Model-based analysis of regulation of gene expression (MARGE) is a framework for interpreting the relationship between the H3K27ac chromatin environment and differentially expressed gene sets. The framework has three main functions: MARGE-potential, MARGE-express, and MARGE-cistrome. MARGE-potential defines a regulatory potential (RP) for each gene as the sum of H3K27ac ChIP-seq signals weighted by a function of genomic distance from the transcription start site. The MARGE framework includes a compendium of RPs derived from 365 human and 267 mouse H3K27ac ChIP-seq data sets. Relative RPs, scaled using this compendium, are superior to superenhancers in predicting BET (bromodomain and extraterminal domain) -inhibitor repressed genes. MARGE-express, which uses logistic regression to retrieve relevant H3K27ac profiles from the compendium to accurately model a query set of differentially expressed genes, was tested on 671 diverse gene sets from MSigDB. MARGE-cistrome adopts a novel semisupervised learning approach to identify cis-regulatory elements regulating a gene set. MARGE-cistrome exploits information from H3K27ac signal at DNase I hypersensitive sites identified from published human and mouse DNase-seq data. We tested the framework on newly generated RNA-seq and H3K27ac ChIP-seq profiles upon siRNA silencing of multiple transcriptional and epigenetic regulators in a prostate cancer cell line, LNCaP-abl. MARGE-cistrome can predict the binding sites of silenced transcription factors without matched H3K27ac ChIP-seq data. Even when the matching H3K27ac ChIP-seq profiles are available, MARGE leverages public H3K27ac profiles to enhance these data. This study demonstrates the advantage of integrating a large compendium of historical epigenetic data for genomic studies of transcriptional regulation.

The cancer transcriptome is remarkably complex, including low-abundance transcripts, many not polyadenylated. To fully characterize the transcriptome of localized prostate cancer, we performed ultra-deep total RNA-seq on 144 tumors with rich clinical annotation. This revealed a linear transcriptomic subtype associated with the aggressive intraductal carcinoma sub-histology and a fusion profile that differentiates localized from metastatic disease. Analysis of back-splicing events showed widespread RNA circularization, with the average tumor expressing 7,232 circular RNAs (circRNAs). The degree of circRNA production was correlated to disease progression in multiple patient cohorts. Loss-of-function screening identified 11.3% of highly abundant circRNAs as essential for cell proliferation; for ∼90% of these, their parental linear transcripts were not essential. Individual circRNAs can have distinct functions, with circCSNK1G3 promoting cell growth by interacting with miR-181. These data advocate for adoption of ultra-deep RNA-seq without poly-A selection to interrogate both linear and circular transcriptomes.

The aberrant activation of the ERG oncogenic pathway due to the TMPRSS2-ERG gene fusion is the major event that contributes to prostate cancer (PCa) development. However, the critical downstream effectors that can be therapeutically targeted remain to be identified. In this study, we have found that the expression of the α1 and β1 subunits of soluble guanylyl cyclase (sGC) was directly and specifically regulated by ERG in vitro and in vivo and was significantly associated with TMPRSS2-ERG fusion in clinical PCa cohorts. sGC is the major mediator of nitric oxide (NO)-cGMP signaling in cells that, upon NO binding, catalyzes the synthesis of cGMP and subsequently activates protein kinase G (PKG). We showed that cGMP synthesis was significantly elevated by ERG in PCa cells, leading to increased PKG activity and cell proliferation. Importantly, we also demonstrated that sGC inhibitor treatment repressed tumor growth in TMPRSS2-ERG-positive PCa xenograft models and can act in synergy with a potent AR antagonist, enzalutamide. This study strongly suggests that targeting NO-cGMP signaling pathways may be a novel therapeutic strategy to treat PCa with TMPRSS2-ERG gene fusion.

We evaluated how variations in sequencing depth and other parameters influence interpretation of chromatin immunoprecipitation-sequencing (ChIP-seq) experiments. Using Drosophila melanogaster S2 cells, we generated ChIP-seq data sets for a site-specific transcription factor (Suppressor of Hairy-wing) and a histone modification (H3K36me3). We detected a chromatin-state bias: open chromatin regions yielded higher coverage, which led to false positives if not corrected. This bias had a greater effect on detection specificity than any base-composition bias. Paired-end sequencing revealed that single-end data underestimated ChIP-library complexity at high coverage. Removal of reads originating at the same base reduced false-positives but had little effect on detection sensitivity. Even at mappable-genome coverage depth of ∼1 read per base pair, ∼1% of the narrow peaks detected on a tiling array were missed by ChIP-seq. Evaluation of widely used ChIP-seq analysis tools suggests that adjustments or algorithm improvements are required to handle data sets with deep coverage.

