whole genome amplification with GenomiPhi, GE Healthcare; Control + : in vitro methylated with SssI. Inset shows the correlation between methylation measurements for all CpGs in two replicates of SW48. b Histogram showing the percentage of the CpG dinucleotides interrogated that have average methylation levels across the 45 cell lines in the indicated intervals The average methylation across all 27,578 CpGs in the majority (50.2%) showed low average methylation across 45 colorectal cancer cell lines investigated showed large the panel of 45 colorectal cancer cell lines (< 0.2 methyla- variability, ranging from 0.22 to 0.43 (Fig. 1a), indicat- tion), while over 30% of the CpGs showed average meth- ing that cancer cells derived from colorectal tumors can ylation levels above 0.5 (Fig. 1b), which again was similar widely vary in their overall levels of methylation. Next, we to the distribution observed in primary colorectal tumors interrogated global methylation in 233 primary tumors (Additional file 1: Fig. S2b). analyzed by The Cancer Genome Atlas (TCGA) project In agreement with the reported association between for which the same HumanMethylation27 platform was colorectal tumors with microsatellite instability (MSI) used [3]. The variability observed across patients’ primary and the CpG methylator phenotype (CIMP) [10], the tumors ranged from 0.19 to 0.35 (Additional file 1: Fig. overall levels of methylation in the panel of cell lines used S2a), indicating that the global DNA methylation levels in our study were significantly higher in MSI cell lines in cell lines reflect that of primary tumors. When the lev- and primary tumors compared to the respective micro- els of methylation of individual CpGs was considered, the satellite stable (MSS) counterpart (Fig. 2a and Additional a MS Status b APC Status c 0.5 0.5 80 ** * Proportion of methylation Proportion of methylation 0.4 0.4 Doubling time (h) 60 0.3 0.3 40 0.2 0.2 20 0.1 0.1 Pearson's r=-0.39 n=29 n=17 n=7 n=20 p=0.01 0.0 0.0 0 MSS MSI WT MUT 0.2 0.3 0.4 0.5 Average CpG methylation Fig. 2 Associations between the average levels of methylation and molecular features of colorectal cancer cell lines. Higher levels of CpG methylation were associated with microsatellite instability (MSI status) (a), wild-type APC status (b) and faster growth (i.e., shorter doubling time) (c) in the panel of 45 colorectal cancer cell lines. Student’s t test *p < 0.05 and **p < 0.01 Bazzocco et al. Clin Epigenet (2021) 13:88 Page 4 of 13 file 1: Fig. S3a). In addition, the average levels of meth- of 9 of these genes in cell lines showing low endogenous ylation across the 27,578 CpGs investigated were sig- expression and high CpG promoter methylation (Addi- nificantly lower in cell lines and primary tumors with tional file 1: Fig. S7), suggesting a direct regulation of APC mutations compared to cell lines and tumors with- gene expression by promoter hypermethylation for most out mutations in this important tumor suppressor gene of these genes. (Fig. 2b and Additional file 1: Fig. S3b). Primary tumors Although the analysis of genes regulated by promoter with BRAF mutations had significantly higher levels of methylation in primary colorectal tumor samples is com- methylation than BRAF wild-type tumors, while TP53 plicated by the presence of varying proportions of non- mutant tumors had lower levels of methylation than tumor cells, we used the cohort of 223 primary colorectal wild-type TP53 tumors, although these differences did tumors from the TCGA repository [3] to further validate not reach statistical significance in cell lines (Additional the association observed between promoter hypermeth- file 1: Fig. S3c–d and g–h). No associations were observed ylation and reduced gene expression. Of the 740 genes between the average methylation levels and mutations in identified as potentially regulated by promoter meth- other frequently mutated driver genes in colorectal can- ylation in the colorectal cancer cell lines, expression and cer, including KRAS or PIK3CA (Additional file 1: Fig. methylation data in the primary tumors were available S3e–f and i–j). In addition, we also observed a significant from 653 genes (88.2%). A significant negative correlation association between higher methylation levels and faster (FDR adjusted Spearman’s p < 0.2) between expression growth rate across the cell lines (Fig. 2c), suggesting that and methylation was observed for 382 of these 653 genes higher levels of promoter CpG methylation may contrib- (58.5%; see Additional file 1: Table S2), identifying a core ute to increased proliferation rates. set of genes potentially silenced by promoter methylation in colorectal cancer. Associations between promoter methylation and gene expression in colorectal cancer Although it is widely accepted that aberrant CpG meth- Methylation outside CpG islands and gene expression ylation is a frequent event during the oncogenic process DNA methylation within CpG islands and flanking and that it plays an important role in the transcriptional regions (‘shores’) is known to be closely linked to tran- regulation of many genes, a detailed list of genes whose scriptional repression [7]. Accordingly, we found that the transcription is frequently regulated by CpG methylation majority of genes that are potentially silenced epigeneti- in colorectal cancer cells is currently lacking. To this end, cally in colorectal cancer cells are associated with CpG we used microarray analysis to assess the levels of expres- islands. Of the 382 genes whose expression was nega- sion of 11,359 (78.47%) of the 14,475 promoters interro- tively associated with CpG methylation, 303 (79.3%) were gated in the HumanMethylation27 BeadChips in a subset associated with a CpG island as defined by Gardiner-Gar- of 30 colorectal cancer cell lines. den and Frommer [16] (Additional file 1: Table S2 and This analysis identified 712 genes (6.3%) with a sig- Figure S7a–b). This is significantly above what would be nificant negative correlation (FDR adjusted Spearman’s expected by chance given the frequency of human genes p < 0.2) between promoter methylation and mRNA with CpG islands associated. In ten randomly selected expression in this panel of 30 colorectal cancer cell lines groups of 