Systemic genome-epigenome analysis captures a lineage-specific super-enhancer for MYB in gastrointestinal adenocarcinoma

Abstract

Gastrointestinal adenocarcinoma is a major cancer type for the digestive system, ranking as the top cause of cancer-related deaths worldwide. While there has been extensive research on mutations in protein-coding regions, the knowledge of the landscape of its non-coding regulatory elements is still insufficient. Combining the analysis of active enhancer profiles and genomic structural variation, we discovered and validated a lineage-specific super-enhancer for MYB in gastrointestinal adenocarcinoma. This super-enhancer is composed of a predominant enhancer e4 and several additional enhancers, whose transcriptional activity is regulated by the direct binding of HNF4A and MYB itself. Suppression of the super-enhancer downregulated the expression of MYB, inhibited downstream Notch signaling and prevented the development of gastrointestinal adenocarcinoma both in vitro and in vivo. Our study uncovers a mechanism driven by non-coding variations that regulate MYB expression in a lineage-specific manner, offering new insights into the carcinogenic mechanism and potential therapeutic strategies for gastrointestinal adenocarcinoma.

Synopsis

By combining the analysis of active enhancer profiles and genomic structural variation, a lineage-specific super enhancer for MYB in gastrointestinal adenocarcinoma was discovered and validated.

• MYB-high/-dependent gastrointestinal adenocarcinoma cells harbor a MYB super-enhancer.

• HNF4A and MYB bind to the enhancer e4 within MYB super-enhancer.

• MYB activation confers a dependency of gastrointestinal adenocarcinoma cell on MYB.

By combining the analysis of active enhancer profiles and genomic structural variation, a lineage-specific super enhancer for MYB in gastrointestinal adenocarcinoma was discovered and validated.

Introduction

The overexpression of oncogenes in patients contributes to the development of cancer, with some of these oncogenes becoming therapeutic targets and inspiring the pharmaceutical industry to develop better cancer therapies (Delmore et al, 2011; Vicente-Duenas et al, 2013; Yang et al, 2024). It has been validated that the variations of non-coding regions in the human genome have significant impacts on health (Groschel et al, 2014; Kataoka et al, 2016; Mansour et al, 2014; Northcott et al, 2014; Peifer et al, 2015). For example, a partial loss of the 3′ UTR of the PD-L1 gene is associated with the upregulation of PD-L1 expression in various cancer types, including T cell leukemia and B cell lymphoma (Kataoka et al, 2016). Somatic mutations in the TAL1 enhancer can introduce de novo MYB binding motifs, leading to the formation of a super-enhancer that drives aberrant TAL1 expression (Mansour et al, 2014). Transcriptional regulation by non-coding regions is highly lineage-specific (Madani Tonekaboni et al, 2019). Thus, further research is needed to better understand the role of non-coding regions in cancer development and progression and to uncover new therapeutic targets for individual cancer type. Gastrointestinal adenocarcinoma is a major cancer type of the digestive system, ranking as the top cause of cancer-related deaths worldwide (Bray et al, 2024). However, knowledge of its lineage-specific variations in non-coding regions remains limited.

Dysregulation of transcriptional programs can be driven by epigenetic alterations targeting non-coding regulatory elements such as enhancers and super-enhancers (SEs) (Leeman-Neill et al, 2023; Zhang and Meyerson, 2020; Zhou and Parsons, 2023). In cancer research, SEs are described as clusters of enhancers in close proximity, marked by H3K27ac, transcription factors or mediator complex to drive the high expression of lineage-specific oncogenes (Hnisz et al, 2013). It has been demonstrated that the recurrent alterations of SEs play an important role in tumorigenesis and malignant phenotypes by interacting with the corresponding promoters via the three-dimensional (3D) structure of genome (Liu et al, 2021; Xing et al, 2019; Zhang et al, 2018; Zhang et al, 2016). For example, SE amplifications increase the expression of oncogenes including MYC, SOX2, USP12, KLF5, ZFP36L2, thereby promoting the progression of various cancers, including endometrioma and gastrointestinal tumors (Liu et al, 2021; Xing et al, 2019; Zhang et al, 2018; Zhang et al, 2016). Considering the diversity of different types of cancer, significant efforts are demanded to understand the transcriptional regulation of oncogenes by the SEs in gastrointestinal adenocarcinoma, especially in colorectal cancer (CRC) and gastric cancer.

MYB is a protein-coding gene, which encodes a protein with a highly conserved DNA-binding domain (Ciciro and Sala, 2021). As a transcription factor, it regulates the expression of several oncogenes, including MYC (Quintana et al, 2011), TAL1 (Mansour et al, 2014) and IGF1R (Andersson et al, 2017). The overexpression of MYB is widely observed in cancer cells and it’s oncogenic activity has been confirmed in leukemia (Pattabiraman and Gonda, 2013), breast cancer (Quintana et al, 2011), prostate cancer (Acharya et al, 2023), and other cancers (Ramsay and Gonda, 2008). Most known mechanisms underlying MYB overexpression involve gene duplication (Lahortiga et al, 2007), translocation (Belloni et al, 2011; Clappier et al, 2007) or intragenic mutations (Hugo et al, 2006). However, the regulation of MYB expression by non-coding variation in human genome has not been fully investigated in gastrointestinal adenocarcinoma.

In this study, we identified a lineage-specific SE in gastrointestinal adenocarcinoma that robustly enhances the expression of MYB and controls the development of tumor. Within this SE, a predominant enhancer e4 was discovered. The activity of e4 is regulated by the direct binding of HNF4A and MYB and it has a joint effect with other enhancers within this SE. Suppressing the activity of e4 significantly reduced MYB expression, inhibited the Notch signaling, decreased the cell viability of CRC and gastric cancer cells and prevented the growth of tumor xenograft. Our study revealed a non-coding variation-based mechanism to affect MYB expression in a lineage-specific manner, providing inspiring insights into the carcinogenic mechanism and potential therapeutic strategies for gastrointestinal adenocarcinoma.

Results

A lineage-specific super-enhancer near the MYB gene is duplicated in gastrointestinal adenocarcinoma

SEs are acquired by cancer cells at key oncogenes (Hnisz et al, 2013). We investigated the landscape of active enhancers in primary tumors and cell lines of colon and rectum adenocarcinoma (COREAD) and stomach adenocarcinoma (STAD) by analyzing the data of chromatin immunoprecipitation-sequencing (ChIP-seq) based on H3K27ac, a marker for active promoters and enhancers, from various datasets deposited in Gene Expression Omnibus (GEO) (Cohen et al, 2017; Okabe et al, 2020; Ooi et al, 2020; Orouji et al, 2022; Zhang et al, 2018). We identified 3433 unique SEs in cancer cell lines and 2137 unique SEs in tumor samples, respectively (Fig. 1A). Considering the shared lineage between colon and stomach, we ought to focus on the SEs common to both COREAD and STAD. Cell lines and tumors shared 82 and 185 SEs, respectively, which are associated with 205 unique protein-coding genes (Fig. 1B and Table EV1), including 34 oncogenes (Fig. 1C). Gene Ontology (GO) (Gene Ontology, 2015) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa and Goto, 2000) pathway analyses revealed that these genes were predominantly involved in transcriptional regulation and cancer-related process (Fig. 1D; Appendix Fig. S1A). Furthermore, Reactome pathway analysis showed that these genes were involved in canonical cancer-related signaling pathways, such as Notch and TGF-β signaling (Appendix Fig. S1B). These findings suggested that the common SEs may play an important role in the development of gastrointestinal adenocarcinoma.