Metastatic prostate cancer remains a major clinical challenge and metastatic lesions are highly heterogeneous and difficult to biopsy. Liquid biopsy provides opportunities to gain insights into the underlying biology. Here, using the highly sensitive enrichment-based sequencing technology, we provide analysis of 60 and 175 plasma DNA methylomes from patients with localized and metastatic prostate cancer, respectively. We show that the cell-free DNA methylome can capture variations beyond the tumor. A global hypermethylation in metastatic samples is observed, coupled with hypomethylation in the pericentromeric regions. Hypermethylation at the promoter of a glucocorticoid receptor gene NR3C1 is associated with a decreased immune signature. The cell-free DNA methylome is reflective of clinical outcomes and can distinguish different disease types with 0.989 prediction accuracy. Finally, we show the ability of predicting copy number alterations from the data, providing opportunities for joint genetic and epigenetic analysis on limited biological samples.

Ribosome biogenesis plays a pivotal role in tumorigenesis by supporting robust protein translation. We investigate the functional and molecular mechanism of Zinc finger protein 545 (ZNF545), a transcriptional repressor for ribosomal RNA (rRNA), in colorectal cancer (CRC). ZNF545 was silenced in CRC compared to adjacent normal tissues (P < 0.0001), implying a tumor-suppressive role. Colon-specific Znf545 knockout in mice accelerated CRC in ApcMin/+ and azoxymethane/dextran sulfate sodium-induced CRC. Mechanistically, we demonstrated that ZNF545 uses its two zinc finger clusters to bind to minimal rDNA promoter, where it assembled transcriptional repressor complex by interacting with KAP1. Znf545 deletion in mouse embryonic fibroblasts not only increased rRNA transcription rate and the nucleolar size and number but also altered the nucleolar composition and architecture with an increased number of fibrillar centers surrounded by net-like dense fibrillar components. Consequently, Znf545 deletion promoted the gene expression of translation machinery, protein translation, and cell growth. Consistent with its tumor-suppressive role, ZNF545 overexpression in CRC cells induced growth arrest and apoptosis. Finally, administration of rRNA synthesis inhibitor, CX-5461, inhibited CRC development in Znf545Δ/ΔApcMin/+ mice. In conclusion, ZNF545 suppresses CRC through repressing rRNA transcription and protein translation. Targeting rRNA biosynthesis in ZNF545-silenced tumors is a potential therapeutic strategy for CRC.

Driver gene mutations that are more prevalent in metastatic, castration-resistant prostate cancer (mCRPC) than localized disease represent candidate prognostic biomarkers. We analyze 1,844 localized (1,289) or mCRPC (555) tumors and quantify the prevalence of 113 somatic driver single nucleotide variants (SNVs), copy number aberrations (CNAs), and structural variants (SVs) in each state. One-third are significantly more prevalent in mCRPC than expected while a quarter are less prevalent. Mutations in AR and its enhancer are more prevalent in mCRPC, as are those in TP53, MYC, ZNRF3 and PRKDC. ZNRF3 loss is associated with decreased ZNRF3 mRNA abundance, WNT, cell cycle & PRC1/2 activity, and genomic instability. ZNRF3 loss, RNA downregulation and hypermethylation are prognostic of metastasis and overall survival, independent of clinical and pathologic indices. These data demonstrate a strategy for identifying biomarkers of localized cancer aggression, with ZNRF3 loss as a predictor of metastasis in prostate cancer.

