New Results

3D genomic analysis reveals novel enhancer-hijacking caused by complex structural alterations that drive oncogene overexpression

Katelyn L. Mortenson, Courtney Dawes, Emily R. Wilson, Nathan E. Patchen, Hailey E. Johnson, Jason Gertz, Swneke D. Bailey, Yang Liu, Katherine E. Varley, Xiaoyang Zhang
doi: https://doi.org/10.1101/2024.01.23.576965

ABSTRACT

Cancer genomes are composed of many complex structural alterations on chromosomes and extrachromosomal DNA (ecDNA), making it difficult to identify non-coding enhancer regions that are hijacked to activate oncogene expression. Here, we describe a 3D genomics-based analysis called HAPI (Highly Active Promoter Interactions) to characterize enhancer hijacking. HAPI analysis of HiChIP data from 34 cancer cell lines identified enhancer hijacking events that activate both known and potentially novel oncogenes such as MYC, CCND1, ETV1, CRKL, and ID4. Furthermore, we found enhancer hijacking among multiple oncogenes from different chromosomes, often including MYC, on the same complex amplicons such as ecDNA. We characterized a MYC-ERBB2 chimeric ecDNA, in which ERBB2 heavily hijacks MYC’s enhancers. Notably, CRISPRi of the MYC promoter led to increased interaction of ERBB2 with MYC enhancers and elevated ERBB2 expression. Our HAPI analysis tool provides a robust strategy to detect enhancer hijacking and reveals novel insights into oncogene activation.

INTRODUCTION

Somatic structural variants, including deletions, duplications, inversions, and translocations, dramatically alter genomic structures in cancer cells. These events can be readily detected by DNA imaging, whole-genome/long-read sequencing, or Hi-C-based 3D genomic assays1–5. Understanding the functional consequences of structural variants is key to uncovering the mechanisms of tumorigenesis and developing therapeutic strategies. For instance, structural variants can create a fusion gene with oncogenic properties such as BCR-ABL1 and EML4-ALK6,7. However, most of the breakpoints of cancer-associated structural variants fall in noncoding regions8. These variants can place enhancer elements adjacent to oncogenes such as MYC, MYCN, MYB, CCND1, EVI1, and GFI1, and activate their expression, events known as enhancer hijacking9–17.

Several analysis approaches have been developed to identify enhancer hijacking. PANGEA and CESAM identify candidate genes based on the correlation of their expression with adjacent structural alterations across tumor samples, which require large cohorts of data and do not utilize chromatin interaction evidence of enhancer hijacking18,19. Neoloop reconstructs Hi-C maps in structurally altered cancer genomes to identify aberrant chromatin loops and their associated genes20, however, it does not prioritize the gene targets and the underlying functional enhancers. As structural variants often affect many enhancer-promoter interactions, most of which are likely passenger events2,3, it is essential to prioritize the functional events. Furthermore, recent work has revealed that complex structural variants such as chromothripsis lead to the formation of circular extrachromosomal DNA (ecDNA)21. These events often cause extensive alterations of enhancer-promoter interactions, posing another challenge to characterize the underlying mechanisms of oncogene regulation22–24.

Here, we propose a 3D genomics-based strategy to identify and characterize enhancer hijacking events based on two assumptions: 1) oncogenes subject to enhancer hijacking should be highly regulated by enhancers; 2) the hijacked enhancers should contribute a substantial proportion of an oncogene’s enhancer usage. To test this, we utilized H3K27ac HiChIP data to map enhancer-promoter interactions and measured their interaction intensities. We applied our strategy to HiChIP data from 34 cancer and two normal cell lines. This approach identified known and novel enhancer hijacking events, prioritized their target genes, and defined the underlying enhancers. We also revealed novel enhancer hijacking events associated with ecDNAs or complex chromosomal amplicons. In these events, multiple oncogenes from different chromosomal regions are translocated near each other, amplified together, and hijack each other’s enhancers to boost their overexpression.

RESULTS

HAPI analysis identifies genes that are highly interactive with enhancers

We first developed an analysis method, Highly Active Promoter Interactions (HAPI), that utilizes H3K27ac HiChIP data to quantify all enhancer interactions of each gene’s promoter in the genome, including intra- and inter-chromosomal interactions. This method ranks genes by 1) their enhancer contact value defined as the number of interacting enhancers and 2) their interaction intensity as defined by the number of HiChIP paired-end tags (PETs). We define highly interactive (HAPI) genes as those that exceed the inflection points of both the ranked enhancer contact value and ranked interaction intensity in a manner similar to previously published methods that define super-enhancers25 (Figure 1A), as exemplified in the prostate cancer cell line LNCaP (Figure 1B and S1A-B).

Figure 1:
Figure 1: HAPI analysis identifies cancer-related genes that are highly interactive with enhancers.

A. Schematic models illustrating the identification of HAPI genes based on enhancer interactions.

B. Plotted are the number of enhancers (x-axis) and interaction intensity (y-axis) for each gene in LNCaP cells. HAPI cutoff is defined by the inflection point of each scoring metric.

C. Spearman correlation clustering of cancer cell lines based on normalized interaction scores of their associated HAPI genes. BRCA: breast cancer; COAD: colorectal adenocarcinoma; LUSC: lung squamous cell carcinoma; PRAD: prostate adenocarcinoma; SCLC: small cell lung cancer; STAD: stomach adenocarcinoma.

D. The percentage of HAPI genes (shared by at least half of the cell lines in each cancer type) that are previously annotated as oncogenes. The numbers outside and within each bar indicate the total number of shared HAPI (red) or other (grey) genes and the ones annotated as oncogenes, respectively. P-values are derived from two-sided Fisher’s exact tests.

We applied HAPI analysis to H3K27ac HiChIP data that we and others generated in 34 cancer cell lines representing multiple cancer types (Supplementary Table 1), which identified on average 538 HAPI genes in each cell line (Supplementary Data 1). In LNCaP cells, HAPI genes are more likely to interact with super-enhancers compared to non-HAPI enhancer-connected genes as expected26,27 (Figure S1C), but there are additional highly interactive HAPI genes that do not utilize super-enhancers (Figure S1C). Conversely, super-enhancers are not necessarily looped to HAPI genes in LNCaP cells, as most of the super enhancers interact with both HAPI and non-HAPI genes, or only non-HAPI genes (Figure S1D). Overexpressed oncogenes such as MYC, FOXA1, TBX3, and PCAT1 were identified as HAPI genes in LNCaP cells. Notably, HAPI genes are not necessarily genes with the highest expression, as highly expressed housekeeping genes ACTB, HPRT1, and GAPDH fell below the HAPI cutoff (Figure 1B and S1E). Furthermore, when we investigated whether copy number amplification had an impact HAPI gene metrics, we found that only a small proportion of the amplified genes are identified as HAPI genes in LNCaP cells (Figure S1F). We did not normalize copy number variants in this analysis, as we and others have shown that focal copy number amplifications contribute to enhancer-promoter interactions for oncogenes – a signal that we want to preserve in our analysis28–33.

We performed unsupervised clustering of the cell lines based on a combination of the enhancer contact and interaction intensity values for each identified HAPI gene (Figure 1C, see Methods for more details). The results showed that cell lines from the same lineages tend to cluster together, highlighting the cancer type-specificity of HAPI genes. We also found significant enrichments of previously annotated oncogenes34–36, and genes involved in cancer-related pathways in our identified HAPI gene lists (Figure 1D and Supplementary Table 2). Overall, HAPI analysis identified known cancer-related genes that are highly interactive with enhancers in a cancer-type specific manner.

HAPI analysis identifies known and novel enhancer-hijacking events

HAPI analysis not only defines genes that are highly interactive with enhancers but also determines the location of their enhancers. This enables us to determine whether the enhancers are in the same genomic region as the HAPI gene or are hijacked from other genomic regions likely through structural alterations (Figure 2A). These include both ‘trans’ hijacking events that relocate enhancers from other chromosomes to a HAPI gene and ‘cis’ hijacking events that relocate distant enhancers (>2Mb, the upper limit of typical TAD domains37) on the same chromosome.

Figure 2:
Figure 2: HAPI analysis identifies known and novel enhancer-hijacking events.

A. Schematic models illustrating the identification of enhancer hijacking HAPI genes based on the enhancer origin.

B. The number of cancer cell lines that are identified to contain trans- and cis-enhancer hijacking events in our cohort.

C. The number of identified trans-enhancer hijacking genes found in each cancer cell line in our cohort. Two immortalized epithelial cell lines are used as negative controls.

D. Plotted are the calculated trans-enhancer contribution (x-axis) and the copy number estimated by SNP-array data (y-axis) for each HAPI gene in selected cell lines. Besides are the expression levels for the highlighted gene in all CCLE cell lines (violin plot), the trans-enhancer hijacking cell line (red), and other cell lines of the same lineage (blue).

E. Same as D but with regards to two cis-enhancer hijacking genes in VCaP and CORL88 cells.

We analyzed H3K27ac HiChIP data of the 34 cancer cell lines to determine the origin of enhancers for HAPI genes. For HAPI genes with over 25% of enhancer activity coming from either trans- or abnormal cis-interactions were treated as candidate genes driven by enhancer hijacking. In total, we identified enhancer hijacking genes in 29 of the 34 cell lines (Figure 2B). In this study, we focus on trans-enhancer hijacking events because they are more likely caused by structural alterations than cis-enhancer hijacking events that may be caused by other types alterations such as epigenetic loss of TAD boundaries38,39. We found trans-enhancer hijacking events in 16 of the cancer cell lines, but in neither of the two normal cell lines we examined (Figure 2C). The calculated enhancer origins, both trans and cis, for all HAPI genes are listed in Supplementary Data 1.