Figure 1. A super‐enhancer near MYB locus is duplicated in gastrointestinal adenocarcinoma.

(A) Hockey stick plots for super-enhancer profiles of COREAD and STAD in cancer cell lines. The ranking was calculated by the H3K27ac ChIP-seq signal enrichment. The super-enhancers associated with MYB were highlighted in red. (B) Venn diagram showing the shared super-enhancers of various gastrointestinal cancer cell lines or tumor samples, respectively. (C) Pie chart showing oncogene classification of 205 SEs nominated protein-coding genes. (D) The enrichment and functional analysis of 205 SEs nominated protein-coding genes. BP: biological process. KEGG: Kyoto Encyclopedia of Genes and Genomes. The P-value was determined by Benjamini correction method. (E) Pie chart showing categorization of 205 SEs nominated protein-coding genes based on enhancer-promoter interaction and expression. (F) The MYB amplicons across gastrointestinal adenocarcinoma samples from PCAWG project and merged enhancer profile from cell lines at the MYB-associated locus in gastrointestinal adenocarcinoma.

With their adjacent SEs, 44 genes had physical interactions with SEs based on H3K27ac HiChIP analysis in HT55 and SNU-719 cells (Table EV2) (Liu et al, 2020; Maestri et al, 2023) and were upregulated in COREAD and STAD based on RNA-seq analysis (Fig. 1E; Appendix Fig. S1C). Among them, 11 genes were previously annotated as oncogenes that have physical interactions with their SEs and were upregulated in COREAD and STAD (Fig. EV1A). As a common event of genomic structural variation in cancer cells, enhancer duplication often drives the abnormally high expression of oncogenes (Zhang and Meyerson, 2020). Among these 11 SEs, only the SEs associated with KLF5 and MYB were located within the genomic regions showing recurrent duplication, as observed in a high-resolution whole-genome sequencing dataset from the Pan-Cancer Atlas of Whole Genomes (PCAWG) Project (Consortium, 2020). The KLF5-SE duplication had been reported previously (Zhang et al, 2018).

Figure EV1. The gastrointestinal specificity of MYB-SE.

(A) Venn diagram showing the overlap between oncogenes associated with SEs and upregulated genes with physical interactions with SEs in gastrointestinal adenocarcinoma. (B) Venn diagram showing the shared SEs and their associated genes between gastrointestinal cancer cell lines and tumor samples, with oncogenes highlighted in red. (C) ChIP-seq signal tracks showing the enrichment of H3K27ac at the MYB-SE locus. The blue denoted COREAD tracks and red denoted STAD tracks; The cell line information provides a detailed description of the cell lines depicted in Fig. 1 F. (D) Focal duplications harboring MYB-SE, shorter than 1 Mb in length, were selected for further analysis across 37 cancer types in the PCAWG project. (E) H3K27ac ChIP-seq tracks at the MYB locus in cell lines representing each cancer type. ESAD: esophageal adenocarcinoma; BLCA: bladder cancer; PADD: pancreatic adenocarcinoma; LIHC: liver hepatocellular carcinoma; UCEC: uterine corpus endometrial carcinoma; LUAD: lung adenocarcinoma; BRCA: breast cancer; PRAD: prostate adenocarcinoma; LAML: acute myeloid leukemia. (F) ChIP-seq signal tracks of H3K27ac at the Myb and Myc loci in mouse CRC cell lines. The region of MYB-SE is conserved between humans and mice except the enhancer e6. However, the activity of human gastrointestinal adenocarcinoma cancer-specific super-enhancer of MYB (converted to mm10 genome) is repressed in this region in mouse CRC cell lines. (G) Expression levels of Myb and Myc in mouse CRC cell lines, based on RNA-seq data. (H) The expression of MYB in cancer tissues and normal tissues, from the TCGA and GTEx databases, was analyzed using Gene Expression Profiling Interactive Analysis 2 (GEPIA2) (Tang et al, 2019). The exact P-value were not provided in GEPIA2. The P-value was determined by two-sided Student’s t-test. ***P ≤ 0.001; Box plots represent the distribution of MYB expression across different groups. The center line within each box denotes the median. The lower and upper bounds of the box indicate the first quartile (Q1, 25th percentile) and third quartile (Q3, 75th percentile), respectively. The whiskers extend to the minimum and maximum values within 1.5× interquartile range (IQR) from Q1 and Q3. Individual points represent paired data and are connected by lines, while potential outliers beyond 1.5× IQR are displayed as separate dots. (I) The expression of MYB in COREAD or STAD samples and their corresponding pair-matched normal tissues. The value for the y-axis is log₂ transformed (count+1). N = 32 for COREAD or STAD. NS: no significance. Box plots represent the distribution of MYB expression across different groups. The center line within each box denotes the median. The lower and upper bounds of the box indicate the first quartile (Q1, 25th percentile) and third quartile (Q3, 75th percentile), respectively. The whiskers extend to the minimum and maximum values within 1.5× interquartile range (IQR) from Q1 and Q3. Individual points represent paired data and are connected by lines, while potential outliers beyond 1.5× IQR are displayed as separate dots. (J) Correlation between MYB expression and MYB copy number, calculated using Spearman’s rank correlation coefficient based on the data from CCLE.

MYB is an oncogene that is widely expressed in acute myeloid leukemia (AML) (Ciciro and Sala, 2021; Ramsay and Gonda, 2008). Especially, MYB is a gene with an SE present in both tumor samples and cell lines of gastrointestinal adenocarcinoma (Fig. EV1B). The SE (MYB-SE) within the amplicon near MYB is shown in Figs. 1F and EV1C. Although duplications containing MYB-SE were found in many cancer types, MYB-SE duplication occurs more frequently in gastrointestinal adenocarcinoma (Fig. EV1D). The information of duplication was derived from WGS data of tumor samples, for which the H3K27ac ChIP-seq data is not available. The information on H3K27ac was derived from cell lines, which have no corresponding WGS data. Therefore, we cannot determine whether the H3K27ac signal of MYB-SE is associated with its amplification.

It has been reported that the activity of SEs is specific to cell type and tumor type (Hnisz et al, 2013; Whyte et al, 2013). Therefore, we examined the lineage specificity of the candidate MYB-SE in other cancer types. Although the genomic region of the candidate MYB-SE was also duplicated in esophageal adenocarcinoma (ESAD), a cancer type closely related in lineage but with remarkable biological differences compared to COREAD and STAD, the ChIP-seq analysis revealed that it did not constitute a SE (Fig. EV1E). In addition, though MYB is considered as an oncogene in AML with a potential super-enhancer (Pelish et al, 2015), the candidate MYB-SE was not present in AML and many other cancer types (Fig. EV1E). We also tested the presence of MYB-SE in other species using a published dataset from mouse (Li et al, 2023). Compared to a well-known oncogene Myc, we observed a low H3K27ac ChIP-seq signal (Fig. EV1F) and low expression of Myb in mouse CRC cell lines (Fig. EV1G), suggesting that MYB-SE is likely specific to humans. We applied liftOver (Kent et al, 2002) to convert the human MYB-SE gene coordinates to mouse coordinates (mm10) and found that the DNA sequence underlying the MYB-SE is conserved between human and mouse except the enhancer e6 (Fig. EV1F).