382 genes, the average percentage of genes and is therefore candidate genes that may be epige- with associated CpG islands was 66.2% (Fisher exact test netically silenced in colorectal cancer cells (Additional p < 0.0001). file 1: Fig. S4). In addition, we identified 28 genes whose However, the role of promoter methylation in the tran- expression was not significantly correlated with the lev- scriptional regulation of genes not associated with CpG els of methylation (FDR adjusted Spearman’s p > 0.2), but islands remains unclear. Here we found that 79 of the showed a nonlinear negative association between expres- 382 genes (20.7%) with significant correlations between sion and methylation levels (see Methods section; Addi- mRNA and methylation levels, were not associated tional file 1: Fig. S5). In total, therefore, 740 genes (6.5%) with a conventionally defined CpG island. This included were identified as candidate genes whose expression may PPP1R14D, BST2 and ZNF550, whose expression was be regulated by promoter methylation in colorectal can- induced by 5-aza-2′-deoxycytidine (decitabine) treat- cer cells. ment in colon cancer cell lines with low basal expression The relative levels of expression of selected genes were and high promoter methylation of these genes (Addi- independently confirmed by qPCR (Additional file 1: Fig. tional file 1: Table S2 and Additional file 1: Fig. S7c–e). S6). In addition, increased mRNA expression follow- These findings suggest that CpG methylation outside ing treatment with the demethylating agents 5-aza-2′- conventionally defined CpG islands may also regulate deoxycytidine (decitabine) was demonstrated for a subset Bazzocco et al. Clin Epigenet (2021) 13:88 Page 5 of 13 gene expression and contribute to colorectal cancer corresponding EGFP control lines (Fig. 3a and Additional progression. file 1: Fig. S10a). Next, the effects of ZBTB18 overexpres- sion on cell growth were assessed using a subcutane- ZBTB18 has tumor suppressor activity in colorectal cancer ous xenograft model in NOD/SCID immunodeficient cells mice. ZBTB18 overexpression resulted in a significant To gain a better understanding of the biological func- growth reduction of both HCT116 and HT29 cells over tion of the 382 genes potentially silenced epigenetically the course of four weeks compared to the correspond- during colorectal tumorigenesis, we performed gene set ing EGFP overexpressing control cell lines (Fig. 3b–c and enrichment analysis and identified 15 enriched annota- Additional file 1: Fig. S10b–c). Consistently, the weight of tion terms (Fisher’s Exact test, FDR p < 0.2; Additional the subcutaneous tumors formed by the ZBTB18 overex- file 1: Table S3). Strikingly, all these largely overlapping pressing HCT116 and HT29 cells was significantly lower categories were related to zinc finger proteins. The larg- than in the tumors formed by the corresponding control est of these categories (zinc finger region:C2H2-type 4) sublines (Fig. 3d and Additional file 1: Fig. S10d), collec- contained 33 zinc finger domain proteins, including a tively demonstrating that reintroduction of ZBTB18 into large proportion of transcriptional regulators (Additional colon cancer cells significantly reduced their growth. file 1: Table S3). Next, we investigated whether ZBTB18 overexpression We postulated that at least some of these epigenetically caused changes in the cell cycle or the number of apop- silenced genes may have tumor suppressor activity in totic cells. No significant changes were observed in the colorectal cancer. As a proof of concept, we investigated distribution of cells along the different phases of the cell the potential tumor suppressive function of the tran- cycle in HT29 or HCT116 cells overexpressing ZBTB18 scriptional repressor ZBTB18 (also known as ZNF238 or compared to the corresponding control cell lines (Fig. 3e RP58; Additional file 1: Figs. S1a, S4a, S6a and Table S2), and Additional file 1: Fig. S10e). However, although which belongs to the zinc finger C2H2 family, and whose ZBTB18 overexpression had no effect on the number of role in colorectal cancer progression has not been previ- apoptotic cells in HT29 cells (Additional file 1: Fig. S10f ), ously addressed. increased apoptosis was observed in HCT116 cells over- First, we showed that treatment of four colorectal expressing ZBTB18 compared to control cells (Fig. 3f ), cancer cell lines with low ZBTB18 expression and high likely contributing to the reduced growth observed. promoter methylation with the demethylating agent We next investigated the role of ZBTB18 on the migra- decitabine (5-aza-2′-deoxycytidine) resulted in a sig- tion and invasion capacity of colon cancer cells. Ectopic nificant increase of ZBTB18 mRNA expression (Addi- expression of ZBTB18 had no effect on the migration tional file 1: Fig. S8a). Ladakamycin (5-azacytidine), a capacity of HT29 or HCT116 cells when assessed in vitro second demethylating agent used clinically, also upregu- in a wound healing assay. However, ZBTB18 overexpres- lated ZBTB18 at the mRNA level, further supporting sion in HT29 cells significantly reduced their Matrigel the notion that ZBTB18 expression is regulated by pro- invasion capacity compared to control HT29 cells moter methylation (Additional file 1: Fig. S8a). Next, we (Fig. 3g–h and Additional file 1: Fig. S10g–h). wondered whether ZBTB18 could also be epigenetically regulated by histone modifications. Contrary to DNA Reduced ZBTB18 expression is associated with poor methyltransferase inhibitors, drugs interfering with his- prognosis of colorectal cancer patients tone acetylation or methylation did not have a significant To further investigate the impact of ZBTB18 silencing in and consistent impact on ZBTB18 expression (Additional colorectal tumorigenesis, we assessed the possible asso- file 1: Fig. S8b). ciation between ZBTB18 expression and patient survival Next, we engineered two isogenic cell line systems with in a cohort of 132 cases with locally advanced (Dukes’ constitutive or doxycycline-inducible overexpression of C) colorectal cancer. Sections of a tissue microarray bicistronic ZBTB18 and EGFP or control EGFP alone containing triplicate samples from these 132 primary in HCT116 and HT29 cells, respectively (Additional tumors, as well as paired normal mucosa samples from file 1: Fig. S9). FACS analysis demonstrated that > 87% a subset of 75 cases and 15 lymph node metastases, were of the cells were EGFP-positive in the HCT116 and immunostained using a rabbit polyclonal anti-ZBTB18 HT29 derivative lines (Additional file 1: Fig. S9a and c). antibody. The specificity of the antibody was confirmed Consistently, increased ZBTB18 mRNA expression was using formalin-fixed, paraffin-embedded samples of confirmed in these derivative cell line systems by qPCR control HCT116 and HT29 (overexpressing EGFP) (Additional file 1: Fig. S9b and d). ZBTB18 overexpres- and the corresponding derivative lines overexpressing sion in HCT116 and HT29 cells resulted in a signifi- ZBTB18 (ZBTB18-IRES-EGFP) grown as subcutane- cant reduction of their growth in vitro compared to the ous xenografts in NOD/SCID immunodeficient mice Bazzocco et al. Clin Epigenet (2021) 13:88 Page 6 of 13 a b 1500 c EV ZBTB18 d 2.5 EV EV 6 Number of cells (x105) ZBTB18 ZBTB18 1250 Tumor size (mm3) 2.0 Tumor weight (g) 1000 4 1.5 ** * 750 * 1.0 * 2 500 * * * 0.5 250 * 0 0 0.0 0 24 48 72 96 0 10 20 30 EV ZBTB18 Time (h) Time (days) e 50 f 20 G1 S G2/M 40 18 *** SubG1 cells (%) Cells (%) 30 16 20 14 10 12 0 10 EV ZBTB18 EV ZBTB18 EV ZBTB18 EV ZBTB18 g 100 h 2.0 EV ZBTB18 Invading Cells (Fold) Wound Closure (%) 80 1.5 60 1.0 40 0.5 20 0 0.0 0 12 24 36 48 EV ZBTB18 Time (h) Fig. 3 Effects of ZBTB18 overexpression on the growth of colon cancer cells. The effects of ZBTB18 overexpression on the growth of HCT116 cells was assessed by directly counting the number of cells in culture over time (a). The growth of HCT116 cells was monitored over time after subcutaneous implantation in immunodeficient NOD/SCID mice (b). At the end of the experiment the tumors were dissected out (c) and the average weight (± SEM) of the tumors was measured (d). Cells in the different phases of the cell cycle (e) and SubG1 cells (f) were assessed to evaluate the effect of ZBTB18 overexpression in cell growth and apoptosis. Migration (g) and invasion (h) properties of parental and ZBTB18 overexpressing cells were assessed in wound healing and Matrigel-coated Boyden chamber assays, respectively. Student’s t test *p < 0.05, **p < 0.01 and ***p < 0.001 Bazzocco et al. Clin Epigenet (2021) 13:88 Page 7 of 13 a b c d e f g h i j 3 *** 3 * ZBTB18 expression ZBTB18 expression *** 2 2 1 1 0 n=15 n=63 n=75 0 N T LN T LN k l 100 Low ZBTB18 (n=112) 100 Low ZBTB18 (n=110) High ZBTB18 (n=20) High ZBTB18 (n=19) Percent survival Percent survival 75 75 50 50 25 25 Logrank p=0.02 Logrank p=0.09 0 0 0 2 4 6 8 10 0 2 4 6 8 10 Overall survival (years) Disease-free survival (years) Fig. 4 ZBTB18 expression in primary colorectal tumors. Normal colonic epithelial cells had high expression of ZBTB18 (a–b). Significant variability was observed in the expression of ZBTB18 in colorectal tumors (c–g). i Histogram showing the average (± SEM) expression of ZBTB18 in normal colonic epithelial cells, primary colorectal tumors and lymph node metastases. j ZBTB18 expression in 13 primary colorectal tumors and paired lymph node metastasis. k–l Overall (k) and disease-free (l) survival as a function of ZBTB18 expression in the tumors from 132 Dukes’ C colorectal cancer patients is shown. N: number of samples. Scale bar: 50 µm. Student’s t test *p < 0.05 and ***p < 0.0001 Bazzocco et al. Clin Epigenet (2021) 13:88 Page 8 of 13 (Additional file 1: Fig. S11). Moreover, ZBTB18 protein mechanism leading to the initiation and/or progression levels assessed with the same antibody in a tissue micro- of different tumor types, including tumors of the colon array containing the 30 colorectal cancer cell lines used and rectum [6]. However, the detailed transcriptional in this study significantly correlated with ZBTB18 mRNA reprograming resulting from aberrant methylation in expression (Additional file 1: Fig. S12), further confirm- colorectal tumors has not been thoroughly characterized. ing the specificity of the primary anti-ZBTB18 antibody The advent of high throughput technologies to interro- used. gate genome-wide levels of mRNA expression and CpG High ZBTB18 immunostaining was observed in nor- methylation now allows for the systematic identification mal colonic epithelial cells (Fig. 4a, b). Significant vari- of genes potentially silenced by DNA methylation. ability in ZBTB18 expression was observed in primary Here, we analyzed the levels of methylation of 27,578 colorectal tumors, ranging from high levels comparable individual CpGs in a panel of 45 colorectal cancer cell to the ZBTB18 expression found in normal colonic epi- lines and compared them with 223 primary colorectal thelial cells (Fig. 4c, d), to low or undetectable ZBTB18 tumors [3]. It has recently been shown that colorectal expression (Fig. 4e, h). Overall, the average expression cancer cell lines faithfully represent and maintain the observed in primary colorectal tumors was significantly genetic diversity of primary tumors [9]. Importantly, we reduced compared to normal colonic epithelial cells now report that the epigenetically regulated genes in (Fig. 4i). Moreover, ZBTB18 protein expression was fur- colon cancer cell lines and primary tumors is also highly ther reduced in lymph node metastases compared to concordant, indicating that colon cancer cell lines consti- primary colorectal tumors (Fig. 4i, j). Importantly, an tute good models to investigate the role of DNA methyla- association was observed between low ZBTB18 expres- tion in this disease. sion in this cohort of 132 Dukes’ C colorectal cancer Higher levels of methylation were associated with patients and shorter overall survival (Fig. 4k). A corre- microsatellite instability, BRAF mutations and wild-type sponding trend to shorter disease-free survival was also APC and TP53 status. This is in good agreement with observed in patients with low ZBTB18 expression in their earlier studies showing that CpG island methylator phe- primary tumors, although it did not reach statistical sig- notype (CIMP) positive tumors, a subset of colorectal nificance (Fig. 4l). In good agreement, low expression of tumors that has high levels of methylation in CpG islands ZBTB18 mRNA was significantly associated with shorter [18], is associated with microsatellite instability, higher survival in an independent cohort of 55 Stage III colorec- BRAF and lower APC and TP53 mutation rates [10, 19, tal cancer patients from the TCGA cohort [3, 17] (Addi- 20]. Hypermethylation of the promoter of MLH1, an tional file 1: Fig. S13). No associations were observed important component of the DNA mismatch repair sys- between ZBTB18 expression and other clinicopathologi- tem, is common in CIMP + tumors and causes the MSI cal features of the patients in both cohorts (Additional phenotype observed in most of the CIMP + tumors [3, file 1: Tables S4-5). 