Cancer stem cells (CSCs) are responsible for tumor progression, recurrence, and drug resistance. To identify genetic vulnerabilities of colon cancer, we performed targeted CRISPR dropout screens comprising 657 Drugbank targets and 317 epigenetic regulators on two patient-derived colon CSC-enriched spheroids. Next-generation sequencing of pooled genomic DNAs isolated from surviving cells yielded therapeutic candidates. We unraveled 44 essential genes for colon CSC-enriched spheroids propagation, including key cholesterol biosynthetic genes (HMGCR, FDPS, and GGPS1). Cholesterol biosynthesis was induced in colon cancer tissues, especially CSC-enriched spheroids. The genetic and pharmacological inhibition of HMGCR/FDPS impaired self-renewal capacity and tumorigenic potential of the spheroid models in vitro and in vivo. Mechanistically, HMGCR or FDPS depletion impaired cancer stemness characteristics by activating TGF-β signaling, which in turn downregulated expression of inhibitors of differentiation (ID) proteins, key regulators of cancer stemness. Cholesterol and geranylgeranyl diphosphate (GGPP) rescued the growth inhibitory and signaling effect of HMGCR/FDPS blockade, implying a direct role of these metabolites in modulating stemness. Finally, cholesterol biosynthesis inhibitors and 5-FU demonstrated antitumor synergy in colon CSC-enriched spheroids, tumor organoids, and xenografts. Taken together, our study unravels novel genetic vulnerabilities of colon CSC-enriched spheroids and suggests cholesterol biosynthesis as a potential target in conjunction with traditional chemotherapy for colon cancer treatment.

N6-methyladenosine (m6A) of mRNAs modulated by the METTL3-METTL14-WTAP-RBM15 methyltransferase complex and m6A demethylases such as FTO play important roles in regulating mRNA stability, splicing, and translation. Here, we demonstrate that FTO-IT1 long noncoding RNA (lncRNA) was upregulated and positively correlated with poor survival of patients with wild-type p53-expressing prostate cancer (PCa). m6A RIP-seq analysis revealed that FTO-IT1 knockout increased mRNA m6A methylation of a subset of p53 transcriptional target genes (e.g., FAS, TP53INP1, and SESN2) and induced PCa cell cycle arrest and apoptosis. We further showed that FTO-IT1 directly binds RBM15 and inhibits RBM15 binding, m6A methylation, and stability of p53 target mRNAs. Therapeutic depletion of FTO-IT1 restored mRNA m6A level and expression of p53 target genes and inhibited PCa growth in mice. Our study identifies FTO-IT1 lncRNA as a bona fide suppressor of the m6A methyltransferase complex and p53 tumor suppression signaling and nominates FTO-IT1 as a potential therapeutic target of cancer.

Expression quantitative trait locus (eQTL) analysis has emerged as an important tool in elucidating the link between genetic variants and gene expression, thereby bridging the gap between risk SNPs and associated diseases. We recently identified and validated a specific case where the methylation of a CpG site influences the relationship between the genetic variant and gene expression.

High-dimensional genomic data analysis is challenging due to noises and biases in high-throughput experiments. We present a computational method matrix analysis and normalization by concordant information enhancement (MANCIE) for bias correction and data integration of distinct genomic profiles on the same samples. MANCIE uses a Bayesian-supported principal component analysis-based approach to adjust the data so as to achieve better consistency between sample-wise distances in the different profiles. MANCIE can improve tissue-specific clustering in ENCODE data, prognostic prediction in Molecular Taxonomy of Breast Cancer International Consortium and The Cancer Genome Atlas data, copy number and expression agreement in Cancer Cell Line Encyclopedia data, and has broad applications in cross-platform, high-dimensional data integration.

Transcription factor cistromes are highly cell-type specific. Chromatin accessibility, histone modifications, and nucleosome occupancy have all been found to play a role in defining these binding locations. Here, we show that hormone-induced DNase I hypersensitivity changes (ΔDHS) are highly predictive of androgen receptor (AR) and estrogen receptor 1 (ESR1) binding in prostate cancer and breast cancer cells, respectively. While chromatin structure prior to receptor binding and nucleosome occupancy after binding are strikingly different for ESR1 and AR, ΔDHS is highly predictive for both. AR binding is associated with changes in both local nucleosome occupancy and DNase I hypersensitivity. In contrast, while global ESR1 binding is unrelated to changes in nucleosome occupancy, DNase I hypersensitivity dynamics are also predictive of the ESR1 cistrome. These findings suggest that AR and ESR1 have distinct modes of interaction with chromatin and that DNase I hypersensitivity dynamics provides a general approach for predicting cell-type specific cistromes.