Could the existence of the candidate MYB-SE impact the expression of MYB in gastrointestinal adenocarcinoma? Based on RNA-seq data of normal samples from the Genotype-Tissue Expression project (GTEx) (Carithers et al, 2015; Consortium, 2015) and tumor samples from the Cancer Genome Atlas (TCGA) (Cancer Genome Atlas Research et al, 2013), the expression of MYB in COAD, READ and STAD tumors was higher compared to normal tissue (Fig. EV1H). Similar results were obtained from paired sample expression profiles from TCGA (Fig. EV1I). Using cell line data from CCLE (Barretina et al, 2012), we found that MYB expression showed only a weak correlation with the copy number of its coding region (Fig. EV1J), indicating the copy number is not a major driver of MYB expression. Thus, we hypothesized that the overexpression of MYB in COREAD and STAD was driven by the presence of the candidate MYB-SE.

To determine the functional significance of MYB in gastrointestinal adenocarcinoma, we analyzed the genome-wide CRISPR screen data released by DepMap (Tsherniak et al, 2017). The CRISPR dependency score was normalized such that, for each cell line, the median effect of a set of known non-essential genes (referred to as “negative controls”) is 0, and the median effect of known essential genes (referred to as “positive controls”) is –1 (Liu et al, 2020). As a result, a lower score indicates a higher likelihood that a gene is dependent in a given cell line, so a cutoff of –0.5 corresponds to half the effect size expected from knocking out genes that are essential for each cell line (Liu et al, 2020). Hence, if a cell line exhibit CRISPR dependency score <−0.5 for a given gene, the cell line is dependent on that gene. The MYB dependency level identified from the data of DepMap was tightly correlated with the results from CRISPR-based genetic screen by the Sanger Institute (Behan et al, 2019) and shRNA-based screen project by the Broad Institute and Novartis (Root et al, 2006), indicating the robustness of the analysis (Fig. EV2A). Importantly, the cancer cell lines with higher expression levels of MYB exhibited an even greater dependence on the MYB gene for many types of cancer, including COREAD and STAD (Fig. EV2B). The expression and dependency of MYB in a subset of COREAD and STAD cancer cell lines are shown in Fig. EV2C. These results highlighted the significance of MYB overexpression in gastrointestinal adenocarcinoma.

Figure EV2. Functional significance of MYB in gastrointestinal adenocarcinoma.

(A) Correlation between MYB dependency scores from DepMap and results from the Sanger Institute CRISPR screen (left) or the Broad Institute shRNA screen (right), calculated using Spearman’s rank correlation coefficient based on the data from CCLE. (B) Correlation between MYB expression and dependency scores across all cancer cell lines in DepMap, calculated using Spearman’s rank correlation coefficient based on the data from CCLE. The results of COREAD and STAD cell lines were highlighted in red and blue, respectively. (C) The expression levels and dependency scores of MYB in the indicated cell lines based on DepMap. Source data are available online for this figure.

3D genomics analysis identified MYB candidate functional enhancers in gastrointestinal adenocarcinomas

We next sought to explore how the candidate MYB-SE regulates the expression of MYB in gastrointestinal cancer. It is widely accepted that the three-dimensional architecture of chromatin plays a crucial role in transcriptional regulation (Davidson and Peters, 2021). Therefore, we first investigated the interaction between MYB promoter and candidate MYB-SE using a published dataset of ChIP-seq based on the CCCTC-binding factor (CTCF) (Boukaba et al, 2022; Maestri et al, 2023; Sahu et al, 2022), which plays an indispensable role in mediating the formation of topologically associating domains (TADs) and chromatin loops (Davidson and Peters, 2021). Interestingly, the MYB promoter and the candidate MYB-SE resided within the same chromatin loop domain, defined by mutually convergent CTCFs binding sites (Fig. 2A), suggesting that the MYB promoter may interact with the candidate MYB-SE through chromatin looping.

Figure 2. Long-range interaction between the super-enhancer and promoter of MYB in gastrointestinal adenocarcinoma.

(A) The CTCF ChIP-seq tracks of multiple gastrointestinal cancer cell lines were presented. The red and green arrows represented the orientation of the CTCF motifs. A schematic of the chromosomal MYB loop domain is shown at the bottom. (B) The chromatin loops based on H3K27ac HiChIP data connected the MYB promoter to multiple distal elements with high-confidence PETs in various gastrointestinal cancer cell lines. (C) The demonstration of the epigenetic features for the MYB associated locus. Top: The chromatin interaction frequency heatmap in the gastrointestinal cancer cell line HT-55 based on H3K27ac HiChIP data. Turquoise box showed the interactions between the MYB promoter and the candidate MYB-SE. Middle: The ATAC-seq tracks of COREAD and STAD samples from TCGA. Bottom: The H3K27ac ChIP-seq tracks of 10 gastrointestinal cancer cell lines. Blue for COREAD tracks and red for STAD tracks.

By analyzing the published H3K27ac HiChIP data from representative cell lines (Table EV2) (Donohue et al, 2022; Johnstone et al, 2020; Karttunen et al, 2023; Liu et al, 2020; Maestri et al, 2023; Mortenson et al, 2024; Orouji et al, 2022), we investigated the physical interactions between the MYB promoter and the enhancers within the candidate MYB-SE by identifying promoter-enhancer loops. The results of the HiChIP analysis indicated that the constituent enhancers of the candidate MYB-SE had highly confident pair-end-tags (false discovery rate (FDR) < 0.05) linking them to the MYB promoter in MYB-high/-dependent cell lines (Fig. 2B) and a MYB-high/-independent HT-29 cell line (Appendix Fig. S2A). In contrast, these pair-end-tags were absent in MYB-low/-independent cell lines (including HCT 116, RKO, AGS cell lines) and normal human primary colonic epithelial cells (Appendix Fig. S2B). The independence of HT-29 on MYB can be attributed to either potential alternative mechanism of survival or the absence of SE (Appendix Fig. S2A).

We next analyzed the chromatin landscape using the Assay for Transposase-Accessible Chromatin with high throughput sequencing (ATAC-seq) of patient samples (Corces et al, 2018) and H3K27ac ChIP-seq of COREAD and STAD cancer cell lines with high or low MYB expression (Fig. 2C). We found that the candidate MYB-SE region exhibited a chromatin-accessible state (Fig. 2C), and MYB-high cancer cell lines displayed strong H3K27ac signals in this region (Fig. 2C). In contrast, MYB-low cancer cell lines showed minimal enrichment of H3K27ac signals in the candidate MYB-SE region (Fig. 2C). The situation became more complex when MYB dependency was considered. We identified the presence of the candidate MYB-SE in both MYB-high/-dependent and some MYB-high/-independent cell lines but not in MYB-low/-independent cell lines (Appendix Fig. S2C). Taken together, these results revealed a correlation between the presence of the candidate MYB-SE and the expression or dependency of MYB in gastrointestinal adenocarcinoma.

To identify the predominant enhancer(s) within the MYB-SE which may drive MYB overexpression, we divided the MYB-SE into seven individual enhancers, based on ATAC-seq and H3K27ac ChIP-seq signals in this region (Fig. 3A). The initial analysis was performed using self-transcribing active regulatory region sequencing (STARR-seq) (Sahu et al, 2022), a method to directly identify enhancers in a genome-wide manner. The results revealed that e4 exhibited stronger transcriptional activity compared to the other enhancers of the MYB-SE in the MYB-high and -dependent GP5d cells (Appendix Fig. S3A). We further validated the significance of e4 in MYB expression with the luciferase assay in HT-55 and SNU-719 cells (Fig. 3B). In contrast, the activity of e4 was not detected in HEK293-FT cells (Appendix Fig. S3B), consistent with the lineage specificity of MYB-SE in gastrointestinal adenocarcinomas. To mimic enhancer duplication in the gastrointestinal adenocarcinomas, we generated a plasmid containing an additional copy of e4 (Appendix Fig. S3C), which indeed exhibited stronger luciferase activity (Fig. 3C).