10]. This is consistent with our results showing that MSI Collectively, these results demonstrate that ZBTB18 cell lines and primary tumors have higher levels of meth- expression is progressively reduced in the transition from ylation. Moreover, in good agreement with our earlier normal colonic epithelial cells to primary colorectal car- findings showing that MSI colorectal cancer cells pro- cinoma and lymph node metastasis and that low expres- liferate faster than MSS tumor cells [21], and given that sion of ZBTB18 in locally advanced primary colorectal MSI tumors have higher levels of methylation, we found tumors is associated with shorter patient survival. here that colon cancer cell lines with higher levels of DNA methylation are associated with faster proliferation rates. Collectively, these results indicate that global pro- Discussion moter methylation in the panel of cell lines derived from Cytosine methylation in CpG dinucleotides is important colorectal tumors used in this study closely resembles the during embryonic development and cell differentiation in heterogeneity of DNA methylation observed in primary higher organisms. These methylation marks lead to chro- colorectal tumors. matin condensation and gene silencing and CpG methyl- CpG promoter hypermethylation was observed ation is inherited by the daughter cells after cell division. throughout the genome of colorectal cancer cell lines Aberrant promoter methylation leading to the epigenetic and primary tumors. However, only the expression of a silencing of tumor suppressor genes is an important subset of 382 of the 9,876 genes investigated (3.9%) was Bazzocco et al. Clin Epigenet (2021) 13:88 Page 9 of 13 negatively associated with the levels of promoter methyl- could contribute to the oncogenic process. As a proof of ation, thus constituting a core set of genes that are poten- concept, we investigated the role of zinc finger and BTB tially regulated by this epigenetic process in colorectal domain containing 18 (ZBTB18/ZNF238/RP58) in colo- cancer (Additional file 1: Table S2). As expected, these rectal tumorigenesis. ZBTB18 binds the consensus DNA genes included important tumor suppressor genes known sequence 5′-(A/C)ACA​TCT​G(G/T)(A/C)-3′, which con- to be frequently silenced through promoter methylated tains the E box core, and acts by recruiting chromatin in colorectal tumors, such as the cell adhesion protein remodeling multiprotein complexes such as histone dea- E-cadherin (CDH1) [22], carbonic anhydrase IX (CA9) cetylases, and repressing the expression of downstream [23–25], the serine/threonine protein kinase DAPK2 [26], genes. ZBTB18 plays a crucial role in supporting myo- the brush border Myosin Ia (MYO1A) [27] and the mis- genesis [39] and brain development [40, 41], although its match repair gene MLH1 [28]. In addition, these analy- role in the development of normal intestinal epithelium ses identified other genes with known tumor suppressor has not been investigated. The expression of ZBTB18 has activity in colorectal cancer that have not been previ- been shown to be negatively regulated by methylation of ously reported to be regulated by promoter methylation the CpG island in its promoter region [42] and to reduce in colorectal tumors. These included the chloride chan- proliferation as well as to promote apoptosis in brain nel that is most frequently mutated in patients with cystic tumors such as medulloblastoma (MB) and glioblas- fibrosis (CFTR) [29, 30] and the negative regulator of toma multiforme (GBM) [42, 43]. Moreover, the loss of Wnt signaling TCF7 [31]. ZBTB18 has been shown to contribute to the aggressive Methylation of CG dinucleotides in CpG islands and phenotype of GBM through regulation of poor progno- CpG island ‘shores’ is known to regulate gene expression sis-associated signatures [42]. However, the role of this [7]. Although the effect of DNA methylation outside the zinc finger protein in colorectal tumorigenesis has not context of a CpG island has not been as thoroughly inves- been investigated. Although ZBTB18 is not frequently tigated, CpG methylation in the promoter of genes with- mutated in colorectal cancer (0.3–2.5% in large exome out a CpG island has previously been shown to be able sequencing studies [3, 44, 45]), here we found that the to regulate gene expression, often by interfering with the expression of ZBTB18 was downregulated by promoter binding of transcriptional regulators [7, 27, 32, 33]. Here, methylation in colorectal cancer cells and that its over- we found that approximately 20% of the 382 genes with expression reduced their growth both in vitro and in sub- expression/methylation correlations are not associated cutaneous xenograft mouse models. Therefore, strategies with the presence of a CpG island near their promoter, aimed at rescuing ZBTB18 expression or downregulating suggesting that CpG methylation in the promoter of some of the downstream ZBTB18 targets genes may have genes not associated with CpG islands may play a signifi- therapeutic potential in colorectal cancer patients. cant role in colorectal cancer tumorigenesis. The expression of several zinc finger proteins, such as The list of 382 genes showing a negative association ZEB1 [46], ZFX [47] and PRDM1 [37], has been reported between the levels of gene expression and promoter to be associated with the survival of colorectal cancer methylation was significantly enriched in zinc finger patients. Moreover, an association has been reported proteins (FDR = 9.2 × ­10–6). Zinc finger proteins consti- between high levels of ZBTB18 promoter methylation tute a large and diverse family of