In addition to enhancer origins, we also examined copy numbers of the HAPI genes using published SNP-array data40. We first focused on HAPI genes that exhibit neutral copies or modest amplifications (log2(CN)<2) (Figure 2D-E). This analysis identified previously reported trans-enhancer hijacking events activating oncogenes CCND1 in the B-cell lymphoma cell line REC110,12 and ETV1 in the prostate cancer cell lines LNCaP and MDAPCA2B20,41,42. We also found novel trans-enhancer hijacking events activating known oncogenes including MYC in the small cell lung cancer cell line NCIH146, NOTCH1 in the T-cell lymphoblastic leukemia cell line CUTLL1, and CCND1 in the breast cancer cell line ZR751. Furthermore, HAPI analysis identified potentially novel oncogenes that hijack enhancers in trans, such as CRKL in the colorectal cancer cell line HT29 and FOXJ2 in the small cell lung cancer cell line NCIH2171 (Figure 2D and S2A). Finally, we found cis-enhancer hijacking events activating potentially novel oncogenes such as ID4 in the prostate cancer cell line VCaP and TXLNA in the small cell lung cancer cell line CORL88 (Figure 2E), which are likely caused by large deletions based on the known patterns of structural variants in the HiChIP signal (Figure S2B-D). Overall, cell lines with these enhancer-hijacking events are associated with higher expression of the involved genes as compared to other cell lines of the same lineage (Figure 2D-E and S2A).

Functional validation of identified enhancer hijacking events

We then chose two representative enhancer hijacking genes identified from HAPI analysis for functional validation. We identified ETV1 (chromosome 7) as a trans-enhancer hijacking gene in both LNCaP and MDAPCA2B cells, with most of its trans enhancer activity coming from chromosome 14 (Figure 3A). This is caused by a cryptic insertion in LNCaP and a balanced translocation in MDAPCA2B43 that connect the two chromosomal regions. A translocation linking these two regions is also identified in a primary prostate cancer patient tumor sample from Pan-Cancer Analysis of Whole Genomes (PCAWG) (Figure S3A). Previous studies have identified various potential enhancers within regions of this translocation, but a consensus regarding which one regulates ETV1 has not yet been reached20,42. Based on H3K27ac HiChIP and ChIP-seq signal, we observed four enhancers e1-e4, upstream of FOXA1 on chromosome 14, hijacked to the ETV1 promoter in LNCaP and MDAPCA2B cells (Figure 3A). CRISPRi assays in LNCaP showed that repression of e3 and e4 caused the most significant decreases in expression of ETV1 (Figure 3B). H3K27ac ChIP-seq performed after CRISPRi confirmed the on-target effects (Figure S3B). As a control, repression of the four enhancers did not affect ETV1 expression in 22RV1 cells that lack the translocation (Figure S3C). Interestingly, these enhancers have little or modest regulatory effects on the nearby gene FOXA1 or TTC6 on their original chromosome in LNCaP cells (Figure 3B and S3D), indicating that these two genes may be dependent on additional enhancers.

Figure 3:
Figure 3: Functional validation of identified trans-enhancer hijacking events.

A. HiChIP and ChIP-seq signal at chromosomal regions containing ETV1 and the hijacked enhancers in MDAPCA2B and LNCaP cells. The H3K27ac ChIP-seq signal in MDAPCA2B cells is derived from HiChIP reads. In LNCaP cells, WGS-identified breakpoints support translocations underlying the enhancer-hijacking events. The DNA position of the left breakpoint on chromosome 7 in LNCaP cells is noted.

B. RT-qPCR measuring expression changes of ETV1 and FOXA1 after CRISPRi of each enhancer e1-e4 or the promoter of ETV1 in LNCaP cells. Expression levels were normalized to cells treated with a non-targeting sgRNA (NT1). N = 3 biological replicates. Data are presented as mean values +/-SEM. P-values were derived from two-sided t-tests. Source data are provided as a Source Data file.

C. HiChIP and ChIP-seq signal at chromosomal regions containing CCND1 and the hijacked enhancers in REC1 and ZR751 cells. In ZR751 cells, WGS-identified breakpoints support translocations underlying the enhancer-hijacking events.

D. RT-qPCR measuring expression changes of CCND1 and DUSP4 after CRISPRi of each enhancer e1-e5, the promoter of CCND1, or a combination of the five enhancers in ZR751 cells. Expression levels were normalized to cells treated with a non-targeting sgRNA (NT1). N = 3 biological replicates. Data are presented as mean values +/-SEM. P-values were derived from two-sided t-tests. Source data are provided as a Source Data file.

ETV1 was reported as an androgen-responsive gene in LNCaP cells promoting cell invasion44. We observed binding of androgen receptor (AR) to both e3 and e4 but not the promoter of ETV1 (Figure 3A and S3E). Additionally, we observed a 40% upregulation of ETV1 expression in response to Dihydrotestosterone (DHT) treatment in LNCaP cells. In contrast, ETV1 expression is minimally affected by DHT in the untranslocated 22RV1 cells, where AR binds the e3 and e4 enhancers in their original locus (Figure S3F). These data suggest that the androgen-induced ETV1 upregulation is caused by these two hijacked AR-bound enhancers.

The second representative enhancer hijacking gene we selected from HAPI analysis for functional validation was CCND1. CCND1 (chromosome 11) is known to hijack enhancers from chromosome 14 in the B-cell lymphoma cell line REC110,12. We found that the same oncogene hijacks different enhancers from chromosome 8 in the breast cancer cell line ZR751 (Figure 3C). WGS analysis revealed a translocation that places these enhancers upstream of CCND1 in ZR751 (Figure 3C). In primary breast cancer patient tumor samples from PCAWG, although we did not observe the same translocation event, we found that the CCND1 locus is highly subject to translocations with diverse regions of chromosome 8 (Figure S3G). We applied CRISPRi to repress five of the strongest enhancers (e1-e5 in Figure 3C) individually or simultaneously in ZR751. While targeting individual enhancers yields little (e1 and e2) or modest effect (e3, e4, and e5) on CCND1 expression, targeting the combination of all enhancers with CRISPRi caused a 45% reduction in CCND1 expression (Figure 3D). Repressing most of these enhancers caused significant decreases in expression of their endogenous target gene DUSP4 on chromosome 8 (Figure 3D). These results together demonstrate that these enhancers activate CCND1, in a combinatorial manner, due to enhancer hijacking while also acting in their canonical role of activating DUSP4. CCND1 is overexpressed in 50% of breast cancers, which has been partially attributed to gene amplification45. It is possible that enhancer hijacking may serve as an additional mechanism for CCND1 overexpression in breast cancer.

The hijacked enhancers on chromosome 8 in ZR751 exhibit H3K27ac enrichment in the breast cancer cell line T47D without the translocation (Figure S4A), but little enrichment in the B-cell lymphoma line REC1 in which CCND1 hijacks enhancers from chromosome 14 (Figure 3C). The reciprocal pattern was observed for the enhancers hijacked in REC1 (Figure S4A). These results indicate that the enhancers involved in hijacking are already active in cancer cells without the translocation and are cancer-type specific. To support this finding, we utilized TCGA ATAC-seq data to estimate potential regulatory activity of translocated enhancers we identified in several breast, prostate, and colon cancer cell lines. Overall, these enhancers have increased ATAC-seq signal in tumors of their respective cancer types compared to others (Figure S4B). These data together suggest that cancer-type specific enhancers may be subject to translocations to activate oncogene expression.

HAPI analysis identifies enhancer hijacking associated with known ecDNAs

Some of the identified enhancer hijacking genes exhibit high copies (log2(CN)>2). Recent studies have revealed that oncogenes amplified in the form of ecDNAs tend to exhibit higher copies compared to their chromosomal amplifications46. As illustrated in Figure 4A, the high copy number of ecDNAs, along with increased chromatin accessibility, results in the overexpression of oncogenes such as EGFR in glioblastoma46. In addition, enhancer hijacking, as a result of complex rearrangements on ecDNAs, also contributes to oncogene expression as exemplified by MYCN in neuroblastoma15. Among our analyzed cell lines, the colorectal cancer cell line COLO320DM and the gastric adenocarcinoma cell line SNU16 are known to contain ecDNAs amplifying the MYC oncogene47. HAPI analysis identified enhancer hijacking events in both the cell lines, which involve multiple oncogenes on the same ecDNA.

Figure 4:
Figure 4: HAPI analysis identifies enhancer hijacking on known ecDNAs.

A. Schematic models illustrating the identification of enhancer hijacking events on ecDNAs that involve multiple oncogenes from different chromosomal regions.

B. Plotted are the calculated trans-enhancer contribution (x-axis) and the copy number estimated by SNP-array data (y-axis) for each HAPI gene in COLO320DM cells.

C. HiChIP, ChIP-seq, and SNP-array-based copy number signal at the DNA segments of the ecDNA found in COLO320DM cells. The HiChIP data and the ecDNA segments in COLO320DM were published by Hung et al47.

D. Enhancer distribution of MYC and the additional oncogenes harbored in the ecDNA found in COLO320DM cells. Enhancers within 2mb of distance to the promoter of MYC or any other gene in the endogenous chromosomal locus will be considered as the gene’s ‘own’ enhancers, otherwise will be considered as ‘other’ enhancers. Source data are provided as a Source Data file.

E. Circos plot: layer 1=chromosome, 2=CNV, 3=enhancers (grey) and super-enhancers (black), 4=HAPI TSS, 5=ecDNA segments, center=E-P loops (black) and WGS-predicted translocations (red) in COLO320DM cells.

F-I. Same as B-E but with regards to the ecDNAs in SNU16 cells. The HiChIP data and the ecDNA segments in SNU16 were published by Hung et al47. Note that the APIP and PDHX genes share the same promoter anchor for the HiChIP analysis. Source data for the panel H are provided as a Source Data file.

COLO320DM contains an ecDNA harboring MYC, a PVT1-MYC fusion, and several other genes including CDX2 and PDX147. As a HAPI gene, MYC has most of its enhancer activity contributed in cis. Interestingly, CDX2, a colorectal cancer lineage-survival oncogene48,49 from chromosome 13, is identified as a trans-enhancer hijacking HAPI gene (Figure 4B), with over 95% of enhancer activity in trans. Indeed, most of the enhancer activity for CDX2 comes from the MYC locus on the ecDNA (Figures 4C-E). These results suggest two separate mechanisms through which the ecDNA causes overexpression of oncogenes in COLO320DM cells – while MYC uses its co-amplified endogenous enhancers, CDX2 hijacks MYC enhancers on the same ecDNA to enhance its expression, representing an “opportunistic” enhancer hijacking event (Figure 4D).