Figure 3. The identification of the predominant enhancer driving MYB expression.

(A) H3K27ac ChIP-seq signal from gastrointestinal adenocarcinoma cancer cell lines and ATAC-seq signal from tumor samples revealed the constituent enhancers e1–e7 for the MYB-SE, sgCtrl_1 and sgCtrl_2: sgRNA targeted negative control region. (B) Luciferase reporter assay measuring the activity of candidate enhancers in HT-55 and SNU-719 cells. The empty pGL3-promoter vector was used as a negative control. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (C) Luciferase reporter assay measuring the activity of single or duplicated enhancer e4 (2 × e4) in HT-55 and SNU-719 cells. The empty pGL3-promoter vector was used as a negative control. N = 3. All values are mean ± S.D. The P-value was determined by One-way ANOVA. (D) Schematic diagram of the strategy to validate the predominant enhancer by CRISPRi. (E) The expression of MYB upon CRISPRi-mediated repression of e4. Two separate sgRNAs, as sg-e4#3 and sg-e4#4, were used to target e4. NC, non-targeted control. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (F) Luciferase reporter assay measuring the activity of single or combinational enhancer in HT-55 and SNU-719 cells. The empty pGL3-promoter vector was used as a negative control. N = 3. All values are mean ± S.D. The P-value was determined by One-way ANOVA. Source data are available online for this figure.

In addition, we tested the function of e4 in regulating MYB expression by manipulating its transcriptional activity endogenously. We performed an improved CRISPR-based interference (CRISPRi) assay (details are provided in the Methods and Appendix Methods) (Liu et al, 2021; Yeo et al, 2018) to suppress the proposed constituent enhancers in four MYB-high gastrointestinal adenocarcinoma cell lines (Fig. 3D). The results revealed that repression of e4 led to a significant reduction in MYB expression across all four cell lines, whereas suppression of the other six enhancers had only minimal effects (Appendix Fig. S3D). Moreover, the effects of CRISPRi targeting e4 on MYB expression were attenuated in the MYB-high/-independent CCK-81 cells (Appendix Fig. S3D). The importance of e4 in MYB expression was further validated using two additional sgRNAs in HT-55 and SNU-719 cells (Fig. 3E). In contrast, activation of e4 by CRISPR-based activation assay (CRISPRa, details are provided in the Methods and Appendix Methods) (Sanson et al, 2018) in AGS and HGC-27 cell lines, two MYB-low/-independent cell lines, dramatically increased MYB expression (Appendix Fig. S3E). These results suggested that e4 within the MYB-SE is the predominant enhancer driving MYB expression.

It is widely accepted that a single enhancer may regulate the expression of multiple genes within the same genomic region (Fukaya et al, 2016). Based on the analysis of H3K27ac HiChIP data, we found that the HiChIP anchor of e4–7 exhibited the strongest physical interactions with the MYB promoter across all five MYB-high/-dependent cell lines (Fig. EV3A). However, we also found four additional genes, HBS1L, BCLAF1, MAP7 and IL20RA, whose promoters interacted with e4–7 in more than one cell line (Fig. EV3A). We therefore investigated whether e4 controls the transcription of genes other than MYB. CRISPRi-mediated suppression of e4 had minimal effects on the expression of HBS1L, BCLAF1, MAP7, and IL20RA as well as other genes surrounding MYB in two cell lines (Fig. EV3B,C). Although the expression of AHI1, a gene located near MYB, was modestly affected, no direct interaction with e4 was identified by the analysis of H3K27ac HiChIP data (Fig. EV3D,E). This suggested that e4 does not directly target AHI1 promoter via a chromatin loop extrusion mechanism. However, other enhancer regulation models, such as phase separation (Hnisz et al, 2017; Tang et al, 2022), could still contribute to the gene’s regulation.

Figure EV3. The enhancer e4 is the predominant element of the MYB-SE.

(A) The number of paired-end tags (PETs) that connecting e4–7 anchor to the promoters of candidate genes in five cell lines, based on H3K27ac HiChIP data. (B, C) The expression of candidate genes following CRISPRi targeting e4 in HT-55 (B) and SNU-719 (C) cells. The genes were annotated with the same color as (A). All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (D, E) The expression of genes adjacent to MYB following CRISPRi targeting e4 in HT-55 (D) and SNU-719 (E) cells. The genes were annotated with the same color as (A). All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (F) The contact heatmap based on H3K27ac HiChIP data revealed that e4 interacted with other candidate enhancers within the MYB-SE in HT-55, GP5d, and SNU-719 cells. (G) The number of PETs that connecting e4–7 anchor to e1–3 or e1 anchor based on H3K27ac HiChIP data in five cell lines. Source data are available online for this figure.

The enhancers within a SE often function cooperatively to regulate the expression of a single gene (Hnisz et al, 2017). Intestinally, we also observed interactions among multiple enhancers within the MYB-SE (Fig. EV3F,G). The luciferase assay demonstrated that the combination of e4 with other constituent enhancers within the MYB-SE exhibited higher activity than e4 alone (Fig. 3F), indicating a joint effect of these enhancers within the MYB-SE on transcriptional regulation. Collectively, our findings revealed the predominant role of e4 within MYB-SE in controlling MYB expression.

HNF4A controls the activity of e4 for the MYB-SE via direct interactions

After confirming the predominant role of e4 within the MYB-SE, we next investigated the potential transcription factors (TFs) that bind to e4 and regulate its activity. Based on the analysis of phastCons scores (Siepel et al, 2005) and DNase I hypersensitive sites (Consortium, 2012) on e4, we observed that this region exhibited high DNA sequence conservation and a chromatin-accessible state across numerous species (Fig. 4A), making it feasible for the downstream analysis. To identify TFs that may bind to e4, we performed motif analysis using publicly available ATAC-seq datasets from tumor samples, including 81 primary COREAD samples (from 38 donors) and 41 primary STAD samples (from 21 donors) from TCGA (Corces et al, 2018). By employing the FIMO (Grant et al, 2011) and TRAP (Thomas-Chollier et al, 2011) (see Methods and Appendix Methods for details), we identified that multiple potential transcription factor families may bind to e4 (Figs. 4A and EV4A).

Figure 4. Discovery of the transcription factors controlling MYB expression through direct interactions with the predominant enhancer.