proteins that contain and shorter progression-free survival of patients with functional domains that require coordination by at least glioblastomas [42]. Here, a validated anti-ZBTB18 anti- one zinc ion [34]. Most zinc finger proteins interact with body was used to assess the expression of this zinc fin- DNA in a sequence-specific manner and regulate gene ger protein in normal colonic samples as well as locally transcription, although many family members regulate advanced primary colorectal tumors and lymph node other cellular functions through its capacity to bind to metastases. While high ZBTB18 expression was observed RNA, PAR (poly-ADP-ribose) and other proteins [35]. in normal colonic epithelial cells, a significant reduction Some zinc finger proteins have been shown to partici- of ZBTB18 immunostaining was frequently observed in pate in colorectal tumorigenesis including ZBP89 [36], primary colorectal tumors. Reduced ZBTB18 expres- PRDM1 [37] and ZNF545 [38]. Here, we hypothesized sion in the primary tumors was associated with shorter that at least some of the 382 genes showing negative patient survival and the average expression of ZBTB18 correlations between the levels of expression and meth- was further reduced in lymph node metastases com- ylation are silenced by this epigenetic mechanism and pared to primary colorectal tumors. These findings are Bazzocco et al. Clin Epigenet (2021) 13:88 Page 10 of 13 consistent with the tumor suppressor activity observed Determination of cell growth when ZBTB18 was overexpressed in colon cancer cells, 3 × ­104 HT29, HCT116 or derivative cells were seeded and indicate that the levels of expression of this tran- in triplicate wells of seven 24-well plates and allowed to scriptional repressor could be useful to identify a subset adhere overnight. Cells were trypsinized and stained with of locally advanced colorectal cancer patients with poor trypan blue at the indicated times. Growth curves pre- prognosis that are good candidates to receive a more sented were plotted with Prism (GraphPad Software). aggressive form of treatment and/or closer follow up. In summary, we used here transcriptomic and methyl- Quantification of mRNA expression omic genome-wide analysis of colorectal cancer cell lines Microarray analysis was used for genome-wide expres- and primary colorectal tumors to identify a subset of 382 sion profiling of 30 colorectal cell lines (Human Genome genes with a negative association between expression and U133 Plus 2.0 Array; Affymetrix, Santa Clara, CA; methylation levels and that likely include epigenetically ArrayExpress E-MTAB-2971) and 223 primary colo- regulated genes driving the oncogenic process. Zinc fin- rectal tumors (Agilent 244 K Custom Gene Expression ger proteins were significantly enriched in this list of 382 G4502A-07–1/2/3) [3]. Relative mRNA levels of selected genes with expression/methylation associations, and we genes were assessed by Real-Time PCR using SYBR demonstrated the tumor suppressor activity of the zinc Green Master Mix (Applied Biosystems, Branchburg, NJ) finger transcriptional repressor ZBTB18 in colon cancer as described before [21]. Primers sequence is available in cells. Consistently, reduced ZBTB18 expression in locally Additional file 1: Table S1. advanced primary colorectal tumors was associated with shorter patient survival. The role of other genes in the Assessment of DNA methylation core set of 382 genes with expression/methylation asso- Methylation at the single-nucleotide resolution level for a ciations on colorectal tumorigenesis remains to be char- total of 27,578 CpG sites was determined using Human- acterized in future investigations. Methylation27 Beadchip (Illumina, San Diego, CA) for a panel of 45 colorectal cancer cell lines (See Supplemen- Methods tary Methods; ArrayExpress MTAB-7867). Promoter Cell lines methylation data for a cohort of 223 primary colorectal A total of 45 colorectal cancer cell lines were used (see tumors (Illumina HumanMethylation27 Beadchip) was Additional file 1). All lines were maintained in MEM (Life available from The Cancer Genome Atlas (TCGA) [3]. In Technologies, Carlsbad, CA) supplemented with 10% this study, tumors from all stages (American Joint Com- fetal bovine serum, 1 × antibiotic/antimycotic, 1 × MEM mittee on Cancer staging system) and histologic grade Non-Essential Amino Acids Solution, and 10 mM HEPES were collected from newly diagnosed patients with colon buffer solution (all from Life Technologies, Carlsbad, CA) or rectum adenocarcinomas undergoing surgical resec- at 37 °C and 5% C ­ O2. All lines were tested to be nega- tion. None of the patients received neoadjuvant chemo- tive for mycoplasma contamination (PCR Mycoplasma therapy or radiotherapy. For bisulfite sequencing analysis, Detection Set, Takara) and possible cell line cross- genomic DNA was bisulfite treated with EZ DNA Meth- contamination was ruled out by clustering analysis of ylation-Gold Kit (Zymo Research, Irvine, CA) following genome-wide mRNA expression and promoter methyla- the manufacturer’s instructions and then PCR-amplified. tion data at the time of these experiments. Primers used are listed in Additional file 1: Table S1. ZBTB18 overexpression in colon cancer cells Associations between mRNA expression and promoter HCT116 cells were transfected with pIRES2-EGFP or methylation levels pZBTB18-IRES-EGFP with Lipofectamine 2000 (Thermo Using the gene symbol as a common identifier, we found Fisher Scientific, Waltham, MA). For tetracycline- that expression data were available for a total of 11,359 inducible overexpression of ZBTB18, HT29 cells were (81.92%) of the 14,475 promoters interrogated in the transduced with pINDUCER20-EGFP or pINDUCER20- HumanMethylation27 arrays. Significant correlations ZBTB18-IRES-EGFP as previously described [48] and between gene expression and promoter methylation were further described in the ‘Supplementary Materials’. identified using Spearman correlations (False Discovery Rate (FDR) < 0.2). Genes showing a nonlinear negative Bazzocco et al. Clin Epigenet (2021) 13:88 Page 11 of 13 association between expression and methylation lev- Immunohistochemistry els (i.e., ‘L-shaped’ in the scattered plots; see Additional Tumor samples excised from mouse xenograft experi- file 1: Fig. S5) were identified using the ad hoc generated ments and tissue microarrays (TMAs) were immu- tool available at http://​cinna.