Chromatin plays a central role in eukaryotic gene regulation. We performed genome-wide mapping of epigenetically marked nucleosomes to determine their position both near transcription start sites and at distal regulatory elements, including enhancers. In prostate cancer cells, where androgen receptor binds primarily to enhancers, we found that androgen treatment dismisses a central nucleosome present at androgen receptor binding sites that is flanked by a pair of marked nucleosomes. A new quantitative model built on the behavior of such nucleosome pairs correctly identified regions bound by the regulators of the immediate androgen response, including androgen receptor and FOXA1. More importantly, this model also correctly predicted previously unidentified binding sites for other transcription factors present after prolonged androgen stimulation, including OCT1 and NKX3-1. Therefore, quantitative modeling of enhancer structure provides a powerful predictive method to infer the identity of transcription factors involved in cellular responses to specific stimuli.

Transcription factors (TFs) are spatially and temporally regulated during gut organ specification. Although accumulating evidence shows aberrant reactivation of developmental programs in cancer, little is known about how TFs drive lineage specification in development and cancer. We first defined gastrointestinal tissue-specific chromatin accessibility and gene expression during development, identifying the dynamic epigenetic regulation of SOX family of TFs. We revealed that Sox2 is not only essential for gastric specification, by maintaining chromatin accessibility at forestomach lineage loci, but also sufficient to promote forestomach/esophageal transformation upon Cdx2 deletion. By comparing our gastrointestinal lineage-specific transcriptome to human gastrointestinal cancer data, we found that stomach and intestinal lineage-specific programs are reactivated in Sox2high /Sox9high and Cdx2high cancers, respectively. By analyzing mice deleted for both Sox2 and Sox9, we revealed their potentially redundant roles in both gastric development and cancer, highlighting the importance of developmental lineage programs reactivated by gastrointestinal TFs in cancer.

Androgen receptor (AR) in prostate cancer (PCa) can drive transcriptional repression of multiple genes including MYC, and supraphysiological androgen is effective in some patients. Here, we show that this repression is independent of AR chromatin binding and driven by coactivator redistribution, and through chromatin conformation capture methods show disruption of the interaction between the MYC super-enhancer within the PCAT1 gene and the MYC promoter. Conversely, androgen deprivation in vitro and in vivo increases MYC expression. In parallel, global AR activity is suppressed by MYC overexpression, consistent with coactivator redistribution. These suppressive effects of AR and MYC are mitigated at shared AR/MYC binding sites, which also have markedly higher levels of H3K27 acetylation, indicating enrichment for functional enhancers. These findings demonstrate an intricate balance between AR and MYC, and indicate that increased MYC in response to androgen deprivation contributes to castration-resistant PCa, while decreased MYC may contribute to responses to supraphysiological androgen therapy.

Androgen receptor (AR) messenger RNA (mRNA) alternative splicing variants (AR-Vs) are implicated in castration-resistant progression of prostate cancer (PCa), although the molecular mechanism underlying the genesis of AR-Vs remains poorly understood. The CDK12 gene is often deleted or mutated in PCa and CDK12 deficiency is known to cause homologous recombination repair gene alteration or BRCAness via alternative polyadenylation (APA). Here, we demonstrate that pharmacological inhibition or genetic inactivation of CDK12 induces AR gene intronic (intron 3) polyadenylation (IPA) usage, AR-V expression, and PCa cell resistance to the antiandrogen enzalutamide (ENZ). We further show that AR binds to the CCNK gene promoter and up-regulates CYCLIN K expression. In contrast, ENZ decreases AR occupancy at the CCNK gene promoter and suppresses CYCLIN K expression. Similar to the effect of the CDK12 inhibitor, CYCLIN K degrader or ENZ treatment promotes AR gene IPA usage, AR-V expression, and ENZ-resistant growth of PCa cells. Importantly, we show that targeting BRCAness induced by CYCLIN K down-regulation with the PARP inhibitor overcomes ENZ resistance. Our findings identify CYCLIN K down-regulation as a key driver of IPA usage, hormonal therapy-induced AR-V expression, and castration resistance in PCa. These results suggest that hormonal therapy-induced AR-V expression and therapy resistance are vulnerable to PARP inhibitor treatment.