(A) Motif analysis of e4 sequence. Top: phastCons scores (0:1 range) representing the conservation of e4 DNA sequence. Middle: Cluster of DNase I hypersensitive sites demonstrating the accessibility of e4. Bottom: Results of JASPAR-based prediction for transcription factors binding motifs in e4. (B) Luciferase reporter assay measuring the activity of e4 upon the deletion of the binding motifs. N = 2–3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (C) ChIP-qPCR analysis indicating the enrichment of transcription factors HNF4A and EBF1 at e4. NCR: Negative control region. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (D) The expression of HNF4A and MYB upon HNF4A knockdown by shRNAs. NC, non-targeted control. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (E) Luciferase reporter assay measuring the activity of e4 upon HNF4A knockdown in HT-55 cells. N = 3. All values are mean ± S.D. The P-value was determined by One-way ANOVA. (F) The expression of HNF4A and MYB upon the treatment of HNF4A antagonist BI-6015 in HT-55 cells. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (G) Luciferase reporter assay measuring the activity of theenhancer e4 following the treatment of HNF4A antagonist BI-6015 in HT-55 cells. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t test. (H) Cell viability of HT-55 cells after 48 h treatment of BI-6015, measured by Cell Counting Kit-8 (CCK-8). N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (I) ChIP-seq tracks for H3K27ac and HNF4A in the MYB-SE associated locus in GP5d cells. Source data are available online for this figure.

Figure EV4. Candidate transcription factors regulating MYB expression through interactions with e4.

(A) Transcription factors predicted to bind to e4 based on TRAP motif analysis; the P-value was determined by Benjamini-Hochberg multiple comparison correction method. (B) Demonstration of the motifs to be disrupted by CRISPR-Cas9. Red triangles indicated the cutting sites. (C) The expression of MYB in HT-55 cells following CRISPR-Cas9 based disruption of the identified motifs. N = 2. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (D) Demonstration of the motifs to be deletion by site-specific mutagenesis for the Luciferase assay. Red sequences indicated the targets of deletion. (E) Western blotting showing a decrease in HNF4A and EBF1 protein levels upon HNF4A or EBF1 knockdown, respectively. (F) The expression of HNF4G and MYB upon HNF4G knockdown. NC, non-targeted control. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (G) The expression of EBF1 and MYB upon EBF1 knockdown. NC, non-targeted control. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (H) Presence of the candidate functional motifs in the e1–e7 enhancers. Source data are available online for this figure.

To test whether the expression of MYB can be regulated by the binding of specific TFs, we performed genome editing using the CRISPR-Cas9 system to disrupt DNA motifs within e4. We found that disruption of the motifs associated with the HNF4, EBF, ASCL, SOX, or SNAI families led to a significant reduction in MYB expression (Fig. EV4B,C). Deletion of specific TF-binding motif within e4 showed that the ASCL, EHF, EBF, HNF4, SNAI, and SOX motifs are important for maintaining e4 activity, as demonstrated by the luciferase assay (Figs. 4B and EV4D). Especially, the depletion of EBF or HNF4 motifs had the most dramatic effects on e4 activity.

Based on the results of motif analysis and genome editing, we focused on two TFs, EBF1 (EBF family motif) and HNF4A (HNF4 family motif), for further investigation. First, we validated the binding of EBF1 and HNF4A to e4 by ChIP-qPCR assay in HT-55 and SNU-719 cells (Fig. 4C). Importantly, knockdown of HNF4A, but not HNF4G, reduced the expression of MYB and potently inhibited the activity of e4 in the luciferase assay (Figs. 4D,E and EV4E,F). In contrast, knockdown of EBF1 did not affect MYB expression (Fig. EV4E,G). Pharmacological inhibition of HNF4A using its antagonist, BI-6015, also suppressed MYB expression and e4 activity (Fig. 4F,G). Interestingly, treatment with BI-6015 also inhibited the growth of HT-55 cells (Fig. 4H). We further validated that HNF4A binds to e4, e5, e6 and e7 in GP5d cells using HNF4A ChIP-seq (Sahu et al, 2022) data. Among these regions, the binding peak at e4 exhibited the highest signal (Fig. 4I). Additionally, we found several candidate functional transcription factor motifs in other enhancers, suggesting that these transcription factors may bind to multiple enhancers to collaboratively regulate the activity of the MYB-SE (Fig. EV4H). In brief, these results suggested that HNF4A can bind to e4 and regulate MYB expression by modulating the activity of e4 within the MYB-SE.

MYB activates cancer-related genes in gastrointestinal adenocarcinomas

To systematically investigate the regulatory functions of MYB in gastrointestinal adenocarcinoma, we performed ChIP-seq analysis for MYB in HT-55 and SNU-719 cells. The canonical binding motif of MYB was identified by the software HOMER (Heinz et al, 2010) and presented at Fig. 5A. We also identified motifs for several transcription factors (TFs) in MYB-binding DNA sequences, including AP1, KLF5, ETS1, FOXA1 and other cancer-related TFs (Appendix Fig. S4A), suggesting that these TFs may work together with MYB in transcriptional regulation. We observed that MYB mainly bound to the promoter and distal intergenic regions (Fig. 5B,C), suggesting that MYB likely binds to regulatory elements, including enhancers and SEs.

Figure 5. MYB contributes to the high expression of cancer-related genes in gastrointestinal adenocarcinoma.

(A) Prediction of MYB’s DNA-binding motifs based on MYB ChIP-seq data. (B) Heatmaps showing the distribution of MYB binding sites cross the whole genome in HT-55 and SNU-719 cells. Genes in the rows were sorted in decreasing order of signal intensity. TSS, transcription start site. TES, transcription end site. (C) Bar charts illustrating the types of MYB binding sites in HT-55 and SNU-719 cells. (D) Functional enrichment analysis of the genes nearest to the MYB binding sites. The P-value was determined by Benjamini correction method. (E) ChIP-seq tracks of MYB in HT-55 and SNU-719 cells showing MYB bound to its super-enhancer. (F) MYB bound to the candidate super-enhancers of various cancer-related genes. (G) Validation of the functions of the candidate super-enhancers for GATA6 and ZFP36L2 by CRISPRi. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. Source data are available online for this figure.

Assigning the MYB binding sites to their nearest genes identified a total of 1240 genes in both HT-55 and SNU-719 cells (Appendix Fig. S4B and Table EV3). The analysis for GO and KEGG term enrichment revealed that these genes were involved in apoptosis, cancer development, and transcription regulation (Fig. 5D). Specifically, we found that MYB bound to e4, e6, and e7 within the MYB-SE based on ChIP-seq and ChIP-qPCR analysis (Fig. 5E; Appendix Fig. S4C). Knocking down MYB significantly inhibited e4 activity in the luciferase assay (Appendix Fig. S4D), indicating that MYB can regulate the transcriptional activity of e4. Thus, an auto-regulatory loop, a typical feature of lineage-specific oncogenes (Durbin et al, 2018; Jiang et al, 2020; Ott et al, 2018; Saint-Andre et al, 2016), may exist between MYB and MYB-SE. In addition, we found that MYB can bind to SEs of other cancer-related genes (Fig. 5F; Appendix Fig. S4E). Specifically, CRISPRi-based suppression of MYB binding sites in HT-55 cells downregulated the expression of GATA6 and ZFP36L2, two well-known genes with oncogenic roles in gastrointestinal adenocarcinoma (Chia et al, 2015; Xing et al, 2019) (Fig. 5G).

To investigate the genes whose expression is regulated by MYB, we performed RNA sequencing (RNA-Seq) analysis on MYB-depleted HT-55 cells. Differential expression analysis identified 532 upregulated genes and 496 downregulated genes in the MYB-depleted cells comparing to cells with a scrambled control (Appendix Fig. S5A). Integrative analysis of RNA-seq and ChIP-seq data revealed 407 genes as potential direct targets of MYB (Appendix Fig. S5B and Table EV4). Gene set enrichment analysis (GSEA) showed that MYB targeted genes, which were downregulated upon MYB depletion, were related to Notch signaling (Appendix Fig. S5C). These findings indicated that MYB directly controlled the expression of multiple cancer-related genes, including itself, which may contribute to the development of gastrointestinal adenocarcinoma.