​upc.​edu:​3838/​alex/​Lheur​ nostained for ZBTB18 protein detection using a istic/ (additional details described in [49]). rabbit polyclonal anti-ZBTB18 antibody at 1:1000 dilu- tion (Biorbyt, Cambridge, UK; cat# orb357631). Gene set enrichment analysis To investigate whether there was an enrichment in the Abbreviations number of genes with significant correlation between MSI: Microsatellite instable; MSS: Microsatellite stable; CIMP: CpG island expression and methylation levels that belonged to differ- methylator phenotype; MEM: Minimum essential medium; PCR: Polymer‑ ent categories of functionally related genes, we used the ase chain reaction; TCGA​: The Cancer Genome Atlas; NOD/SCID: Nonobese diabetic/severe combined immunodeficiency; FDR: False discovery rate; EGFP: Database for Annotation, Visualization and Integrated Enhanced green fluorescent protein. Discovery (DAVID) v6.7 [50]. Supplementary Information Subcutaneous xenograft mouse model The online version contains supplementary material available at https://​doi.​ org/​10.​1186/​s13148-​021-​01070-0. Subcutaneous tumors were established in NOD/SCID mice by injection HCT116 or HT29 derivative cells Additional file 1: Supplementary Figures: Supplementary Figure S1. resuspended in 100 μl sterile PBS. All animals bearing CpG methylation levels assessed by HumanMethylation27 beadChips tumors from HCT116 derivative cells received 1 mg/ml (Illumina) and direct bisulfite sequencing. Supplementary Figure S2. doxycycline (Sigma) in the drinking water and tumor CpG methylation in primary colorectal tumors. Supplementary Figure S3. Associations between the average levels of methylation and molecular size followed over time. At the experimental end point, features of primary colorectal tumors and cell lines. Supplementary all animals were euthanized, tumors were surgically Figure S4. Representative examples of genes showing significant correla‑ removed, and the tumor weight was determined. Tumor tions between mRNA and methylation levels. Supplementary Figure S5. representative genes showing a non-linear negative association between specimens were formalin-fixed and paraffin-embedded mRNA expression and methylation levels. Supplementary Figure S6. for histological analysis. All experiments were carried out Validation of the differences in gene expression observed by microarray under observance of the protocol approved by the Ethi- analysis. Supplementary Figure S7. Effects of decitabine treatment on the expression of genes with promoter methylation. Supplementary Fig‑ cal Committee for Animal Experimentation from the Vall ure S8. Effects of DNMT (Decitabine and Ladakamycin), HDAC (Vorinostat) d’Hebron University Hospital Institute of Research (Bar- KDM (OG-L002) and KMT (Tazemetostat) inhibitors on the expression of celona, Spain). ZBTB18. Supplementary Figure S9. Validation of ZBTB18 overexpression in colon cancer cell line systems. Supplementary Figure S10. Effects of ZBTB18 overexpression on the growth of HT29 colon cancer cells. Sup‑ Determination of cell cycle and apoptosis plementary Figure S11. ZBTB18 antibody validation. Supplementary Cell cycle and hypodiploid (subG1) cells were quantified Figure S12. ZBTB18 protein expression in colorectal cell lines. Sup‑ plementary Figure S13. Survival of Stage III colorectal cancer patients by flow cytometry after propidium iodide staining. Full as a function of ZBTB18 mRNA expression. Supplementary Tables: procedure is detailed in the ‘Supplementary Materials’. Supplementary Table S1. PCR primer s used in this study. Supplemen‑ tary Table S2. Details of the 382 gene s showing significant correlation between expression and methilation levels in 30 colorectal cell lines Cell migration and invasion assays and 223 primary tumors. Supplementary Table S3. Group enrichment To investigate changes in the motility and invasion of analysis: Functional group enrichment analysis carried out with DAVID HCT116 and HT29 cells upon ZBTB18 upregulation, (https://​doi.​org/​10.​1186/​gb-​2003-4-​9-​r60). Supplementary Table S4 Clinicopathological features of the 132 Dukes C colorectal cancer patients a wound healing and a Matrigel Boyden chamber assay in this study as a function of ZBTB18 protein expression as determined were used, respectively [48]. Further details are described by immunohistochemistry. Supplementary Table S5 Clinicopathologi‑ in Additional file 1: ‘Supplementary Materials’. cal features of the 55 colorectal Stage III cancer patients from the TCGA as a function of ZBTB 18 mRNA expression as determined by microarray analysis (Agilent). Supplementary Methods. Tissue microarrays analysis TMAs from colorectal cancer cell lines (Supplementary Acknowledgements Methods) and colorectal cancer patients were used [51]. RN is supported by a PERIS (Pla Estratègic de Recerca i Innovació en Salut) Briefly, 30 colorectal cancer cell lines, 75 samples of nor- fellowship from Generalitat de Catalunya Health Department, and EG-V is sup‑ mal colon tissues, 132 primary adenocarcinomas from ported by Ayudas para la promoción de empleo joven e implantación de la garantía juvenil en I+D+I from the Spanish Ministry of Health and Innovation. colon and rectum and 15 lymph node metastases from EA, LJ and PGM are recipients of fellowships by the Agència de Gestió d’Ajuts Dukes’ C Finnish colorectal cancer patients were forma- Universitaris i de Recerca (AGAUR), Conselho Nacional de Desenvolvimento lin-fixed, paraffin-embedded and arrayed in triplicate. Científico e Tecnológico (CNPq) and the Chinese Scholarship Council (CSC), respectively. Bazzocco et al. Clin Epigenet (2021) 13:88 Page 12 of 13 Authors’ contributions 6. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. DA conceived experiments, analyzed data and wrote the manuscript. SB, HD, Nat Rev Genet. 2002;3(6):415–28. AM-B, EA, RN, VM, EG-V, JL, PR, PGM, JMM carried out experiments and ana‑ 7. Irizarry RA, Ladd-Acosta C, Wen B, Wu Z, Montano C, Onyango P, Cui lyzed data. SB, HD and AM-B participated in writing the manuscript. SS, LAA, H, Gabo K, Rongione M, Webster M, et al. The human colon cancer AS and JMM analyzed data and revised the manuscript. methylome shows similar hypo- and hypermethylation at conserved tissue-specific CpG island shores. Nat Genet. 