Gastric cancer (GC) is one of the most common deadly cancers in the world. Although patient genomic data have identified AT-rich interaction domain 1A (ARID1A), a key chromatin remodeling complex subunit, as the second most frequently mutated gene after TP53, its in vivo role and relationship to TP53 in gastric tumorigenesis remains unclear. Establishing a novel mouse model that reflects the ARID1A heterozygous mutations found in the majority of human GC cases, we demonstrated that Arid1a heterozygosity facilitates tumor progression through a global loss of enhancers and subsequent suppression of the p53 and apoptosis pathways. Moreover, mouse genetic and single-cell analyses demonstrated that the homozygous deletion of Arid1a confers a competitive disadvantage through the activation of the p53 pathway, highlighting its distinct dosage-dependent roles. Using this unique vulnerability of Arid1a mutated GC cells, our combined treatment with the epigenetic inhibitor, TP064, and the p53 agonist, Nutlin-3, inhibited growth of Arid1a heterozygous tumor organoids, providing a novel therapeutic option for GC.

BRCA1 mutations are associated with increased breast and ovarian cancer risk. BRCA1-mutant tumors are high-grade, recurrent, and often become resistant to standard therapies. Herein, we performed a targeted CRISPR-Cas9 screen and identified MEPCE, a methylphosphate capping enzyme, as a synthetic lethal interactor of BRCA1. Mechanistically, we demonstrate that depletion of MEPCE in a BRCA1-deficient setting led to dysregulated RNA polymerase II (RNAPII) promoter-proximal pausing, R-loop accumulation, and replication stress, contributing to transcription-replication collisions. These collisions compromise genomic integrity resulting in loss of viability of BRCA1-deficient cells. We also extend these findings to another RNAPII-regulating factor, PAF1. This study identifies a new class of synthetic lethal partners of BRCA1 that exploit the RNAPII pausing regulation and highlight the untapped potential of transcription-replication collision-inducing factors as unique potential therapeutic targets for treating cancers associated with BRCA1 mutations.

Dysregulation of histone lysine methyltransferases and demethylases is one of the major mechanisms driving the epigenetic reprogramming of transcriptional networks in castration-resistant prostate cancer (CRPC). In addition to their canonical histone targets, some of these factors can modify critical transcription factors, further impacting oncogenic transcription programs. Our recent report demonstrated that LSD1 can demethylate the lysine 270 of FOXA1 in prostate cancer (PCa) cells, leading to the stabilization of FOXA1 chromatin binding. This process enhances the activities of the androgen receptor and other transcription factors that rely on FOXA1 as a pioneer factor. However, the identity of the methyltransferase responsible for FOXA1 methylation and negative regulation of the FOXA1-LSD1 oncogenic axis remains unknown. SETD7 was initially identified as a transcriptional activator through its methylation of histone 3 lysine 4, but its function as a methyltransferase on nonhistone substrates remains poorly understood, particularly in the context of PCa progression. In this study, we reveal that SETD7 primarily acts as a transcriptional repressor in CRPC cells by functioning as the major methyltransferase targeting FOXA1-K270. This methylation disrupts FOXA1-mediated transcription. Consistent with its molecular function, we found that SETD7 confers tumor suppressor activity in PCa cells. Moreover, loss of SETD7 expression is significantly associated with PCa progression and tumor aggressiveness. Overall, our study provides mechanistic insights into the tumor-suppressive and transcriptional repression activities of SETD7 in mediating PCa progression and therapy resistance.

Cancer cells exhibit phenotypical plasticity and epigenetic reprogramming, which allows them to evade lineage-dependent targeted treatments by adopting lineage plasticity. The underlying mechanisms by which cancer cells exploit the epigenetic regulatory machinery to acquire lineage plasticity and therapy resistance remain poorly understood. We identified Zinc Finger Protein 397 (ZNF397) as a bona fide co-activator of the androgen receptor (AR), essential for the transcriptional program governing AR-driven luminal lineage. ZNF397 deficiency facilitates the transition of cancer cell from an AR-driven luminal lineage to a Ten-Eleven Translocation 2 (TET2)-driven lineage plastic state, ultimately promoting resistance to therapies inhibiting AR signaling. Intriguingly, our findings indicate that TET2 inhibitor can eliminate the AR targeted therapies resistance in ZNF397-deficient tumors. These insights uncover a novel mechanism through which prostate and breast cancers acquire lineage plasticity via epigenetic rewiring and offer promising implications for clinical interventions designed to overcome therapy resistance dictated by lineage plasticity.