Targeting e4 suppressed the growth of gastrointestinal cancer cells in vitro and in vivo

Considering the significance of MYB in regulating the expression of MYB and multiple oncogenes, we hypothesized that MYB is essential for the high proliferation rate of gastrointestinal cancer cells. To validate this, we first used an shRNA system to knock down MYB in HT-55 and SNU-719 cells. We observed that suppressing the expression of MYB significantly inhibited cell growth (Figs. 6A and EV5A,B). To characterize the effects of e4 on cellular function, we suppressed its activity by CRISPRi system and observed a significant decrease of cell growth as well (Figs. 6B and EV5C). Additionally, suppression of e4 prevented colony formation (Fig. 6C) and reduced the positive rate of EdU staining, a marker for cell proliferation, in HT-55 cells (Fig. 6D). Mechanistically, MYB suppression downregulated the expression of genes related to Notch signaling (Fig. EV5D). Notably, treatment with the Notch antagonist BMS-986115 also inhibited HT-55 cell growth (Fig. 6E).

Figure 6. The predominant enhancer of MYB as a master regulator in the development of gastrointestinal adenocarcinoma.

(A, B) Cell viability of HT-55 and SNU-719 cells upon MYB knockdown (A) or repression of e4 (B), measured by Cell Counting Kit-8 (CCK-8). N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (C) The effects of repression of e4 on colony formation of HT-55 cells. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (D) The effects of repression of e4 on cell proliferation of HT-55 cells as measured by EdU staining. N = 8. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (E) Cell viability of HT-55 cells upon the treatment of Notch antagonist BMS-986115, measured by Cell Counting Kit-8 (CCK-8). N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (F–H) The growth of xenografts derived from HT-55 cells was significantly inhibited after repression of e4. N = 6. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (I) Representative pictures of Ki-67 staining in xenografts. Scale bar, 250 µm. (J) The average number of Ki-67 positive cells per 10 mm² of xenograft tumor tissue. N = 6. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (K) Schematic diagram: HNF4A and MYB bind to e4 within MYB-SE to activate the expression of MYB. The overexpressed MYB protein directly upregulated its target genes and promoted the development of gastrointestinal adenocarcinoma. Source data are available online for this figure.

Figure EV5. Repression of MYB expression in vivo inhibited the development of gastrointestinal adenocarcinoma.

(A) The expression of MYB upon MYB knockdown in HT-55 and SNU-719 cells. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (B, C) Western blotting showing the decreased MYB protein upon MYB knockdown (B) or CRSPRi targeting e4 (C) in HT-55 and SNU-719 cells. The ACTB protein served as a loading control. (D) RT-qPCR analysis demonstrated the alterations in Notch signaling-related genes following MYB knockdown. N = 3. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (E–G) The growth of the xenografts derived from HT-55 cells was significantly inhibited following MYB knockdown. N = 3–5. All values are mean ± S.D. The P-value was determined by two-sided Student’s t-test. (H) Immunohistochemical staining for Ki-67 in xenograft samples following MYB knockdown in HT-55 cells. Scale bar = 250 μm. Source data are available online for this figure.

We then tested the significance of MYB for gastrointestinal adenocarcinoma in vivo using xenograft model with HT-55 cells in nude mice. We observed that both the volume and weight of the tumor xenografts were significantly reduced by suppressing e4 activity or knockdown of the endogenous MYB (Figs. 6F–H and EV5E–G). The expression of Ki-67, a marker of cell proliferation, was decreased following e4 inhibition or MYB knockdown in the xenografts (Figs. 6I–J and EV5H). Taking these results collectively, we concluded that MYB functions as a potential oncogene in promoting the development of gastrointestinal adenocarcinomas.

Discussion

Deciphering the biological and pathological basis of gastrointestinal adenocarcinoma could improve its diagnosis and therapy. By systematically investigating the active enhancer profiles of tumor samples as well as cell lines, we identified a lineage-specific super-enhancer for MYB, which contains a dominant enhancer e4 in gastrointestinal adenocarcinoma. Through an auto-regulatory loop, MYB regulates gastrointestinal adenocarcinoma development by directly controlling the transcription of multiple carcinogenic genes related to Notch signaling. These results suggested the specific role of MYB in the carcinogenesis of gastrointestinal adenocarcinoma (Fig. 6K).

Various types of genomic alterations at the MYB locus have been detected in human cancer, such as duplication (Lahortiga et al, 2007), translocation (Belloni et al, 2011; Clappier et al, 2007), gene-fusion (Belloni et al, 2011), and intragenic mutations (Hugo et al, 2006). However, these genomic alterations can only account for MYB expression in a subset of cancers. Increasing evidence suggests that the abnormal activation of enhancers represents a major epigenetic aberration in various types of cancer (Liu et al, 2021; Morton et al, 2019; Terekhanova et al, 2023; Zhang et al, 2018; Zhang et al, 2016). In the present study, we performed a comprehensive analysis of a super-enhancer (SE) that forms a loop with the MYB promoter and specifically regulates its expression in gastrointestinal adenocarcinoma. Especially, the predominant enhancer e4 interacts with other enhancers within MYB-SE to regulate its transcriptional activity. These results confirmed the existence of “facilitators”, as the elements contributing to the classical enhancer function without inherent activities (Blayney et al, 2023). It is possible that some of low-activity enhancers act as facilitators for e4 within the MYB-SE. We also found MYB-SE is frequently duplicated in gastrointestinal adenocarcinoma in comparison to other cancer types. Although the duplication for genomic region of MYB-SE has been reported in some of the T-ALL samples (Lahortiga et al, 2007; O’Neil et al, 2007), the MYB-SE region in ALL is not functional for MYB expression, as the enhancers in ALL for MYB are located 116 kb away from the MYB-SE (Li et al, 2021; Stadhouders et al, 2014).

The MYB protein is a critical therapeutic target for various cancer types, such as T cell acute lymphoblastic leukemia (T-ALL) (Pattabiraman and Gonda, 2013), breast cancer (Li et al, 2016), and prostate cancer (Acharya et al, 2023). In this study, we expanded the understanding of MYB by uncovering its essential role in gastrointestinal cancer. Our findings indicated that MYB binds to e4 within the MYB-SE to regulate its expression, thereby forming a self-regulatory circuit characteristic of lineage-specific master transcription factors (Durbin et al, 2018; Jiang et al, 2020; Ott et al, 2018; Saint-Andre et al, 2016). We also found that HNF4A regulated the transcription of MYB as well as the growth of gastrointestinal adenocarcinoma via binding to e4 within the MYB-SE. Given that an mRNA-based strategy has been developed to target HNF4A in preclinical models for liver fibrosis (Yang et al, 2021), similar strategy can be used to control the development of gastrointestinal adenocarcinoma with HNF4A-dependent overexpression of MYB.

In addition to its functions in cancer progression, MYB is also involved in the developmental process of neural tissues (Malaterre et al, 2008), colonic crypts (Malaterre et al, 2007), breast tissue (Li et al, 2016) as well as hematopoietic system (Jiang et al, 2006). We showed that the suppression of e4 within the MYB-SE, which downregulated the expression of MYB, was able to reduce the growth of gastrointestinal adenocarcinoma in vitro and in vivo. This finding highlighted the significance of e4 in MYB function and may inspire further studies about the MYB associated non-coding variants in the development of non-malignant tissues at the population level.