SNU16 contains multiple ecDNA isoforms harboring the MYC and FGFR2 oncogenes separately or together47,50. The MYC ecDNA harbors three other genes from chromosome 11, including APIP, PDHX, and CD44, which are involved in the development of gastric adenocarcinoma51,52. HAPI analysis identified these three genes as enhancer hijacking genes (Figure 4F), which have 34%-42% of enhancer activity contributed from enhancers near MYC on chromosome 8 that are now on the same ecDNA (Figures 4G-I). Vice versa, MYC also has most of the enhancer activity contributed in trans from other chromosomal regions on the same ecDNA (Figure 4G-I). These reciprocal enhancer-oncogene interactions represent a “mutualistic” enhancer hijacking event on ecDNAs (Figure 4H). Interestingly, although most of the FGFR2 ecDNA does not include regions from other chromosomes based on previous reports47, it has around half of enhancer activity contributed in trans (Figure 4F-G). This is likely caused by the reported intermolecular interactions between MYC and FGFR2 ecDNAs as well as the presence of chimeric MYC-FGFR2 ecDNA in a small proportion of the ecDNAs in this cell line47.

Enhancer hijacking among multiple oncogenes is prevalent on ecDNAs and complex chromosomal amplicons

We reasoned that enhancer hijacking, together with high copy numbers of the involved genes, may indicate a unique signature for chimeric ecDNAs or complex chromosomal amplicons that contain multiple DNA segments in our HAPI analysis results. Indeed, we found three additional cell lines NCIH2170, NCIH446, and MCF7 that have clusters of highly amplified HAPI genes hijacking enhancers from different chromosomes (Figure 5A-C). We analyzed their WGS data using AmpliconArchitect, an algorithm to reconstruct amplicons53, which showed extensive rearrangements among the amplified regions harboring the enhancer-hijacking genes (Figure S5A-C).

Figure 5:
Figure 5: Enhancer hijacking is prevalent on complex amplicons.

A-C. Plotted are the calculated trans-enhancer contribution (x-axis) and the copy number estimated by SNP-array data (y-axis) for each HAPI gene in NCIH2170, NCIH446, and MCF7 cells.

D. HiChIP, H3K27ac ChIP-seq (derived from HiChIP reads), and copy number profiles at DNA segments of the MYC ecDNA in NCIH2170 cells.

E. Enhancer distribution of MYC and the additional oncogenes harbored in the ecDNA found in NCIH2170 cells. Source data are provided as a Source Data file.

F. HiChIP, H3K27ac ChIP-seq, and copy number profiles at DNA segments of the MYC amplicon in NCIH446 cells.

G. Enhancer distribution of MYC and the additional oncogenes harbored in the amplicon found in NCIH446 cells. Source data are provided as a Source Data file.

H. Summary of AmpliconArchitect data from PCAWG tumor samples: most of oncogene amplicons contain additional trans or cis DNA segments. 37.6% of these complex amplicons contain multiple oncogenes originating from different DNA segments.

In the lung squamous cell carcinoma cell line NCIH2170, we identified a cluster of trans enhancer hijacking genes from chromosomes 8 and 17 (Figure 5A). AmpliconArchitect analysis of WGS data predicted an ecDNA containing most of these HAPI genes and their hijacked enhancers (Figure S6A-B). In particular, we found that ERBB2 (chromosome 17) and its surrounding oncogenes MIEN1 and IKZF3 strongly interact with MYC’s enhancers (Figure 5D-E). ERBB2 was known to be an amplified and overexpressed gene in this cell line54,55. Our analysis showed that, in addition to the copy number amplification, hijacking MYC enhancers on the same ecDNA may also contribute to ERBB2 and its surrounding oncogenes’ overexpression.

Additionally, in the small-cell lung cancer cell line NCIH446, HAPI analysis identified several genes including known oncogenes MYC and NFIB with trans enhancer hijacking from regions of chromosomes 8 and 9 (Figure 5F-G). WGS analysis showed translocations among these gene loci (Figure S6C). A recent study reported that the MYC locus in NCIH446 resides on the chromosome based on DNA Fluorescent In Situ Hybridization (FISH) results56, which together with our results suggest that complex chromosomal amplicons are also capable of causing multiple oncogenes to hijack each other’s enhancers.

Lastly, we identified 36 HAPI genes with trans enhancer hijacking in the ERα-positive breast cancer cell line MCF7, most of which are clustered within genomic regions of chromosomes 17 and 20 (Figure S7A). These include known oncogenes TBX2 and BRIP1 from chromosome 17, and ZNF217 from chromosome 20.

Accordingly, ampliconArchitect analysis predicted an ecDNA harboring most of these HAPI genes and their hijacked enhancers (Figures S7B-C). These oncogenes are highly interactive with each other’s enhancers, indicating extensive enhancer hijacking spanning the ecDNA segments (Figures S7D).

In the aforementioned five cell lines that harbor ecDNAs or chromosomal complex amplicons, we observed both “opportunistic” (COLO320DM and NCIH2170) and “mutualistic” (MCF7, SNU16 and NCIH446) enhancer hijacking phenomena among multiple oncogenes on the same amplicon, which are associated with overexpression of the associated oncogenes (Figure S8A). Using publicly available AmpliconArchitect results from WGS data of primary patient tumors included in the PCAWG project57, we identified 1299 amplicons containing previously annotated oncogenes. We found that most of these amplicons, either in the extrachromosomal or chromosomal forms, harbor additional DNA segments originating from other chromosomal regions. In 37.6% of these complex amplicons, we found multiple oncogenes from separate DNA segments that are now on the same amplicon (Figure 5H). One amplicon is predicted to be a MYC-ERBB2 chimeric ecDNA, similar to what we report in NCIH2170 cells (Figure S8B). These analyses suggest that many of these amplicons could enable extensive enhancer hijacking among their harbored oncogenes in patient tumors.

Functional exploration of chimeric ecDNA-associated enhancer hijacking

To validate the presence of chimeric ecDNAs that contain multiple oncogenes from different chromosomes, we performed DNA FISH in NCIH2170 cells using two probes targeting the MYC and ERBB2 regions separately. DNA FISH on metaphase spreads showed the presence of ecDNAs containing both MYC and ERBB2 across examined cells (Figure 6A). However, some of the ecDNAs contain only MYC or ERBB2, suggesting the intracellular heterogeneity of ecDNA compositions similar to what was reported previously in SNU1647. DNA FISH in interphase nuclei showed similar overlap between MYC and ERBB2 signal across the examined cells (Figure S9A). We also observed the presence of ecDNA aggregates in some of the interphase nuclei, indicative of ecDNA hubs as previously reported in other cancer cell lines47 (Figure S9B). These results demonstrate the presence of highly abundant MYC-ERBB2 chimeric ecDNA in NCIH2170 cells.

Figure 6:
Figure 6: Functional exploration of the MYC-ERBB2 chimeric ecDNA.

A. Representative images of metaphase chromosome spreads showing highly amplified copies of MYC (red) and ERBB2 (green) colocalizing outside of chromosomal DNA (blue, DAPI staining) in NCIH2170 cells. In total, 20 metaphase spreads of high quality were collected from 7 separate images, all indicating the presence of MYC-ERBB2 chimeric ecDNAs in high abundance. Scale bar, 5 µm. Four enlarged regions of interest (ROI) were highlighted. Scale bar, 2 µm.

B. AUC values representing the sensitivity of lung squamous carcinoma cell lines to Neratinib and Lapatinib. Source data are provided as a Source Data file.

C. Schematic illustration of the experimental design of the MYC-promoter CRISPRi assay.

D. Virtual 4C signal (derived from HiChIP data) showing interactions with the MYC promoter at the chromosome 8 locus (now on ecDNA) in NCIH2170 cells (sgNT2, replicate 1).

E. Virtual 4C signal between the promoter of MYC (upper) or ERBB2 (lower) and five highlighted enhancer regions representing different levels of interaction intensity in NCIH2170 cells with and without MYC promoter repression (sgMYC-prom vs. sgNT2). SEM: standard error of the mean. N=2 biological replicates.

F. The averaged virtual 4C signal between the top 50 MYC-interacting regions from the MYC locus (ranked by their interactions with the MYC promoter) and the promoter of MYC (left) or ERBB2 (right).

G. A volcano plot presenting RNA-seq results from NCIH2170 cells with and without MYC promoter repression. Two separate sgRNAs were used for each condition (sgNT1/sgNT2, sgMYC-prom#1/ sgMYC-prom#2). Highlighted are the genes on the MYC-ERBB2 ecDNA that are significantly differentially expressed.

The presence of chimeric MYC ecDNAs may provide novel therapeutic opportunities. For instance, although MYC as a transcription factor is difficult to target by small molecules, ERBB2 is considered highly targetable. Indeed, NCIH2170 exhibits the strongest sensitivity to HER2 inhibitors lapatinib and neratinib among all the lung squamous cancer cell lines tested in the Cancer Therapeutics Response Portal (CTRP) CTD2 project (Figure 6B).

We then sought to assess the regulatory function of enhancers on the MYC-ERBB2 ecDNA. Due to the large number of potential MYC enhancers that are shared by these genes (Figure 5D), individually repressing these enhancers may not reflect their combinatorial activity. As an alternative strategy, we used CRISPRi to repress the MYC promoter and released all MYC enhancers from MYC to assess their effect on expression of ERBB2 and its neighboring genes that also use these enhancers (illustrated in Figure 6C). HiChIP assays showed that CRISPRi resulted in an overall decrease in interactions between the MYC promoter and its enhancers, as expected (Figure 6D-F). In contrast, ERBB2 gained interactions with these enhancers after CRISPRi of the MYC promoter (Figure 6E-F). RNA-seq showed that CRISPRi represses MYC expression and causes a significant increase in expression of ERBB2 (2.7 folds) and its neighboring genes MIEN1, PGAP3, and GRB7 (Figure 6G). For the remaining affected genes, we found enrichments of MYC up- and down-regulated gene sets, suggesting that they are secondary effects of MYC repression (Figure S9C). All these results demonstrate that, on the chimeric MYC-ERBB2 ecDNA in NCIH2170 cells, while MYC retains its interaction with its endogenous enhancers, ERBB2 and its neighboring genes opportunistically hijack MYC enhancers on the same ecDNA for their transcriptional activation.

