Targeting the Noncoding Genome: Superenhancers Meet Their Kryptonite

Summary

In this study, McKeown and colleagues carried out a genome-wide characterization and stratification of the enhancer landscape in acute myeloid leukemia (AML). The authors’ analysis led to the discovery of a novel RARA superenhancer found in a subset of patients with AML, rendering these leukemia cells highly sensitive to SY-1425, a highly potent RARA agonist able to induce myeloid differentiation in these high-expressing RARA AML subtypes.

Noncoding regulatory DNA regions known as enhancers are characterized by transcription factor (TF)–binding sites and are capable of regulating genes over large genomic distances through promoter–enhancer interactions. These enhancer sites are defined by TF occupancy, which facilitates the recruitment of the Mediator complex and chromatin-modifying enzymes, leading to deposition of histone H3 lysine 27 acetylation (H3K27ac) and H3 lysine 4 monomethylation (H3K4me1). In 2013, Richard Young’s group classified DNA-regulatory regions as “superenhancers,” which are defined as long stretches of enhancers with elevated levels of enhancer-associated histone marks and Mediator binding. Superenhancers have been shown to be critical in embryonic stem cells and other cell types by regulating master TFs necessary for cell fate determination (1, 2). Because of the importance of superenhancers in transcriptional processes, several studies have also implicated their involvement in disease mechanisms, including the regulation of key oncogenic drivers in cancers. This led to the first therapeutic approach of using the prototypical BET bromodomain inhibitor JQ1 to selectively target the bromodomain protein BRD4, a major constituent of the MYC superenhancer that is found in several hematologic malignancies (3–8). Altogether, these were some of the studies that highlighted the oncogenic dependency of superenhancers in cancer and opened a new avenue for therapeutic targeting of TFs (9).

In this issue of Cancer Discovery, McKeown and colleagues continue this quest to target superenhancers, this time in acute myeloid leukemia (AML; ref. 10). Initially, the authors set out to define the superenhancer landscape and the transcriptional networks in patients with AML and AML cell lines. To this end, the authors performed high-throughput chromatin immunoprecipitation sequencing (ChIP-seq) for the H3K27ac mark, found on active enhancers, in 66 AML patient donors and 28 AML cell lines. Using de novo analysis to profile the enhancer environment, the authors classified 6 novel superenhancer clusters in AML that were distinct from normal hematopoietic progenitor cells (HSPC) or normal myeloid (monocyte) populations. Genetic mutation analysis and cytogenetic studies of these patients with AML determined one subclass of superenhancers to be heavily associated with mixed lineage leukemia translocation leukemia; additionally, NPM1 and FLT3 mutations were found to be associated with several superenhancer signatures. However, in a majority of the cases, the mutational background did not correlate with a single superenhancer cluster. Subsequent analysis of each superenhancer cluster also demonstrated differential overall median survival of patients with AML.

TF binding and activity are commonly deregulated in leukemia and are major classes of cancer dependencies; therefore, the authors decided to focus their efforts on generating a network analysis of superenhancers that sustain the expression of key transcription factors. This led them to identify, from one of their TF clusters, a superenhancer found within the RARA gene locus leading to high expression of RARA. Ironically, one of the successful examples of targeting TFs in leukemia is the use of the FDA-approved drug all-trans retinoic acid (ATRA) for treating patients with acute promyelocytic leukemia (APL), a subtype of AML characterized by a t(15;17) chromosomal translocation. However, McKeown and colleagues found the RARA superenhancer was highly expressed in patients with AML or myelodysplastic syndrome (MDS) who lack the t(15;17) translocation, which shows these leukemia cells are highly addicted to elevated levels of RARA. Furthermore, the RARA superenhancer was observed in only a subset of their patients with AML and was absent from normal HSPCs.

This led the authors to suggest using a selective RARA agonist, SY-1425 (also known as tamibarotene). To test this approach, the authors selected a panel of AML cell lines characterized with strong or weak RARA superenhancer activity. As predicted, they observed that RARA-high cell lines treated with the SY-1425 agonist exhibited profound antileukemic effects compared with RARA-low cell lines. The authors further validated their findings using in vivo AML models by transplanting non-APL AML cell lines and patient-derived mouse xenograft models (PDX) that exhibited RARA-high or RARA-low phenotypes. Using PDX mouse models, McKeown and colleagues were able to show that oral dosing of SY-1425 in RARA-high AML led to depletion of AML blast counts, ultimately leading to disease regression and survival benefit. Additional analysis of AML cell lines and ex vivo AML patient samples showed an increase of CD38 mRNA expression in RARA-high treated with SY-1425, implicating that blocking RARA binding consequently leads to myeloid maturation.

The authors further investigated the transcriptional and epigenetic impact of SY-1425 in RARA-high AML. Differential expression profiling of SY-1425 treatment demonstrated that RARA-high AML cell lines exhibited greater dynamic changes in gene expression compared with RARA-low AML, as would be expected due to their differential sensitivity to the RARA agonist. Gene set enrichment analysis of genes upregulated in RARA-high AML treated with SY-1425 showed processes implicative of myeloid differentiation, whereas downregulated genes were associated with MYC target genes. Interestingly, H3K27ac ChIP-seq analysis in RARA-high AML treated with SY-1425 demonstrated an increase in H3K27ac accumulation near strong RARA-occupied sites, suggesting a nucleation model in which RARA occupancy spreads and upregulates nearby genes and thereby promoting myeloid differentiation. This model proposed by McKeown and colleagues suggests SY-1425 acts as an agonist and converts RARA repressor into a transcriptional activator, a mechanism that is also conserved in patients with APL treated with ATRA.

Overall, this study provides a proof of principle of using superenhancers (and active enhancers in general) as clinical biomarkers to predict sensitivities of targeting specific oncogenic transcription factors driving superenhancer activity. McKeown and colleagues have demonstrated the efficacy of utilizing SY-1425 in RARA superenhancer AML and have shown the therapeutic advantage of SY-1425 over the conventional use of ATRA in PDX AML models. SY-1425 is already approved in Japan for treating patients with relapsed APL and is currently undergoing clinical phase II trials under Syros Pharmaceuticals as a single drug and also as combination therapy with azacitidine for the treatment of patients with relapsed AML and high-risk patients with MDS. However, such studies introduce a number of exciting questions, including what is an enhancer, how its activity is controlled, and whether histone marks are enough to detect active enhancers or such approaches have to be coupled to RNA sequencing (to identify eRNA transcription) and 3-D chromosomal architecture approaches (to identify enhancer-interacting partners).