In summary, our study identified a MYB associated lineage-specific super-enhancer, containing a predominant enhancer e4 and a few potential facilitators, which regulated the development of gastrointestinal adenocarcinoma. These findings underscore the potential of non-coding elements, such as super-enhancers, as viable therapeutic targets for cancer treatment.

Methods

Reagents and tools table

Reagent/Resource	Reference or Source	Identifier or Catalog Number
Experimental models
HEK 293-FT cells (H. sapiens)	Cobioer	CBP60438
HT-55 cells (H. sapiens)	Cobioer	CBP60012
SNU-719 cells (H. sapiens)	Cobioer	CBP60511
GP5d cells (H. sapiens)	Sigma	95090715
CCK-81 cells (H. sapiens)	Cobioer	CBP60581
AGS cells (H. sapiens)	Cell Bank of Chinese Academy of Sciences	SCSP-5262
HGC-27 cells (H. sapiens)	Cell Bank of Chinese Academy of Sciences	TCHu 22
BALB/c-Nude (M. musculus)	GemPharmatech	D000521
Recombinant DNA
pLKO.1-puro	Biofeng	8453
Lenti-dCas9-KRAB-MeCP2-Blast	Liu et al (2021)	N/A
Lenti-dCas9-VP64-Blast	Addgene	61425
Lenti-Cas9-Blast	Addgene	52962
LentiGuide-puro	Addgene	52963
pXPR-502	Addgene	96923
pGL3-Promoter	Addgene	212939
pRL Renilla	Promega	E2271
pGL3-MYB-e1	This study	N/A
pGL3-MYB-e2	This study	N/A
pGL3-MYB-e3	This study	N/A
pGL3-MYB-e4	This study	N/A
pGL3-MYB-e5	This study	N/A
pGL3-MYB-e6	This study	N/A
pGL3-MYB-e7	This study	N/A
pGL3-MYB-e4×2	This study	N/A
pGL3-MYB-e3 + e4	This study	N/A
pGL3-MYB-e4 + e5 + e6 + e7	This study	N/A
Additional plasmids	This study	Appendix Table S1
Antibodies
Rabbit anti-c-MYB	Abcam	ab45150
Rabbit anti-c-MYB	Proteintech	17800-1-AP
Rabbit anti-HNF4α	Abcam	ab181604
Rabbit anti-EBF1	Millipore	AB10523
Mouse anti-ACTB	CWBio	CW0096M
Goat anti-rabbit HRP	ABclonal	AS014
Goat anti-mouse HRP	ABclonal	AS003
Mouse anti-Ki-67	Cell Signaling Technology	9449S
Oligonucleotides and sequence-based reagents
All primers and more information	This study	Table EV5
All sgRNAs, shRNA and more information	This study	Table EV5
Chemicals, enzymes, and other reagents
DMEM high glucose	Gibco	11995065
RPMI 1640	Gibco	11875119
Fetal bovine serum (FBS)	Gibco	10099141C
Penicillin-Streptomycin	Gibco	15140122
0.25% Trypsin-EDTA	Gibco	25200114
Trizol	Takara	9109
CalPhos Mammalian Transfection Kit	Clontech	631312
DynaGreen™ Protein A/G	Thermo Fisher Scientific	80105G
FastDigest Esp3 I	Thermo Fisher Scientific	FD0454
FastAP	Thermo Fisher Scientific	EF0654
Lipofectamine 3000 transfection reagent	Thermo Fisher Scientific	L3000015
Opti-MEM	Gibco	31985
PEI	Polysciences	24765
Kpn I-HF®	New England Biolabs	R3142S
Mlu I-HF®	New England Biolabs	R3198S
Xho I	New England Biolabs	R0146S
T4 Polynucleotide Kinase (T4 PNK)	New England Biolabs	M0201L
T4 DNA Ligase Reaction Buffer	New England Biolabs	B0202S
T7 DNA ligase	New England Biolabs	M0318L
Luna® Universal qPCR Master Mix	New England Biolabs	M3003E
LunaScript RT SuperMix	New England Biolabs	M3010L
Q5® Hot Start High-Fidelity DNA	New England Biolabs	M0493L
Q5 Site-Directed Mutagenesis kit	New England Biolabs	E0554S
10 mM dNTPs	New England Biolabs	N0446S
NEBuffer™ 2	New England Biolabs	B7002S
EcoRI-HF®	New England Biolabs	R3101L
AgeI-HF®	New England Biolabs	R3552L
0.5 M DTT	Beyotime	ST041-10ml
Dual Luciferase Reporter Gene Assay Kit	Beyotime	RG027
Cell Counting Kit-8	Beyotime	C0040
Ampicillin	Beyotime	ST008
Puromycin Dihydrochloride	Beyotime	ST551-250mg
Polybrene	Beyotime	C0351-1ml
Blasticidin S HCl	Beyotime	ST018-5ml
DEPC Water	Beyotime	R0022
DH5α Chemically Competent Cell	Tsingke	TSC-C14
DMOS	Sigma-Aldrich	D2650-100ML
Pierce™ DAB Substrate Kit	Thermo Fisher Scientific	34002
Pierce™ BCA Protein assay Kit	Thermo Fisher Scientific	23227
Cell-Light EdU Apollo488 In Vitro Kit	RIBOBIO	C10310-3
BI-6015	MedChemExpress	HY-108469
BMS-986115	MedChemExpress	HY-12860
Matrigel	Corning	354248
Software
GraphPad Prism 9	GraphPad	https://www.graphpad.com/
R	Bioconductor	https://www.r-project.org/
SAMtools	Li et al (2009)	http://samtools.sourceforge.net/
Bowtie2	Langmead and Salzberg (2012)	https://bowtie-bio.sourceforge.net/index.shtml
DEseq2	Love et al (2014)	https://github.com/thelovelab/DESeq2
HiC-Pro	Servant et al (2015)	https://github.com/nservant/HiC-Pro
HOMER	Heinz et al (2010)	http://homer.ucsd.edu/homer/
Liftover	Kent et al (2002)	https://genome.ucsc.edu/cgi-bin/hgLiftOver
MACS2	Zhang et al (2008)	https://github.com/macs3-project/MACS
STAR	Dobin et al (2013)	https://github.com/alexdobin/STAR
hichipper	Lareau and Aryee (2018)	https://hichipper.readthedocs.io/en/latest/index.html
DAVID	Sherman et al (2022)	https://davidbioinformatics.nih.gov/
BEDTools	Quinlan and Hall (2010)	https://bedtools.readthedocs.io/en/latest/index.html
ChIPseeker	Yu et al (2015)	https://github.com/YuLab-SMU/ChIPseeker
gTrack	Hadi et al (2020)	https://github.com/mskilab-org/gTrack
HTSeq	Anders et al (2015)	https://github.com/simon-anders/htseq

Cell lines and cell culture

AGS and HGC-27 cells were obtained from the Cell Bank of Chinese Academy of Sciences. HT-55, HEK293-FT, SNU-719, and CCK-81 cells were obtained from Nanjing Cobioer Biosciences company. AGS, HGC-27, and SNU-719 cells were cultured in RPMI 1640 medium (Gibco, #11875119) supplemented with 10% FBS (Gibco, #10099141C) and 1% penicillin-streptomycin (Gibco, #15140122). HEK293-FT, HT-55, and CCK-81 cells were cultured in DMEM high glucose medium (Gibco, #11995065) supplemented with 10% FBS and 1% penicillin-streptomycin. All cells were verified by short tandem repeat analysis and tested negative for mycoplasma using the Mycoplasma qPCR Detection Kit (Beyotime, #C0303S).