DISCUSSION

Enhancer hijacking caused by structural alterations is an oncogenic event in cancer; however, it remains difficult to detect due to the complexity of cancer genomes. Here, we presented a two-step strategy, namely HAPI, to robustly detect enhancer hijacking events based on: 1) oncogenes subject to enhancer hijacking should be highly regulated by enhancers, 2) the hijacked enhancers should contribute to a substantial proportion of an oncogene’s enhancer activity. This assumption is supported by recent work using CRISPR to generate de novo enhancer translocations to the MYC locus, which showed that the combination of enhancer activity and contact from the translocated regions determines MYC activation58. Although our analysis was based on H3K27ac HiChIP data, we expect it to be compatible with other types of high-resolution enhancer-promoter contact mapping assays such as PLAC-seq, Micro-C, and ChIA-PET59–61.

We first found hundreds of HAPI genes in each cell line whose promoters are highly interactive with enhancers. Previous work focused on neural cells found that genes that specify lineage development exhibit this pattern62. Our analysis in cancer cell lines found that oncogenes are also enriched in this category. We then prioritized HAPI genes that have a substantial amount of enhancer activity originating from other chromosomes or abnormally distant intra-chromosomal regions. This led to the identification of known and potentially novel enhancer-hijacking events regardless of the underlying structural variants such as balanced translocation, cryptic insertion, large deletion, as well as ecDNA formation. Although several of the identified oncogenes such as MYC, ETV1, and CCND1 have been previously linked to enhancer hijacking, we identified their novel hijacked enhancers that are cancer type specific. The analysis also identified potentially novel enhancer-hijacking oncogenes such as CRKL and ID4. We have made the HAPI analysis tool publicly available (URL: https://codeocean.com/capsule/2804329) and anticipate its broader application in cancer driver discovery and mechanism investigation.

In addition to individual enhancer-hijacking events caused by simple DNA rearrangements, we identified nested enhancer-hijacking on complex amplicons, such as ecDNAs, that involve multiple oncogenes originating from different chromosomal regions. Our analysis of previously characterized as well as our newly identified ecDNAs shows that ecDNAs may contain multiple oncogenes from different chromosomes, which we refer to as chimeric ecDNAs. These structures were predicted by WGS analysis and validated by either optical imaging from previous work47,63 or dual-color DNA FISH in this study. We revealed two types of enhancer hijacking events associated with chimeric ecDNAs, as exemplified by MYC ecDNAs in this study. For the MYC ecDNAs found in COLO320DM and NCIH2170 cells, while MYC mainly uses its own enhancers, the additional oncogenes such as CDX2 and ERBB2 that are translocated to the MYC ecDNAs heavily use MYC enhancers for their activation, representing an “opportunistic” enhancer hijacking event. For the MYC ecDNA found in SNU16 cells, MYC and the co-amplified CD44 gene use each other’s enhancers in a more balanced manner, representing a potentially “mutualistic” enhancer hijacking event. These results provide important insights about another layer of complexity for oncogene amplification and overexpression.

Chimeric ecDNAs may be prevalent in cancer. A recent work reported that 33.1% of ecDNAs found in Barret’s oesophagus and esophageal adenocarcinomas contain multiple oncogenes on the same molecule based on WGS analysis64. In addition to ecDNAs, chromosomal translocations often precede amplifications, also causing co-amplification two oncogenes from distinct chromosomal regions65. Characterization of these chromosomal or extrachromosomal complex amplicons may reveal novel therapeutic strategies. As exemplified in this study, although MYC is a notoriously difficult drug target, ERBB2 that is co-amplified with MYC is highly targetable.

Our observation of potentially “mutualistic” or “opportunistic” enhancer hijacking events on different chimeric ecDNAs has important implications for the future of CRIPSR-based therapeutics being developed to silence oncogenes. Targeting the shared enhancers on chimeric ecDNAs that contribute to overexpression of multiple oncogenes is an appealing therapeutic strategy to repress multiple oncogenes simultaneously. However, we also demonstrate that repressing the MYC promoter on a chimeric ecDNA released the shared enhancers from MYC and increased ERBB2’s opportunistic enhancer hijacking and expression. This could be a strategy to further sensitize the tumor to HER2 inhibitors, or it could be disastrous in increasing growth factor driven proliferation. These examples illustrate the importance of identifying and characterizing these complex regulatory events involving multiple oncogenes in order to develop effective CRISPR-based therapeutic strategies.

One limitation of our work is that we primarily used cell lines as our experimental models because they are relatively homogenous and easy to engineer. We used WGS data from primary human tumors in PCAWG to support our conclusions about the presence of complex rearrangements involving multiple oncogenes but the functional validation of enhancer hijacking in primary human tumors requires further development of 3D genomics assays and analysis methods that account for their intrinsic heterogeneity. In addition, we focused on trans-enhancer hijacking events that are caused by structural alterations. Many cis-enhancer hijacking events may be caused by other types of genomic alterations such as point mutations and small indels as well as epigenetic alterations such as DNA methylation that are capable of disrupting TAD domains, which fall outside the scope of this manuscript. Future efforts integrating additional model systems and genomic/epigenomic analysis may address these limitations.

In summary, applying a two-step HAPI analysis, we identified and characterized enhancer-hijacking events that are caused by chromosomal and extrachromosomal structural alterations in cancer. Our findings reveal enhancer-hijacking events underlying the activation of multiple oncogenes that reside together on complex amplicons such as ecDNAs.

METHODS

Cell lines

Prostate cancer cell lines LNCaP, MDAPCA2B, 22Rv1, and VCaP, ERα-positive breast cancer cell lines ZR751 and MCF7, lymphoma cell lines REC1 and DOHH2, lung squamous cell line NCIH2170, and small cell lung cancer cell lines NCIH446 and NCIH1876 were obtained from the Cancer Cell Line Encyclopedia (CCLE)66,67. The lymphoma cell line CUTLL1 was obtained from Sigma. VCaP, MCF7, and MDAPCA2B cells were grown in DMEM media supplemented with 10% FBS. NCIH1876 cells were grown in DMEM/F12 media supplemented with 4 µg/ml Hydrocortisone, 5 ng/ml murine EGF, 1× Insulin-Transferrin-Selenium, 10 nM of beta-estradiol, and 10% FBS. The rest of the cell lines were grown in RPMI media supplemented with 10% FBS. The identities of the cell lines were verified by either SNP fingerprinting as previously described in the CCLE project66,67 or short tandem repeat analysis through IDEXX. All the cell lines were tested negative for mycoplasma.

HiChIP and loop calling

HiChIP assays were performed for the following cell lines: CUTLL1, REC1, VCaP, MDAPCA2B, 22Rv1, ZR751, MCF7, DOHH2, NCIH446, NCIH1876, and NCIH2170 as described in the published protocol68 with minor modifications described previously69. Cross-linked chromatin was digested with the MboI restriction enzyme, filled with dNTPs including biotin labeled dATP, ligated with T4 DNA ligase, and sonicated to obtain chromatin fragments averaging ∼1 kb in length using Qsonica (Q800). To enrich for chromatin interactions at active regulatory elements, antibodies of H3K27ac (Abcam, ab4729, rabbit polyclonal, 7.5 µg/HiChIP) were used for DNA fragment capture. Magnetic streptavidin beads were applied to pull down fully ligated DNA fragments and HiChIP libraries were prepared using Illumina Tagment DNA Enzyme and Buffer Kit. Sequencing was done with Illumina NextSeq or NovaSeq. HiChIP data from the other cell lines included in the study was obtained from the NCBI Gene Expression Omnibus (GEO) (Supplementary Table 1).

The paired-end HiChIP sequencing reads were aligned to human genome hg19 with the HiC-Pro pipeline70. Hichipper was then used to call chromatin loops where DNA anchors were assigned, and the number of PETs connecting the anchors71. Loop anchors were either called by hichipper based on enrichment of HiChIP sequencing reads or derived from broadpeaks called by MACS2 using publicly available H3K27ac ChIP-seq data in the corresponding cell lines, as indicated in Supplementary Table 1. If multiple biological replicates of HiChIP data are available, the reads of replicates were merged prior to analysis. Chromatin loops were removed if their PETs <3 or anchors overlap with ENCODE blacklist regions72. HiChIP data were visualized using the R package gTrack3 (URL: https://github.com/mskilab/gTrack) or RCircos73. The gene positions in gTrack plots are based on the GENCODE database.

HAPI Analysis

We grouped the loops, called by hichipper, into ‘normal,’ ‘abnormal cis’, and ‘trans’ groups. Intra-chromosomal chromatin interactions spanning more than 5 kb and less than 2 mb were considered normal loops. Intra-chromosomal interactions, whose anchors are separated by at least 2 mb were considered abnormal long-range cis chromatin loops. Inter-chromosomal interactions were treated as trans chromatin loops. The H3K27ac HiChIP anchors were then intersected with +/-2.5 kb of transcription start sites (NCBI RefSeq gene list downloaded from UCSC browser in 2019) using bedtools74 to identify promoters and enhancers. We consider a HiChIP loop that has a promoter in only one of its anchors as an enhancer-promoter loop. As promoters may act as enhancers or repressors for another gene75–77, we excluded promoter-promoter interactions for our analysis.

To identify highly interactive genes, all types of enhancer-promoter interactions (normal, abnormal cis, and trans) described above were combined. Then for every gene, two scores were calculated to account for the number and strength of interactions between a promoter and its enhancers. The first score, defined as an enhancer contact value, is the number of enhancer anchors that are looped to the promoter of a gene. The second score, defined as interaction intensity value, is the summation of the paired-end tags (PETs) that support the interactions of the promoter and its enhancer anchors. In the case of genes that have multiple promoters, the promoter with the most PETs stemming from it is used as the primary anchor for enhancer-promoter interactions. Genes are then ranked based on these scores individually and are presented in hockey stick plots. The inflection points of these plots, defined as the point where the line with a slope of ((max value-min value)/number of values) is tangent to the curve, are the cutoff values to define genes with a significant enhancer contacts values or interaction intensity. Genes with both scores above their respective cutoffs are defined as HAPI genes.

To determine what types of enhancers are interacting with HAPI genes, the trans and long-range cis interaction data sets with enhancer contact values and interaction intensity calculated for each gene as described above are used. Then a percent enhancer contribution is calculated for each HAPI gene using the following formula for both trans and long-range cis interaction data: Embedded Image Using this formula, genes with a higher trans enhancer contribution have a higher number of interacting enhancers and stronger interaction intensity coming from a different chromosome. Similarly, genes with a higher abnormal enhancer contribution have a higher number of enhancers and stronger interaction intensity coming from regions greater than 2 mb away on the same chromosome. We focus on the genes that have more than 25% of enhancer activity contributed in trans or cis.