2009;41(2):178–86. Funding 8. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human This study was partially funded by grants of the Instituto de Salud Car‑ cancers. Nature. 1998;396(6712):643–9. los III and the European Regional Development Fund (ERDF; PI16/00540, 9. Mouradov D, Sloggett C, Jorissen RN, Love CG, Li S, Burgess AW, Arango PI19/00993, AC20/00022 and AC15/00066) and the Spanish Association D, Strausberg RL, Buchanan D, Wormald S, et al. Colorectal cancer cell Against Cancer (AECC GCA15152966ARAN) to Diego Arango. lines are representative models of the main molecular subtypes of primary cancer. Cancer Res. 2014;74(12):3238–47. Availability of data and material 10. Weisenberger DJ, Siegmund KD, Campan M, Young J, Long TI, Faasse MA, Microarray and methylation data of the cell lines are available in ArrayExpress Kang GH, Widschwendter M, Weener D, Buchanan D, et al. CpG island E-MTAB-2971 and MTAB-7867. methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nat Genet. 2006;38(7):787–93. Declarations 11. Boland CR, Goel A. Microsatellite instability in colorectal cancer. Gastroen‑ terology. 2010;138(6):2073–87. Ethics approval 12. van Rijnsoever M, Grieu F, Elsaleh H, Joseph D, Iacopetta B. Characterisa‑ All experiments were carried out under observance of the protocol approved tion of colorectal cancers showing hypermethylation at multiple CpG by the Ethical Committee for Animal Experimentation from the Vall d’Hebron islands. Gut. 2002;51(6):797–802. University Hospital Institute of Research (Barcelona, Spain). Informed consent 13. Hawkins NJ, Ward RL. Sporadic colorectal cancers with microsatellite for genetic analysis of the human tumor samples was obtained from each instability and their possible origin in hyperplastic polyps and serrated patient according to the Human Investigations and Ethical Committee– adenomas. J Natl Cancer Inst. 2001;93(17):1307–13. approved research proposal. 14. Simons CCJM, Hughes LAE, Smits KM, Khalid-de Bakker CA, de Bruïne AP, Carvalho B, Meijer GA, Schouten LJ, van den BrandT PA, Weijenberg MP, Consent for publication et al. A novel classification of colorectal tumors based on microsatellite Not applicable. instability, the CpG island methylator phenotype and chromosomal insta‑ bility: implications for prognosis. Ann Oncol. 2013;8:2048–56. Competing interests 15. IJspeert JEG, Vermeulen L, Meijer GA, Dekker E. Serrated neoplasia-role in The authors declare that they have no competing interests. colorectal carcinogenesis and clinical implications. Nat Rev Gastroenterol Hepatol. 2015;12(7):401–9. Author details 16. Gardiner-Garden M, Frommer M. CpG islands in vertebrate genomes. J 1 Group of Biomedical Research in Digestive Tract Tumors, CIBBIM‑Nanomed‑ Mol Biol. 1987;196(2):261–82. icine, Vall d’Hebron University Hospital, Research Institute (VHIR), Universitat 17. Zhang B, Wang J, Wang X, Zhu J, Liu Q, Shi Z, Chambers MC, Zimmerman Autònoma de Barcelona, Passeig Vall d’Hebron, 119‑129, 08035 Barcelona, LJ, Shaddox KF, Kim S, et al. Proteogenomic characterization of human Spain. 2 Group of Drug Delivery and Targeting, CIBBIM‑Nanomedicine, Vall colon and rectal cancer. Nature. 2014;513(7518):382–7. d’Hebron University Hospital, Research Institute (VHIR), Universitat Autònoma 18. Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB, Issa JP. CpG de Barcelona, Passeig Vall d’Hebron, 119‑129, 08035 Barcelona, Spain. 3 CIBER island methylator phenotype in colorectal cancer. Proc Natl Acad Sci USA. de Bioingeniería, Biomateriales Y Nanomedicina (CIBER-BBN), Madrid, Spain. 1999;96(15):8681–6. 4 Department of Medical Genetics, Medicum, University of Helsinki, Biomedi‑ 19. Hawkins N, Norrie M, Cheong K, Mokany E, Ku SL, Meagher A, O’Connor cum Helsinki, 00290 Helsinki, Finland. 5 Departament d’Estadísitica, Facultat de T, Ward R. CpG island methylation in sporadic colorectal cancers Biologia, Universitat de Barcelona, 08028 Barcelona, Spain. 6 Olivia Newton- and its relationship to microsatellite instability. Gastroenterology. John Cancer Research Institute, Heidelberg, VIC 3084, Australia. 7 School 2002;122(5):1376–87. of Cancer Medicine, La Trobe University, Melbourne 3086, Australia. 8 Group 20. Samowitz WS, Slattery ML, Sweeney C, Herrick J, Wolff RK, Albertsen of Molecular Oncology, IRBLleida, 25198 Lleida, Spain. H. APC mutations and other genetic and epigenetic changes in colon cancer. Mol Cancer Res. 2007;5(2):165–70. Received: 17 July 2019 Accepted: 7 April 2021 21. Bazzocco S, Dopeso H, Carton-Garcia F, Macaya I, Andretta E, Chionh F, Rodrigues P, Garrido M, Alazzouzi H, Nieto R, et al. Highly expressed genes in rapidly proliferating tumor cells as new targets for colorectal cancer treatment. Clin Cancer Res. 2015;21(16):3695–704. 22. Wheeler JM, Kim HC, Efstathiou JA, Ilyas M, Mortensen NJ, Bodmer WF. References Hypermethylation of the promoter region of the E-cadherin gene (CDH1) 1. Martinez-Barriocanal A, Arango D, Dopeso H. PVT1 long non-coding RNA in sporadic and ulcerative colitis associated colorectal cancer. Gut. in gastrointestinal cancer. Front Oncol. 2020;10:38. 2001;48(3):367–71. 2. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global 23. Nakamura J, Kitajima Y, Kai K, Hashiguchi K, Hiraki M, Noshiro H, Miyazaki cancer statistics 2018: GLOBOCAN estimates of incidence and mor‑ K. Expression of hypoxic marker CA IX is regulated by site-specific DNA tality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. methylation and is associated with the histology of gastric cancer. Am J 2018;68(6):394–424. Pathol. 2011;178(2):515–24. 3. Cancer Genome Atlas Network. Comprehensive molecular characteriza‑ 24. Jakubickova L, Biesova Z, Pastorekova S, Kettmann R, Pastorek J. Methyla‑ tion of human colon and rectal cancer. Nature. 2012;487(7407):330–7. tion of the CA9 promoter can modulate expression of the tumor-asso‑ 4. Jung G, Hernandez-Illan E, Moreira L, Balaguer F, Goel A. Epigenetics of ciated carbonic anhydrase IX in dense carcinoma cell lines. Int J Oncol. colorectal cancer: biomarker and therapeutic potential. Nat Rev Gastro‑ 2005;26(4):1121–7. enterol Hepatol. 