Luciferase reporter assay

Luciferase reporter assays were performed as previously described (Zhang et al, 2016). Firstly, e4 was amplified from the HT-55 cell’s genomic DNA, then e4 regions were cloned upstream of the pGL3-promoter vector (Addgene, #212939) using XhoI (NEB, #R0146S) and MluI (NEB, #R3198S) restriction enzyme sites. For cloning the 2 × e4, e4 region was cloned using XhoI and MluI restriction enzyme sites, followed the another e4 region was cloned using KpnI (NEB, #R3142S) and MluI restriction enzyme sites. For the combination of e4 and other constituent enhancers, other enhancers were cloned using XhoI and MluI restriction enzyme sites into 2×e4 vector. The enhancer luciferase reporter constructs were co-transfected with a control luciferase construct, named Renilla (Promega, #E2271), into cells using PEI (Polysciences, #24765) or lipofectamine 3000 (Invitrogen, #L3000015). The luciferase signal was first normalized to the Renilla luciferase signal and then normalized to the signal from the empty pGL3-promoter vector. Primers used for cloning are listed in Table EV5.

CRISPR-mediated DNA motif cutting, enhancer repression and activation

According to the ATAC-seq peaks within the candidate MYB-SE, the DNA sequence of the enhancers was obtained. Then, CRISPR/Cas9 sgRNAs were designed using the CRISPick tool from the Broad Institute (Doench et al, 2016; Sanson et al, 2018) and negative control, non-targeting sgRNAs were used as previously reported (Liu et al, 2021). For DNA motif cutting, cells were first infected with lenti-Cas9-blast vector (Addgene, #52962) and selected with 10 μg/ml of blasticidin (Beyotime, #ST018-5ml) for at least 7 days. Cells stably expressing Cas9 then subsequently infected with LentiGuide-Pour vector (Addgene, #52963) carrying either non-targeting sgRNAs or sgRNAs targeting the DNA motif and selected with 2 μg/ml puromycin (Beyotime, #ST551-250mg) for at least 5 days before any molecular or cellular assays. For enhancer repression, cells were first infected with lenti-dCas9-KRAB-MeCP2-blast vector (Liu et al, 2021) and selected with 10 μg/ml of blasticidin for at least 7 days. Cells stably expressing dCas9-KRAB-MeCP2 then subsequently infected with LentiGuide-Pour vector carrying either non-targeting sgRNAs or sgRNAs targeting the MYB enhancers and selected with 2 μg/ml puromycin for at least 5 days before any molecular or cellular assays. For enhancer activation, cells were first infected with lenti-dCas9-VP64-blast vector (Addgene, #61425) and selected with 10 μg/ml of blasticidin for at least 7 days. Cells stably expressing dCas9-VP64 then subsequently infected with pXPR-502 vector (Addgene, #96923) carrying either negative control sgRNAs or sgRNAs targeting e4 and selected with 2 μg/ml puromycin for at least 5 days before any molecular or cellular assays. All sgRNA sequences are listed in Table EV5.

Cell growth assays and colony formation assays

For cell growth assays, selected cells were collected. 5000 cells were seeded into 96-well plates (100 µL/well). After 5 days, 10 µL of Cell Counting Kit-8 (CCK-8) (Beyotime, #C0040) was added into each well according to the manufacturer’s instructions and incubated for 1 h. Optical density was measured at 450 nm using a microplate reader. For colony formation assays, 5000 cells were seeded into 6-well plates (2 mL/well). The cell culture medium was changed every three days. After culture for 18 days, cells were fixed with methanol (100%) and stained with crystal violet solution (25% methanol and 75% crystal violet) for 20 min at room temperature.

In vivo xenograft tumor assays

All animal experiments were performed in accordance with procedures approved by the Fudan University ethical committee (IDM2024011). All the mice were maintained under a 12 h light/12 h dark cycle at constant temperature (23 °C) and specific pathogen-free conditions with free access to food and water. A total 1 million HT-55 cells with or without CRISPRi system mediated e4 repression or shRNA system mediated MYB knock-down were resuspended in 100 μL serum-free DMEM and supplement with 100 μL Matrigel (Coring, #FAL-354248) (total volume 200 μL). Then the cells were subcutaneously injected into the flanks of female nude mice (BALB/c, Nu/Nu, 5–6 weeks old, female from GemPharmatech). Tumor growth was examined every 4–5 days, and tumor length and width were measured using calipers. Until tumor volume reached approximately 1000–1500 mm³ (the tumor size not exceeded 2 cm in any dimension), xenograft tumor-bearing mice were sacrificed, the tumor xenografts were isolated and weighted. Tumor volume was calculated suing the following formula: (length × width²) × 0.52. Animals that developed infections or severe side effects during the course of the experiment were intentionally excluded to maintain the integrity of the data and to avoid confounding factors that could impact the interpretation of the results.

Public data usage

This study used TCGA publicly available ATAC-seq data were downloaded from NCI Genomic Data Commons data portal (UTR: https://gdc.cancer.gov/about-data/publications/ATACseq-AWG), publicly PCAWG SV data were downloaded from PCAWG data portal (UTR: https://dcc.icgc.org/releases/PCAWG/consensus_sv), the remaining publicly available data were downloaded from GEO. Accession numbers of public dataset used in this study listed in Table EV2.

Quantification and statistical analysis

All statistical comparisons between two groups were performed by GraphPad Prism software 9.0 using a two-tailed unpaired t-test, unless specified. The variance between the statistically compared groups was similar.

Author contributions

Fuyuan Li: Conceptualization; Data curation; Formal analysis; Validation; Investigation; Visualization; Methodology; Writing—original draft; Writing—review and editing. Shangzi Wang: Conceptualization; Data curation; Formal analysis; Validation; Investigation; Visualization; Methodology; Writing—original draft; Writing—review and editing. Lian Chen: Conceptualization; Data curation; Formal analysis; Validation; Investigation; Visualization; Methodology; Writing—original draft; Writing—review and editing. Ning Jiang: Conceptualization; Data curation; Formal analysis; Validation; Investigation; Visualization; Methodology; Writing—original draft; Writing—review and editing. Xingdong Chen: Conceptualization; Data curation; Formal analysis; Supervision; Funding acquisition; Validation; Investigation; Visualization; Methodology; Writing—original draft; Writing—review and editing. Jin Li: Conceptualization; Data curation; Formal analysis; Supervision; Funding acquisition; Validation; Investigation; Visualization; Methodology; Writing—original draft; Writing—review and editing.

Source data underlying figure panels in this paper may have individual authorship assigned. Where available, figure panel/source data authorship is listed in the following database record: biostudies:S-SCDT-10_1038-S44320-025-00098-1.

Data availability

The datasets produced in this study are available in the following databases: MYB ChIP-seq data: Gene Expression Omnibus GSE234202. RNA-seq data: Gene Expression Omnibus (GEO) GSE234203.

The source data of this paper are collected in the following database record: biostudies:S-SCDT-10_1038-S44320-025-00098-1.

Disclosure and competing interests statement

The authors declare no competing interests.

Supplementary information

Expanded view data, supplementary information, appendices are available for this paper at 10.1038/s44320-025-00098-1.