To calculate the contributions of enhancers hijacked from the MYC locus to a target gene (Figures 4D, 4H, 5E, and 5G), we used the same formula but classified the enhancers into “own enhancers” (cis-interacting enhancers within 2 mb of the target gene’s promoter), “hijacked MYC enhancers” (trans-interacting enhancers hijacked from regions within 2 mb of the MYC promoter), and “other enhancers”.

Copy number analysis

SNP-array-based copy number segment data of cancer cell lines were downloaded from the Broad Institute CCLE Portal (URL: https://data.broadinstitute.org/ccle_legacy_data/dna_copy_number/CCLE_copynumber_2013-12-03.seg.txt; DOI: https://doi.org/10.1038/ng.3984; downloaded in November, 2020). For the CUTLL1 cell line that is not included in the CCLE copy number dataset, we used Neoloop20 to infer the copy number profile using HiChIP reads.

Clustering analysis

An interaction score was calculated for all HAPI genes in each cell line by multiplying the log10 of the interaction intensity (PETs) by the enhancer contact value (interaction score = log10(interaction intensity) * (enhancer contact value)). The clustering analysis was based on Spearman pairwise sample correlation. Heatmap was generated using the heatmap.2 function in the gplots R package.

Pathway analysis and oncogene enrichment analysis

For HAPI genes that are shared by at least half of the cell lines in each cancer type, we performed pathway analysis and oncogene enrichment analysis. For stomach adenocarcinoma, since only one cell line (SNU16) was included in our analysis, all its HAPI genes are used to represent this cancer type. GSEA analysis (MSigDB)78,79 was performed to identify the top KEGG legacy pathways80,81, Reactome pathways82, and Gene Ontology molecular functions83,84 enriched in the shared HAPI genes. As the GSEA ‘compute overlaps’ tool requires no more than 500 genes as the input, for prostate cancer cell lines that have 548 shared HAPI genes identified, we ranked the genes based on their interaction scores (averaged across the corresponding cell lines) and used the top 500 genes for the analysis. The shared HAPI genes and other genes in each cancer type were also compared to a list of known oncogenes merged from three previously published oncogene data sets34–36. Fisher exact tests were used to calculate oncogene enrichment in the HAPI genes.

RT-qPCR

The Zymo Quick-RNA miniprep kit paired with on-column DNase I treatment was used to extract total RNA. To make cDNA for RT-qPCR, 1μg of RNA was used with the NEB LunaScript SuperMix kit. Real-time PCR was completed with NEB Luna Universal qPCR Master Mix on a Bio-Rad CFX96 qPCR instrument with technical replicates. The qPCR signal was first normalized to the internal reference gene CTCF and then the non-targeting sgRNA (sgNT1). RT-qPCR primers are listed in Supplementary Table 3.

RNA-seq

CCLE RNA-seq data were downloaded from the DepMap portal (URL: https://depmap.org/portal/; downloaded in May, 2022)66. Additional RNA-seq data for the CUTLL1 cell line was downloaded from GSE61999 and aligned to hg19 using Bowtie285 and gene expression was calculated using RSEM86. The log2(TPM+1) values of CUTLL1 were merged with the CCLE dataset for comparison.

For determining androgen dependence of ETV1 expression in LNCaP and 22RV1, RNA-seq data for both cell lines grown in androgen-depleted media treated with DHT or vehicle were downloaded from GSE92574. Counts per million reads were used to determine the log2 fold change in expression between control and treated cells. We generated RNA-seq data for the CRISPRi experiment in NCIH2170 cells and performed edgeR analysis87 to identify differentially expressed genes (FDR<0.05; Fold change>1.5).

ChIP-seq

ChIP-seq was performed in LNCaP-dCas9-KRAB-MeCP2 cells treated with enhancer-targeting or non-targeting sgRNAs, as previously described69. Five million cells were crosslinked with 1% formaldehyde (diluted in 1× PBS) and lysed with Lysis Buffer I (5 mM PIPES pH 8.0, 85 mM KCl, 0.5% NP40) and then Lysis Buffer II (1× PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate) supplemented with protease inhibitors. Chromatin extract was sonicated with Bioruptor (20 min; pulse: 30 sec on/30 sec off; high amplitude) and immunoprecipitated with H3K27ac antibody (Abcam, ab4729, rabbit polyclonal, 4 µg/ChIP) premixed with Dynabeads A and G. ChIP-seq libraries were prepared using NEBNext DNA Ultra II library prep kit and sequenced by Illumina MiSeq.

Remaining ChIP-seq data used in the study were obtained from previously published datasets (Supplementary Table 1). ChIP-seq reads and genomic input reads, if available, were aligned to the hg19 using Bowtie285. Samtools88 was used to sort and index aligned reads followed by MACS289 to call significant peaks (q value < 0.05) and calculate the ChIP-seq signal. Super-enhancers were called using H3K27ac ChIP-seq signal using the Homer pipeline90. We focused on super-enhancers that contain at least one enhancer anchor of a HiChIP loop and linked them to gene promoters utilizing the pairToBed function of bedtools74.

CRISPR-mediated enhancer repression

To repress enhancers via CRISPR, we used a lentiviral dCas9-KRAB-MeCP2 vector that was previously generated28. Cells were infected with the vector and selected with blasticidin (10 µg/ml) for >5 days to stably express dCas9-KRAB-MeCP2. sgRNAs were designed to target the peaks of DNase I hypersensitive regions and the trough of H3K27ac ChIP-seq profiles for either enhancer or promoter regions of interest. The promoter- or enhancer-targeting sgRNAs and non-targeting sgRNAs with no genome recognition sites were cloned into LentiGuide-Puro (Addgene: 52963). The cells stably expressing dCas9-KRAB-MeCP2 were infected with these vectors and then selected with puromycin (2 µg/ml) for at least 3 days before extracting RNA. All sgRNA sequences used are listed in Supplementary Table 3.

WGS data processing and structural variant calling

WGS data for LNCaP, MCF7, ZR751, NCIH446, NCIH2170, COLO320DM, and SNU16 were downloaded from published studies (Supplementary Table 1). Raw FASTQ files were aligned to hg19 using BWA-MEM91 and the aligned reads were sorted and indexed with Samtools88. Structural variants and their breakpoints were called using GRIDSS92,93. The output VCF files were converted to bedpe files and filtered by QUAL ≥ 1000, AS>0, and RAS>0 for translocation visualization using the Bioconductor package StructuralVariantAnnotation94.

PCAWG structural variants

PCAWG consensus call sets of structural variants in tumor samples were downloaded from the ICGC Data Portal (URL: https://dcc.icgc.org/releases/PCAWG/consensus_sv; release date: Nov 25, 2019) and presented using RCircos73.

For PCAWG complex amplicon analysis, we used the publicly available AmpliconArchitect results released from the AmpliconRepository (URL: https://ampliconrepository.org; downloaded in October 2023). We used the “aggregated results” file in the “PCAWG – full” dataset released on August 29, 2023, which listed the DNA segments included in each amplicon. To identify DNA segments harboring oncogenes, we used a list of known oncogenes merged from three previously published oncogene data sets34–36.

ecDNA calling

To call ecDNAs in CCLE cell lines we employed the AmpliconSuite pipeline (URL: https://github.com/jluebeck/AmpliconSuite-pipeline). Briefly, WGS data were aligned to hg19 using BWA91 and amplified regions were called with cnvkit95. Seed regions for AmpliconArchitect53 were identified using the default settings of the prepareAA script. AmpliconArchitect53 was used to identify ecDNAs. AmpliconClassifier64 was used to determine if the amplicons represent ecDNAs and CycleViz (URL: https://github.com/jluebeck/CycleViz) was used to visualize the reconstructed ecDNA.

TCGA ATAC analysis

We downloaded the TCGA ATAC-seq data96 from the UCSC Xena Browser (URL: https://xenabrowser.net/datapages/?dataset=TCGA_ATAC_peak_Log2Counts_dedup_sample&host=https%3A%2F%2Fatacseq.xenahubs.net&removeHub=https%3A%2F%2Fxena.treehouse.gi.ucsc.edu%3A443; version: 2018-06-21), which contains log2-transformed normalized ATAC-seq insertion counts for each tested sample across the identified pan-cancer ATAC peaks. To compare chromatin openness of the hijacked enhancers across different cancer types, we first took enhancers that are hijacked by HAPI genes in each selected cell line, and then summed the log2 normalized ATAC-seq counts from ATAC peaks overlapping with the hijacked enhancers to determine its overall chromatin openness for each sample included in the tested cancer types.

Dual-color DNA FISH

Metaphase arrest of NCIH2170 cells was performed by incubation with KaryoMAX (Gibco) at 0.1 µg/mL overnight. Single cell suspension of the metaphase arrested cells was resuspended in 1xPBS followed by incubation with 75 mM KCL for 15-30 min. The cells were fixed with 3:1 methanol to glacial acetic acid overnight and then washed three times with the fixative before being dropped on a coverslip. The coverslip was dried for 5 min at 60°C before being immersed in 2xSSC in preparation for DNA FISH.

Coverslips with fixed cells in metaphase were washed twice for 5 min with 2xSSC. The coverslips were then washed with 50% formamide/1×SSC for 4 min and dehydrated through a series of ethanol incubations (70%, 85%, and 100%) for 2 min each. Next, the coverslips were placed on slides containing a 1:1 ratio of ERBB2 to MYC FISH probes diluted in the manufacturer’s hybridization buffer (Empire Genomics). The coverslips were sealed with rubber cement and denatured in a heated incubator at 82°C for 1 min, followed by incubation in a humidified chamber at 37°C overnight. The coverslip was washed once with 0.4×SSC/0.3% Igepal at 73°C and twice with 2×SSC/0.1% Igepal each for 2min. A final two washes with 2xSSC were performed before mounting the coverslip using Prolong Antifade Mountants with DAPI (Thermo Fisher).

For interphase DNA FISH, NCIH2170 cells were seeded on collagen-coated coverslips and fixed with 4% PFA in 1×PBS for 10 min, followed by quenching for 10 min with 0.1 M Glycine in 1×PBS. The cell membrane was then permeabilized for 10 min with 0.5% Triron-X 100. An additional three 5 min rinses with 1xPBS proceeded RNase A treatment at 0.1 ug/mL for 1 hour at 37°C, followed by three more rinses with 1×PBS before a 10 min incubation with 0.7% Triton X-100 in 0.1 M HCl on ice. After three rinses with PBS, the coverslips were incubated for 30 min in 2x SSC with 50% formamide. During the incubation, the hybridization buffer containing 2×SSC, 1 µl each of MYC and ERBB2 FISH probes (Empire Genomics), 50% (v/v) formamide, 2% dextran sulfate, and 0.2 µl 10 mg/ml salmon sperm DNA was prepared and 10 µl was placed on the slide. The coverslips were then placed on the hybridization buffer on the slide and sealed with rubber cement. Denaturation was performed in a heated incubator at 78°C for 4 min followed by incubation overnight at 37°C in a humidified chamber. The coverslips were then washed three times with 50% formamide in 2×SSC for 5 min each at 37°C. An additional three washes with 2×SSC at 37°C were performed before mounting with Prolong Antifade Mountants with DAPI (Thermo Fisher).

All DNA FISH samples were imaged using a Nikon Eclipse Ti-2 microscope coupled with a CSU-W1 spinning disk confocal scanner unit (Yokogawa). Specifically, the fluorescent cell samples were excited using three laser lines: 405 nm, 488 nm, and 561 nm. The fluorescence signals were captured and detected using a Plan Apo 60x objective (Nikon, NA 1.40) and a high-speed Kinetix sCMOS camera (Photometrics), respectively. This microscope utilizes the Nikon Perfect Focus System, an interferometry-based focus lock that allows the capture of multipoint images without loss of focus. All imaging data were examined and quantified using Fiji and Cell Profiler.

Virtual 4C Analysis

For HiChIP data from NCIH2170 cells with and without CRISPRi-mediated MYC promoter repression, we presented and compared their virtual 4C signal anchored at MYC or ERBB2 promoters. Briefly, we first identified all valid read pairs (HiC-Pro output) that have one end mapped to a 2kb window centered at MYC or ERBB2 TSS. We then used the other end of the read pairs to generate a bedgraph file at a 10bp resolution. The bedgraph signal was then scaled based on the total number of PETs identified in each condition. As most of enhancers used by MYC and ERBB2 come from the chromosome 8 locus surrounding MYC, we presented Virtual 4C signal at five regions from the MYC locus with a range of enhancer activity as well as the averaged Virtual 4C signal for the top 50 HiChIP signal-derived peaks at the MYC locus that are ranked based on their interactions with the MYC promoter.

DATA AVAILABILITY

The publicly available H3K27ac HiChIP data used in this study were obtained from the NCBI Gene Expression Omnibus (GEO) at GSE227242 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE227242], GSE97585 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE97585], GSE157381 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE157381], GSE147854 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE147854], GSE166232 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE166232], GSE151002 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE151002], GSE159985 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE159985], and GSE188401 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE188401] (specific cell lines are described in Supplementary Table 1). The publicly availably ChIP-seq data used in this study were downloaded from GEO at GSE73994 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE73994], GSE63109 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE63109], GSE69558 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE69558], GSE105760 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE105760], GSE105627 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE105627], GSE85158 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE85158], GSE57436 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE57436], GSE86743 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE86743], GSE73319 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE73319], GSE115123 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE115123], and GSE159972 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE159972] for H3K27ac, and GSE174109 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE174109] and GSE94013 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE94013] for AR ChIP-seq (specific cell lines are described in Supplementary Table 1). TCGA copy number segment data for all applicable cell lines were downloaded from the Broad Institute Portal (URL: https://data.broadinstitute.org/ccle_legacy_data/dna_copy_number/CCLE_copynumber_2013-12-03.seg.txt).

CCLE RNA-seq data were downloaded from the DepMap portal (URL: https://depmap.org/portal/). RNA-seq data for CUTLL1 was downloaded from GSE61999 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE61999]. Publicly available WGS data for cell lines were downloaded from NCBI SRA: PRJNA361316 [https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA361316], PRJNA523380 [https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA523380], and PRJNA506071 [https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA506071] (specific cell lines are described in Supplementary Table 1). PCAWG consensus call sets of structural variants in tumor samples were downloaded from the ICGC Data Portal (URL: https://dcc.icgc.org/releases/PCAWG/consensus_sv). Complex amplicon data from PCAWG were downloaded from the AmpliconRepository (https://ampliconrepository.org). Publicly available TCGA ATAC-seq data were downloaded from the UCSC Xena Browser (URL: https://xenabrowser.net/datapages/?dataset=TCGA_ATAC_peak_Log2Counts_dedup_sample&host=https%3A%2F%2Fatacseq.xenahubs.net&removeHub=https%3A%2F%2Fxena.treehouse.gi.ucsc.edu%3A443). The ChIP-seq, HiChIP, and RNA-seq data generated in this study have been deposited to GEO under the series GSE228247 [https://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/geo/query/acc.cgi?acc=GSE228247]. The remaining data are available within the Article, Supplementary Information, Supplementary Data, or Source Data file. Source data are provided with this paper.

CODE AVAILABILITY

HAPI analysis is available for public use with test examples as a module in Code Ocean at https://codeocean.com/capsule/2804329

AUTHOR CONTRIBUTIONS

KLM, ERW, HEJ, SDB, and XZ performed the computational analysis for the manuscript, and KLM, CD, NEP, and XZ performed the wet-laboratory experiments. KEV and XZ designed the study. JG, SDB, YL, KEV, and XZ supervised the study. KLM, KEV, and XZ wrote the manuscript with input from other co-authors.

COMPETING INTERESTS

The authors declare no competing interest regarding this work.

ACKNOWLEDGEMENTS

We thank Matthew Meyerson, John Pulice, and Dean Tantin for insightful discussions.

The support and resources from the Center for High Performance Computing and the HSC Imaging Core at the University of Utah are gratefully acknowledged.

Funding for this work was provided by R00CA215244 and R37CA263505 from the National Cancer Institute (XZ), R35GM150941 from the National Institute of General Medical Sciences (YL), 132596-RSG-18-197-01-DMC from the American Cancer Society (KEV), Huntsman Cancer Institute’s National Cancer Institute Cancer Center Support Grant (P30CA042014) and the Huntsman Cancer Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Footnotes

  • This version of the manuscript has been revised to update the corrections and modifications we have recently made in the text and figures.

REFERENCES

  1. 1.↵
    Zhang, X. & Meyerson, M. Illuminating the noncoding genome in cancer. Nat Cancer 1, 864–872 (2020). doi:10.1038/s43018-020-00114-3
    OpenUrlCrossRef
  2. 2.↵
    Li, Y. et al. Patterns of somatic structural variation in human cancer genomes. Nature 578, 112–121 (2020). doi:10.1038/s41586-019-1913-9
    OpenUrlCrossRefPubMed
  3. 3.↵
    Hadi, K. et al. Distinct Classes of Complex Structural Variation Uncovered across Thousands of Cancer Genome Graphs. Cell 183, 197–210 e132 (2020). doi:10.1016/j.cell.2020.08.006
    OpenUrlCrossRefPubMed
  4. 4.
    Garraway, L. A. & Lander, E. S. Lessons from the cancer genome. Cell 153, 17–37 (2013). doi:10.1016/j.cell.2013.03.002
    OpenUrlCrossRefPubMedWeb of Science
  5. 5.↵
    Stratton, M. R., Campbell, P. J. & Futreal, P. A. The cancer genome. Nature 458, 719–724 (2009). doi:10.1038/nature07943
    OpenUrlCrossRefPubMedWeb of Science
  6. 6.↵
    Soda, M. et al. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature 448, 561–566 (2007). doi:10.1038/nature05945
    OpenUrlCrossRefPubMedWeb of Science
  7. 7.↵
    Bartram, C. R. et al. Translocation of c-ab1 oncogene correlates with the presence of a Philadelphia chromosome in chronic myelocytic leukaemia. Nature 306, 277–280 (1983). doi:10.1038/306277a0
    OpenUrlCrossRefPubMedWeb of Science
  8. 8.↵
    Rheinbay, E. et al. Analyses of non-coding somatic drivers in 2,658 cancer whole genomes. Nature 578, 102–111 (2020). doi:10.1038/s41586-020-1965-x
    OpenUrlCrossRefPubMed
  9. 9.↵
    Groschel, S. et al. A single oncogenic enhancer rearrangement causes concomitant EVI1 and GATA2 deregulation in leukemia. Cell 157, 369–381 (2014). doi:10.1016/j.cell.2014.02.019
    OpenUrlCrossRefPubMed
  10. 10.↵
    Ryan, R. J. et al. Detection of Enhancer-Associated Rearrangements Reveals Mechanisms of Oncogene Dysregulation in B-cell Lymphoma. Cancer Discov 5, 1058–1071 (2015). doi:10.1158/2159-8290.CD-15-0370
    OpenUrlAbstract/FREE Full Text
  11. 11.
    Northcott, P. A. et al. Enhancer hijacking activates GFI1 family oncogenes in medulloblastoma. Nature 511, 428–434 (2014). doi:10.1038/nature13379
    OpenUrlCrossRefPubMed
  12. 12.↵
    Tsujimoto, Y. et al. Clustering of breakpoints on chromosome 11 in human B-cell neoplasms with the t(11;14) chromosome translocation. Nature 315, 340–343 (1985). doi:10.1038/315340a0
    OpenUrlCrossRefPubMed
  13. 13.
    Battey, J. et al. The human c-myc oncogene: structural consequences of translocation into the IgH locus in Burkitt lymphoma. Cell 34, 779–787 (1983). doi:10.1016/0092-8674(83)90534-2
    OpenUrlCrossRefPubMedWeb of Science
  14. 14.
    Dubois, F. P. B. et al. Structural variants shape driver combinations and outcomes in pediatric high-grade glioma. Nat Cancer 3, 994–1011 (2022). doi:10.1038/s43018-022-00403-z
    OpenUrlCrossRef
  15. 15.↵
    Helmsauer, K. et al. Enhancer hijacking determines extrachromosomal circular MYCN amplicon architecture in neuroblastoma. Nat Commun 11, 5823 (2020). doi:10.1038/s41467-020-19452-y
    OpenUrlCrossRef
  16. 16.
    Bandopadhayay, P. et al. MYB-QKI rearrangements in angiocentric glioma drive tumorigenicity through a tripartite mechanism. Nat Genet 48, 273–282 (2016). doi:10.1038/ng.3500
    OpenUrlCrossRefPubMed
  17. 17.↵
    Drier, Y. et al. An oncogenic MYB feedback loop drives alternate cell fates in adenoid cystic carcinoma. Nat Genet 48, 265–272 (2016). doi:10.1038/ng.3502
    OpenUrlCrossRefPubMed
  18. 18.↵
    Weischenfeldt, J. et al. Pan-cancer analysis of somatic copy-number alterations implicates IRS4 and IGF2 in enhancer hijacking. Nat Genet 49, 65–74 (2017). doi:10.1038/ng.3722
    OpenUrlCrossRefPubMed
  19. 19.↵
    He, B. et al. Diverse noncoding mutations contribute to deregulation of cis-regulatory landscape in pediatric cancers. Sci Adv 6, eaba3064 (2020). doi:10.1126/sciadv.aba3064
    OpenUrlFREE Full Text
  20. 20.↵
    Wang, X. et al. Genome-wide detection of enhancer-hijacking events from chromatin interaction data in rearranged genomes. Nat Methods 18, 661–668 (2021). doi:10.1038/s41592-021-01164-w
    OpenUrlCrossRef
  21. 21.↵
    Shoshani, O. et al. Chromothripsis drives the evolution of gene amplification in cancer. Nature 591, 137–141 (2021). doi:10.1038/s41586-020-03064-z
    OpenUrlCrossRefPubMed
  22. 22.↵
    Hung, K. L., Mischel, P. S. & Chang, H. Y. Gene regulation on extrachromosomal DNA. Nat Struct Mol Biol 29, 736–744 (2022). doi:10.1038/s41594-022-00806-7
    OpenUrlCrossRef
  23. 23.
    Wu, S., Bafna, V., Chang, H. Y. & Mischel, P. S. Extrachromosomal DNA: An Emerging Hallmark in Human Cancer. Annu Rev Pathol 17, 367–386 (2022). doi:10.1146/annurev-pathmechdis-051821-114223
    OpenUrlCrossRef
  24. 24.↵
    Yi, E., Chamorro Gonzalez, R., Henssen, A. G. & Verhaak, R. G. W. Extrachromosomal DNA amplifications in cancer. Nat Rev Genet 23, 760–771 (2022). doi:10.1038/s41576-022-00521-5
    OpenUrlCrossRef
  25. 25.↵
    Whyte, W. A. et al. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell 153, 307–319 (2013). doi:10.1016/j.cell.2013.03.035
    OpenUrlCrossRefPubMedWeb of Science
  26. 26.↵
    Hnisz, D. et al. Super-enhancers in the control of cell identity and disease. Cell 155, 934–947 (2013). doi:10.1016/j.cell.2013.09.053
    OpenUrlCrossRefPubMedWeb of Science
  27. 27.↵
    Loven, J. et al. Selective inhibition of tumor oncogenes by disruption of super-enhancers. Cell 153, 320–334 (2013). doi:10.1016/j.cell.2013.03.036
    OpenUrlCrossRefPubMedWeb of Science
  28. 28.↵
    Liu, Y. et al. A predominant enhancer co-amplified with the SOX2 oncogene is necessary and sufficient for its expression in squamous cancer. Nat Commun 12, 7139 (2021). doi:10.1038/s41467-021-27055-4
    OpenUrlCrossRef
  29. 29.
    Zhang, X. et al. Identification of focally amplified lineage-specific super-enhancers in human epithelial cancers. Nat Genet 48, 176–182 (2016). doi:10.1038/ng.3470
    OpenUrlCrossRefPubMed
  30. 30.
    Zhang, X. et al. Somatic Superenhancer Duplications and Hotspot Mutations Lead to Oncogenic Activation of the KLF5 Transcription Factor. Cancer Discov 8, 108–125 (2018). doi:10.1158/2159-8290.CD-17-0532
    OpenUrlAbstract/FREE Full Text
  31. 31.
    Takeda, D. Y. et al. A Somatically Acquired Enhancer of the Androgen Receptor Is a Noncoding Driver in Advanced Prostate Cancer. Cell 174, 422–432 e413 (2018). doi:10.1016/j.cell.2018.05.037
    OpenUrlCrossRef
  32. 32.
    Viswanathan, S. R. et al. Structural Alterations Driving Castration-Resistant Prostate Cancer Revealed by Linked-Read Genome Sequencing. Cell 174, 433–447 e419 (2018). doi:10.1016/j.cell.2018.05.036
    OpenUrlCrossRefPubMed
  33. 33.↵
    Morton, A. R. et al. Functional Enhancers Shape Extrachromosomal Oncogene Amplifications. Cell 179, 1330–1341 e1313 (2019). doi:10.1016/j.cell.2019.10.039
    OpenUrlCrossRef
  34. 34.↵
    Chakravarty, D. et al. OncoKB: A Precision Oncology Knowledge Base. JCO Precis Oncol 2017, 1–16 (2017). doi:10.1200/PO.17.00011
    OpenUrlCrossRefPubMed
  35. 35.
    Liu, Y., Sun, J. & Zhao, M. ONGene: A literature-based database for human oncogenes. J Genet Genomics 44, 119–121 (2017). doi:10.1016/j.jgg.2016.12.004
    OpenUrlCrossRefPubMed
  36. 36.↵
    Sondka, Z. et al. The COSMIC Cancer Gene Census: describing genetic dysfunction across all human cancers. Nat Rev Cancer 18, 696–705 (2018). doi:10.1038/s41568-018-0060-1
    OpenUrlCrossRefPubMed
  37. 37.↵
    Dekker, J. & Heard, E. Structural and functional diversity of Topologically Associating Domains. FEBS Lett 589, 2877–2884 (2015). doi:10.1016/j.febslet.2015.08.044
    OpenUrlCrossRefPubMed
  38. 38.↵
    Flavahan, W. A. et al. Altered chromosomal topology drives oncogenic programs in SDH-deficient GISTs. Nature 575, 229–233 (2019). doi:10.1038/s41586-019-1668-3
    OpenUrlCrossRefPubMed
  39. 39.↵
    Flavahan, W. A. et al. Insulator dysfunction and oncogene activation in IDH mutant gliomas. Nature 529, 110–114 (2016). doi:10.1038/nature16490
    OpenUrlCrossRefPubMed
  40. 40.↵
    Meyers, R. M. et al. Computational correction of copy number effect improves specificity of CRISPR– Cas9 essentiality screens in cancer cells. Nature Genetics 49, 1779–1784 (2017). doi:10.1038/ng.3984
    OpenUrlCrossRefPubMed
  41. 41.↵
    Dixon, J. R. et al. Integrative detection and analysis of structural variation in cancer genomes. Nat Genet 50, 1388–1398 (2018). doi:10.1038/s41588-018-0195-8
    OpenUrlCrossRefPubMed
  42. 42.↵
    Parolia, A. et al. Distinct structural classes of activating FOXA1 alterations in advanced prostate cancer. Nature 571, 413–418 (2019). doi:10.1038/s41586-019-1347-4
    OpenUrlCrossRefPubMed
  43. 43.↵
    Tomlins, S. A. et al. Distinct classes of chromosomal rearrangements create oncogenic ETS gene fusions in prostate cancer. Nature 448, 595–599 (2007). doi:10.1038/nature06024
    OpenUrlCrossRefPubMedWeb of Science
  44. 44.↵
    Cai, C. et al. ETV1 is a novel androgen receptor-regulated gene that mediates prostate cancer cell invasion. Mol Endocrinol 21, 1835–1846 (2007). doi:10.1210/me.2006-0480
    OpenUrlCrossRefPubMedWeb of Science
  45. 45.↵
    Ortiz, A. B. et al. Prognostic significance of cyclin D1 protein expression and gene amplification in invasive breast carcinoma. PLoS One 12, e0188068 (2017). doi:10.1371/journal.pone.0188068
    OpenUrlCrossRef
  46. 46.↵
    Wu, S. et al. Circular ecDNA promotes accessible chromatin and high oncogene expression. Nature 575, 699–703 (2019). doi:10.1038/s41586-019-1763-5
    OpenUrlCrossRefPubMed
  47. 47.↵
    Hung, K. L. et al. ecDNA hubs drive cooperative intermolecular oncogene expression. Nature 600, 731–736 (2021). doi:10.1038/s41586-021-04116-8
    OpenUrlCrossRef
  48. 48.↵
    Salari, K. et al. CDX2 is an amplified lineage-survival oncogene in colorectal cancer. Proc Natl Acad Sci U S A 109, E3196–3205 (2012). doi:10.1073/pnas.1206004109
    OpenUrlAbstract/FREE Full Text
  49. 49.↵
    Werling, R. W., Yaziji, H., Bacchi, C. E. & Gown, A. M. CDX2, a highly sensitive and specific marker of adenocarcinomas of intestinal origin: an immunohistochemical survey of 476 primary and metastatic carcinomas. Am J Surg Pathol 27, 303–310 (2003). doi:10.1097/00000478-200303000-00003
    OpenUrlCrossRefPubMedWeb of Science
  50. 50.↵
    Hung, K. L. et al. Targeted profiling of human extrachromosomal DNA by CRISPR-CATCH. Nat Genet 54, 1746–1754 (2022). doi:10.1038/s41588-022-01190-0
    OpenUrlCrossRef
  51. 51.↵
    Hong, S. H. et al. APIP, an ERBB3-binding partner, stimulates erbB2-3 heterodimer formation to promote tumorigenesis. Oncotarget 7, 21601–21617 (2016). doi:10.18632/oncotarget.7802
    OpenUrlCrossRef
  52. 52.↵
    Hou, W., Kong, L., Hou, Z. & Ji, H. CD44 is a prognostic biomarker and correlated with immune infiltrates in gastric cancer. BMC Med Genomics 15, 225 (2022). doi:10.1186/s12920-022-01383-w
    OpenUrlCrossRef
  53. 53.↵
    Deshpande, V. et al. Exploring the landscape of focal amplifications in cancer using AmpliconArchitect. Nat Commun 10, 392 (2019). doi:10.1038/s41467-018-08200-y
    OpenUrlCrossRef
  54. 54.↵
    Estrada-Bernal, A. et al. Tarloxotinib Is a Hypoxia-Activated Pan-HER Kinase Inhibitor Active Against a Broad Range of HER-Family Oncogenes. Clin Cancer Res 27, 1463–1475 (2021). doi:10.1158/1078-0432.CCR-20-3555
    OpenUrlAbstract/FREE Full Text
  55. 55.↵
    Ise, N., Omi, K., Nambara, D., Higashiyama, S. & Goishi, K. Overexpressed HER2 in NSCLC is a Possible Therapeutic Target of EGFR Inhibitors. Anticancer Res 31, 4155–4161 (2011).
    OpenUrlAbstract/FREE Full Text
  56. 56.↵
    Pongor, L. S. et al. Extrachromosomal DNA Amplification Contributes to Small Cell Lung Cancer Heterogeneity and Is Associated with Worse Outcomes. Cancer Discov 13, 928–949 (2023). doi:10.1158/2159-8290.CD-22-0796
    OpenUrlCrossRef
  57. 57.↵
    Aaltonen, L. A. et al. Pan-cancer analysis of whole genomes. Nature 578, 82–93 (2020). doi:10.1038/s41586-020-1969-6
    OpenUrlCrossRefPubMed
  58. 58.↵
    Xu, Z. et al. Structural variants drive context-dependent oncogene activation in cancer. Nature 612, 564–572 (2022). doi:10.1038/s41586-022-05504-4
    OpenUrlCrossRef
  59. 59.↵
    Hsieh, T. H. et al. Mapping Nucleosome Resolution Chromosome Folding in Yeast by Micro-C. Cell 162, 108–119 (2015). doi:10.1016/j.cell.2015.05.048
    OpenUrlCrossRefPubMed
  60. 60.
    Fang, R. et al. Mapping of long-range chromatin interactions by proximity ligation-assisted ChIP-seq. Cell Res 26, 1345–1348 (2016). doi:10.1038/cr.2016.137
    OpenUrlCrossRefPubMed
  61. 61.↵
    Fullwood, M. J. et al. An oestrogen-receptor-alpha-bound human chromatin interactome. Nature 462, 58–64 (2009). doi:10.1038/nature08497
    OpenUrlCrossRefPubMedWeb of Science
  62. 62.↵
    Song, M. et al. Cell-type-specific 3D epigenomes in the developing human cortex. Nature 587, 644–649 (2020). doi:10.1038/s41586-020-2825-4
    OpenUrlCrossRef
  63. 63.↵
    Li, Y., Zhou, S., Schwartz, D. C. & Ma, J. Allele-Specific Quantification of Structural Variations in Cancer Genomes. Cell Syst 3, 21–34 (2016). doi:10.1016/j.cels.2016.05.007
    OpenUrlCrossRef
  64. 64.↵
    Luebeck, J. et al. Extrachromosomal DNA in the cancerous transformation of Barrett’s esophagus. Nature 616, 798–805 (2023). doi:10.1038/s41586-023-05937-5
    OpenUrlCrossRefPubMed
  65. 65.↵
    Lee, J. J. et al. ERα-associated translocations underlie oncogene amplifications in breast cancer. Nature 618, 1024–1032 (2023). doi:10.1038/s41586-023-06057-w
    OpenUrlCrossRef
  66. 66.↵
    Ghandi, M. et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature 569, 503–508 (2019). doi:10.1038/s41586-019-1186-3
    OpenUrlCrossRefPubMed
  67. 67.↵
    Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 483, 603–607 (2012). doi:10.1038/nature11003
    OpenUrlCrossRefPubMedWeb of Science
  68. 68.↵
    Mumbach, M. R. et al. HiChIP: efficient and sensitive analysis of protein-directed genome architecture. Nat Methods 13, 919–922 (2016). doi:10.1038/nmeth.3999
    OpenUrlCrossRefPubMed
  69. 69.↵
    Liu, Y. et al. Chromatin Looping Shapes KLF5-Dependent Transcriptional Programs in Human Epithelial Cancers. Cancer Res 80, 5464–5477 (2020). doi:10.1158/0008-5472.CAN-20-1287
    OpenUrlAbstract/FREE Full Text
  70. 70.↵
    Servant, N. et al. HiC-Pro: an optimized and flexible pipeline for Hi-C data processing. Genome Biol 16, 259 (2015). doi:10.1186/s13059-015-0831-x
    OpenUrlCrossRefPubMed
  71. 71.↵
    Lareau, C. A. & Aryee, M. J. hichipper: a preprocessing pipeline for calling DNA loops from HiChIP data. Nat Methods 15, 155–156 (2018). doi:10.1038/nmeth.4583
    OpenUrlCrossRefPubMed
  72. 72.↵
    Amemiya, H. M., Kundaje, A. & Boyle, A. P. The ENCODE Blacklist: Identification of Problematic Regions of the Genome. Sci Rep 9, 9354 (2019). doi:10.1038/s41598-019-45839-z
    OpenUrlCrossRefPubMed
  73. 73.↵
    Zhang, H., Meltzer, P. & Davis, S. RCircos: an R package for Circos 2D track plots. BMC Bioinformatics 14, 244 (2013). doi:10.1186/1471-2105-14-244
    OpenUrlCrossRefPubMed
  74. 74.↵
    Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010). doi:10.1093/bioinformatics/btq033
    OpenUrlCrossRefPubMedWeb of Science
  75. 75.↵
    Oudelaar, A. M. et al. A revised model for promoter competition based on multi-way chromatin interactions at the alpha-globin locus. Nat Commun 10, 5412 (2019). doi:10.1038/s41467-019-13404-x
    OpenUrlCrossRef
  76. 76.
    Cho, S. W. et al. Promoter of lncRNA Gene PVT1 Is a Tumor-Suppressor DNA Boundary Element. Cell 173, 1398-1412 e1322 (2018). doi:10.1016/j.cell.2018.03.068
    OpenUrlCrossRefPubMed
  77. 77.↵
    Wijgerde, M., Grosveld, F. & Fraser, P. Transcription complex stability and chromatin dynamics in vivo. Nature 377, 209-213 (1995). doi:10.1038/377209a0
    OpenUrlCrossRefPubMedWeb of Science
  78. 78.↵
    Liberzon, A. et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics 27, 1739–1740 (2011). doi:10.1093/bioinformatics/btr260
    OpenUrlCrossRefPubMedWeb of Science
  79. 79.↵
    Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102, 15545–15550 (2005). doi:10.1073/pnas.0506580102
    OpenUrlAbstract/FREE Full Text
  80. 80.↵
    Kanehisa, M., Furumichi, M., Sato, Y., Kawashima, M. & Ishiguro-Watanabe, M. KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res 51, D587–D592 (2023). doi:10.1093/nar/gkac963
    OpenUrlCrossRefPubMed
  81. 81.↵
    Kanehisa, M. & Goto, S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res 28, 27–30 (2000). doi:10.1093/nar/28.1.27
    OpenUrlCrossRefPubMedWeb of Science
  82. 82.↵
    Jassal, B. et al. The reactome pathway knowledgebase. Nucleic Acids Res 48, D498–d503 (2020). doi:10.1093/nar/gkz1031
    OpenUrlCrossRefPubMed
  83. 83.↵
    Ashburner, M. et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25, 25–29 (2000). doi:10.1038/75556
    OpenUrlCrossRefPubMedWeb of Science
  84. 84.↵
    Gene Ontology, C. The Gene Ontology resource: enriching a GOld mine. Nucleic Acids Res 49, D325–D334 (2021). doi:10.1093/nar/gkaa1113
    OpenUrlCrossRefPubMed
  85. 85.↵
    Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat Methods 9, 357–359 (2012). doi:10.1038/nmeth.1923
    OpenUrlCrossRefPubMedWeb of Science
  86. 86.↵
    Li, B. & Dewey, C. N. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12, 323 (2011). doi:10.1186/1471-2105-12-323
    OpenUrlCrossRefPubMed
  87. 87.↵
    Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2009). doi:10.1093/bioinformatics/btp616
    OpenUrlCrossRefPubMedWeb of Science
  88. 88.↵
    Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009). doi:10.1093/bioinformatics/btp352
    OpenUrlCrossRefPubMedWeb of Science
  89. 89.↵
    Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137 (2008). doi:10.1186/gb-2008-9-9-r137
    OpenUrlCrossRefPubMed
  90. 90.↵
    Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 38, 576–589 (2010). doi:10.1016/j.molcel.2010.05.004
    OpenUrlCrossRefPubMedWeb of Science
  91. 91.↵
    Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26, 589–595 (2010). doi:10.1093/bioinformatics/btp698
    OpenUrlCrossRefPubMedWeb of Science
  92. 92.↵
    Cameron, D. L. et al. GRIDSS2: comprehensive characterisation of somatic structural variation using single breakend variants and structural variant phasing. Genome Biol 22, 202 (2021). doi:10.1186/s13059-021-02423-x
    OpenUrlCrossRef
  93. 93.↵
    Cameron, D. L. et al. GRIDSS: sensitive and specific genomic rearrangement detection using positional de Bruijn graph assembly. Genome Res 27, 2050–2060 (2017). doi:10.1101/gr.222109.117
    OpenUrlAbstract/FREE Full Text
  94. 94.↵
    StructuralVariantAnnotation: Variant annotations for structural variants (2019).
  95. 95.↵
    Talevich, E., Shain, A. H., Botton, T. & Bastian, B. C. CNVkit: Genome-Wide Copy Number Detection and Visualization from Targeted DNA Sequencing. PLoS Comput Biol 12, e1004873 (2016). doi:10.1371/journal.pcbi.1004873
    OpenUrlCrossRefPubMed
  96. 96.↵
    Corces, M. R. et al. The chromatin accessibility landscape of primary human cancers. Science 362, eaav1898 (2018). doi:10.1126/science.aav1898
    OpenUrlAbstract/FREE Full Text
Back to top
Posted June 21, 2024.
Download PDF

Supplementary Material

Email
3D genomic analysis reveals novel enhancer-hijacking caused by complex structural alterations that drive oncogene overexpression
Katelyn L. Mortenson, Courtney Dawes, Emily R. Wilson, Nathan E. Patchen, Hailey E. Johnson, Jason Gertz, Swneke D. Bailey, Yang Liu, Katherine E. Varley, Xiaoyang Zhang
bioRxiv 2024.01.23.576965; doi: https://doi.org/10.1101/2024.01.23.576965
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area