2020;17(2):111–30. 25. Zatovicova M, Jelenska L, Hulikova A, Csaderova L, Ditte Z, Ditte P, Golias‑ 5. Feinberg AP, Vogelstein B. Hypomethylation distinguishes genes ova T, Pastorek J, Pastorekova S. Carbonic anhydrase IX as an anticancer of some human cancers from their normal counterparts. Nature. therapy target: preclinical evaluation of internalizing monoclonal anti‑ 1983;301(5895):89–92. body directed to catalytic domain. Curr Pharm Des. 2010;16(29):3255–63. Bazzocco et al. Clin Epigenet (2021) 13:88 Page 13 of 13 26. Mittag F, Kuester D, Vieth M, Peters B, Stolte B, Roessner A, Schneider- 40. Okado H, Ohtaka-Maruyama C, Sugitani Y, Fukuda Y, Ishida R, Hirai S, Miwa Stock R. DAPK promotor methylation is an early event in colorectal A, Takahashi A, Aoki K, Mochida K, et al. The transcriptional repressor RP58 carcinogenesis. Cancer Lett. 2006;240(1):69–75. is crucial for cell-division patterning and neuronal survival in the develop‑ 27. Mazzolini R, Dopeso H, Mateo-Lozano S, Chang W, Rodrigues P, Bazzocco ing cortex. Dev Biol. 2009;331(2):140–51. S, Alazzouzi H, Landolfi S, Hernandez-Losa J, Andretta E, et al. Brush 41. Xiang C, Baubet V, Pal S, Holderbaum L, Tatard V, Jiang P, Davuluri RV, border myosin Ia has tumor suppressor activity in the intestine. Proc Natl Dahmane N. RP58/ZNF238 directly modulates proneurogenic gene levels Acad Sci U S A. 2012;109(5):1530–5. and is required for neuronal differentiation and brain expansion. Cell 28. Kane MF, Loda M, Gaida GM, Lipman J, Mishra R, Goldman H, Jessup JM, Death Differ. 2012;19(4):692–702. Kolodner R. Methylation of the hMLH1 promoter correlates with lack of 42. Fedele V, Dai F, Masilamani AP, Heiland DH, Kling E, Gatjens-Sanchez expression of hMLH1 in sporadic colon tumors and mismatch repair- AM, Ferrarese R, Platania L, Soroush D, Kim H, et al. Epigenetic regula‑ defective human tumor cell lines. Cancer Res. 1997;57(5):808–11. tion of ZBTB18 promotes glioblastoma progression. Mol Cancer Res. 29. Than BLN, Linnekamp JF, Starr TK, Largaespada DA, Rod A, Zhang Y, 2017;15(8):998–1011. Bruner V, Abrahante J, Schumann A, Luczak T, et al. CFTR is a tumor 43. Tatard VM, Xiang C, Biegel JA, Dahmane N. ZNF238 is expressed in suppressor gene in murine and human intestinal cancer. Oncogene. postmitotic brain cells and inhibits brain tumor growth. Cancer Res. 2017;36(24):3504. 2010;70(3):1236–46. 30. Koh J, Sferra TJ, Collins FS. Characterization of the cystic fibrosis trans‑ 44. Seshagiri S, Stawiski EW, Durinck S, Modrusan Z, Storm EE, Conboy CB, membrane conductance regulator promoter region. Chromatin context Chaudhuri S, Guan Y, Janakiraman V, Jaiswal BS, et al. Recurrent R-spondin and tissue-specificity. J Biol Chem. 1993;268(21):15912–21. fusions in colon cancer. Nature. 2012;488(7413):660–4. 31. Jung K, Yoon K, Song J, Lee S, Eun J, Noh J, Kim J, Bae H, Lee J, Kim S, et al. 45. Giannakis M, Mu XJ, Shukla SA, Qian ZR, Cohen O, Nishihara R, Bahl S, Cao Loss-of-function mutations in the Transcription Factor 7 (T cell factor-1) Y, Amin-Mansour A, Yamauchi M, et al. Genomic correlates of immune- gene in hepatogastrointestinal cancers. Mol Cell Toxicol. 2010;6:271–8. cell infiltrates in colorectal carcinoma. Cell Rep. 2016;17(4):1206. 32. Han H, Cortez CC, Yang X, Nichols PW, Jones PA, Liang G. DNA meth‑ 46. Zhang GJ, Zhou T, Tian HP, Liu ZL, Xia SS. High expression of ZEB1 corre‑ ylation directly silences genes with non-CpG island promoters and lates with liver metastasis and poor prognosis in colorectal cancer. Oncol establishes a nucleosome occupied promoter. Hum Mol Genet. Lett. 2013;5(2):564–8. 2011;20(22):4299–310. 47. Yan X, Yan L, Su Z, Zhu Q, Liu S, Jin Z, Wang Y. Zinc-finger protein X-linked 33. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies is a novel predictor of prognosis in patients with colorectal cancer. Int J and beyond. Nat Rev Genet. 2012;13(7):484–92. Clin Exp Pathol. 2014;7(6):3150–7. 34. Laity JH, Lee BM, Wright PE. Zinc finger proteins: new insights into struc‑ 48. Rodrigues P, Macaya I, Bazzocco S, Mazzolini R, Andretta E, Dopeso H, tural and functional diversity. Curr Opin Struct Biol. 2001;11(1):39–46. Mateo-Lozano S, Bilic J, Carton-Garcia F, Nieto R, et al. RHOA inactivation 35. Cassandri M, Smirnov A, Novelli F, Pitolli C, Agostini M, Malewicz M, enhances Wnt signalling and promotes colorectal cancer. Nat Commun. Melino G, Raschella G. Zinc-finger proteins in health and disease. Cell 2014;5:5458. Death Discov. 2017;3:17071. 49. Sánchez A, Miró B, Castellano P, Carmona F, Bazzocco S, Arango D. A 36. Essien BE, Sundaresan S, Ocadiz-Ruiz R, Chavis A, Tsao AC, Tessier AJ, heuristic algorithm to select L-shaped genes, potentially regulated by Hayes MM, Photenhauer A, Saqui-Salces M, Kang AJ, et al. Transcription methylation. Submitted. 2020. factor ZBP-89 drives a feedforward loop of beta-catenin expression in 50. da Huang W, Sherman BT, Lempicki RA. Systematic and integrative analy‑ colorectal cancer. Cancer Res. 2016;76(23):6877–87. sis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 37. Liu C, Banister CE, Weige CC, Altomare D, Richardson JH, Contreras CM, 2009;4(1):44–57. Buckhaults PJ. PRDM1 silences stem cell-related genes and inhibits 51. Arango D, Laiho P, Kokko A, Alhopuro P, Sammalkorpi H, Salovaara R, proliferation of human colon tumor organoids. Proc Natl Acad Sci USA. Nicorici D, Hautaniemi S, Alazzouzi H, Mecklin JP, et al. Gene-expression 2018;115(22):E5066–75. profiling predicts recurrence in Dukes’ C colorectal cancer. Gastroenterol‑ 38. Xiang S, Xiang T, Xiao Q, Li Y, Shao B, Luo T. Zinc-finger protein 545 is ogy. 2005;129(3):874–84. inactivated due to promoter methylation and functions as a tumor sup‑ pressor through the Wnt/beta-catenin, PI3K/AKT and MAPK/ERK signaling pathways in colorectal cancer. Int J Oncol. 2017;51(3):801–11. Publisher’s Note 39. Yokoyama S, Ito Y, Ueno-Kudoh H, Shimizu H, Uchibe K, Albini S, Mitsuoka Springer Nature remains neutral with regard to jurisdictional claims in pub‑ K, Miyaki S, Kiso M, Nagai A, et al. A systems approach reveals that the lished maps and institutional affiliations. myogenesis genome network is regulated by the transcriptional repres‑ sor RP58. Dev Cell. 2009;17(6):836–48. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions