Chromatin Variants Reveal the Genetic Determinants of Oncogenesis in Breast Cancer

Abstract

Breast cancer presents as multiple distinct disease entities. Each tumor harbors diverse cell populations defining a phenotypic heterogeneity that impinges on our ability to treat patients. To date, efforts mainly focused on genetic variants to find drivers of inter- and intratumor phenotypic heterogeneity. However, these efforts have failed to fully capture the genetic basis of breast cancer. Through recent technological and analytical approaches, the genetic basis of phenotypes can now be decoded by characterizing chromatin variants. These variants correspond to polymorphisms in chromatin states at DNA sequences that serve a distinct role across cell populations. Here, we review the function and causes of chromatin variants as they relate to breast cancer inter- and intratumor heterogeneity and how they can guide the development of treatment alternatives to fulfill the goal of precision cancer medicine.

Figure 1.

Differences in chromatin states differentially delineate breast cancer subtype-specific gene regulatory landscapes. Cis-regulatory elements such as promoters and enhancers can harbor different combinations of chromatin modifications, known as chromatin states. These can denote active enhancers (e.g., H3K27ac, H3K4me1), active promoters (e.g., H3K27ac, H3K4me3), and inactive enhancers or promoters (e.g., H3K27me3, H3K9me3). Additionally, active cis-regulatory elements typically reside in accessible chromatin, whereas inactive ones are hidden in closed chromatin. Differences in chromatin state over enhancers and promoters of ESR1, HER2 (ERBB2), and PGR (top) can be associated with differences in gene expression of the luminal A, luminal B, HER2, and basal/triple-negative breast cancer (TNBC) (n = 15) subtypes. High ESR1 and PGR gene expression were typical of luminal A (n = 18) and B (n = 25) and high ERBB2 for HER2 (n = 11) (bottom). Breast cancer samples were obtained from The Cancer Genome Atlas (TCGA) cohort with matched PAM50 subtyping and processed RSEM normalized gene expression levels were downloaded from cBioPortal (Cerami et al. 2012; Corces et al. 2018). Figure created with BioRender.com.

Figure 2.

Chromatin variants capture cell-state population shifts across oncogenesis. Oncogenesis involves normal cells acquiring tumor-initiating properties from accumulating cancer-initiating genetic and chromatin variants. Finding regions of the genome differing in chromatin states across cancer cells in different states, such as primary versus drug-resistant or metastatic state, identifies cell-state-associated chromatin variants. Cancer cell-state-associated chromatin variants commonly define cis-regulatory elements permissive to transcription factor (TF)-binding guiding cell-state-specific transcriptional programs. Figure created with BioRender.com.

Figure 3.

Regulatory plexuses and cistromes define noncoding cancer drivers. (A) Cancer driver regulatory plexuses are defined based on the excess of mutations (blue dots) across the sum of all cis-regulatory elements of a gene, inclusive of all enhancers (yellow boxes) and the promoter (red box) regulating the expression from one gene (green box). The excess of mutations relies on comparing the frequency of mutations across regulatory plexuses in a cell state. (B) Cancer driver cistromes are defined based on the excess of mutations across the sum of all cis-regulatory elements bound by the same transcription factor, such as FOXA1, in a cell state. Not all cistromes are excessively burdened by mutations (e.g., TFAP4 cistrome). Figure created with BioRender.com.

Figure 4.

Chromatin variants can arise from disruption to metabolism and pioneer factor activity. (A) S-Adenosylmethionine (SAM) is the universal methyl donor. It is used by DNA methyltransferase (DNMT) enzymes to methylate DNA, which defines a repressive chromatin state over repetitive DNA sequences, such as transposable elements (TEs). The chemotherapeutic agent paclitaxel, used for the treatment of TNBCs, alters the methionine pathway responsible for SAM production. This leads to decreased SAM levels limiting DNMT activity and DNA hypomethylation giving rise to chromatin variants. Paclitaxel-resistant cancer cells emerge by compensating DNA hypomethylation over TEs with an alternative chromatin state relying on EZH2-dependent H3K27me3, thereby establishing chemotherapy-associated chromatin variants. (B) Pioneer factors, such as FOXA1, can access their binding sites (forkhead motif [FKH] for FOXA1) within closed chromatin and initiate nucleosome displacement to create chromatin variants by rendering the chromatin accessible to other transcription factors. In luminal breast cancer subtypes, this pioneering activity of FOXA1 exposes DNA-recognition motifs for the estrogen receptor α (ERα), known as EREs. Under estrogen stimulation (red circle), ERα can then bind those sequences and impact target gene expression. Figure created with BioRender.com.

Figure 5.

Perturbing chromatin variants can provide a means for delivering precision medicine. (A) Chromatin modifier enzymes deposit modifications on histones or DNA. DNA methyltransferases (DNMTs) methylate DNA. Histone lysine methyltransferases (KMTs) transfer one, two, or three methyl groups to lysine or arginine residues on histones. Histone lysine deacetylases (KDACs) remove acetyl groups from lysine residues of histones, whereas histone lysine acetylases (KATs) add acetyl groups. Chromatin modifiers can be targets of chemical inhibitors to alter chromatin states, defining epigenetic therapy. (B) Chromatin variants can be directly manipulated by using genome hacking approaches, such as CRISPRi, CRISPRa, CRISPRoff, or CRISPRon. Catalytically dead Cas9 nuclease (dCas9) can be used to modify a target site such as a putative chromatin variant on the genome specified by the sgRNA. CRISPRi involves the fusion of dCas9 to a repressor domain such as MeCP2 and KRAB. This changes the active chromatin state to an inactive chromatin state at target regions, which can serve to repress gene expression. CRISPRa consists of dCas9 fused to a transcriptional activator that can activate cis-regulatory elements of choice to increase target gene expression. CRISPRoff and CRISPRon introduce heritable changes to the chromatin state at target regions. The former consists of a dCas9-KRAB-DNMT construct that methylates DNA while the latter uses dCas9-TET to remove methyl groups on DNA. (CRISPRi) CRISPR interference, (KRAB) Krüppel-associated box, (sgRNA) short-guide RNA, (MeCP2) methyl-CpG-binding protein 2. Figure created with BioRender.com.