Can Identifying the Cell-of-Origin and Epigenetic Targets Halt Prostate Cancer in its Tracks?

By Stuart P. Atkinson

Basal prostate cells expand into a fast-growing intermediate cell type that coexpresses basal and luminal markers and appears in invasive tumors; immunofluorescence shows these intermediate (IM) cells — and the tumor luminal cells — express the ERG oncogene and androgen receptor, consistent with ERG-driven tumor formation.

Just Where Does ERG-expressing Prostate Cancer Originate?

Cell-tracing and organoid experiments show that, when ERG is active, rare basal and IM cells shift toward a luminal cell identity over time — dividing both symmetrically and asymmetrically — producing daughter cells with luminal features and linking ERG activity to a luminal tumor outcome.

Multiple studies have suggested that the ERG(ETS-related gene) translocations present in a considerable number of prostate cancer cases in Western cohorts represent an initiating disease event, with uniform ERG expression noted in the lesions that precede prostate cancer development. But just how does the ERG transcription factor function in these cases, and do tumors originate from basal or luminal prostate cells?

Unfortunately, transcriptomic and epigenetic studies have failed to reveal ERG-induced oncogenic events, although bulk RNA-seq-based studies have revealed the possibly pro-oncogenic activation of a pro-luminal epithelial differentiation program associated with the loss of basal epithelial cells. Notably, the bulk nature of chromatin-binding analysis in related studies may have limited the examination of ERG cooperativity with the androgen receptor regarding target gene activation in prostate cancer, potentially overlooking specific events in rare cell populations.

Researchers led by Charles L. Sawyers (Memorial Sloan Kettering Cancer Center) recently sought to explore how ERG translocations induce prostate cancer tumorigenicity by employing lineage tracing in mouse models. Now, their new Nature Genetics article identifies basal prostate cells expressing luminal marker genes as the cell-of-origin for ERG-driven prostate cancer and defines an important role for the histone H3K4 methyltransferase Kmt2a and the histone H3K79 methyltransferase Dot1l (Feng, Ladewig, Lange, Salsabeel, and Zhao et al.). Could this exciting new epigenetics study provide a new means of halting prostate cancer in its tracks and improving outcomes for affected patients?

Paired-Tag technology from Epigenome Technologies enables the simultaneous profiling of transcriptomics and epigenetics in single cells. Could the integration of this approach in this fascinating new study have yielded additional insights and provided more depth to this study of the origins of prostate cancer? Paired-Tag technology from Epigenome Technologies generates joint epigenetic and transcriptomic profiles at single-cell resolution and detects histone modifications and RNA transcripts in individual nuclei with comparable efficiency to single-nucleus RNA-seq/ChIP-seq assays while avoiding the need for cell sorting. Could Paired-Tag support the further exploration of the roles of Kmt2a/Dot1l and histone methylation in prostate cancer initiation, progression, and response to potential therapeutics, thereby defining potentially targetable pathways that could halt this critically important form of cancer in its tracks?

Single-cell chromatin profiling reveals ERG-expressing IM cells carry a distinct open-chromatin landscape enriched for STAT3 and other transcription-factor motifs (NF-κB, AP-1, NFAT), highlighting a chromatin context that may cooperate with ERG to drive tumorigenesis.

Can We Target Epigenetic Modulators in the Cell-of-Origin to Put a Quick End to Prostate Cancer?

The authors first employed mouse organoid transplantation and lineage tracing to investigate whether ERG-mediated prostate tumorigenesis arises from basal or luminal prostate cells via cell-type-specific ERG activation. They discovered that tumorigenesis did not arise in the much larger population of ERG-expressing luminal cells of the prostate; instead, their approach implicated a rare subpopulation of basal cells in the healthy mouse prostate that coexpressed a subset of luminal marker genes (such as Tmprss2 and Nkx3.1).

In short, ERG activation in a genetically engineered mouse model provided said “BasalLum” cells with the ability to produce a highly proliferative intermediate cell type with stem cell-like properties that coexpressed basal, luminal, hillock, and club marker genes; this population then gave rise to an invasive luminal adenocarcinoma that accounts for the bulk of cells in mouse invasive prostate tumors.

Single-cell transcriptomic (single-cell RNA sequencing) and chromatin accessibility (single-cell assay for transposase-accessible chromatin with high-throughput sequencing) profiling of disease initiation and progression in the genetically engineered mouse model revealed a unique chromatin landscape of ERG-expressing intermediate cells. In brief, they displayed increased chromatin accessibility at binding sites for the Stat3 transcription factor (as well as for ERG, NF-κB, AP-1, and NFATC1) and the elevated expression of the Kmt2a and Dot1l methyltransferases. An ERG-dependent in vivo tumorigenicity assay that efficiently scores the consequences of perturbing candidates via CRISPR deletion revealed a requirement for Stat3, Kmt2a, and Dot1l for ERG-dependent prostate tumorigenicity, suggesting the potential for pharmacologic interventions with currently available clinical-grade inhibitors (Tsimberidou et al. and Feng et al.).

Finally, and perhaps most importantly, the authors reported the existence of the gene expression signatures of ERG-expressing BasalLum cells and intermediate cells identified in mouse models in ERG-expressing human prostate cancers; furthermore, the presence of the intermediate cell gene expression signature correlated with shorter disease-free survival, suggesting that the cell-of-origin for the ERG translocation significantly impacts patient outcomes. Meanwhile, a limited analysis of human prostate cell lines suggested that ERG-expressing prostate cancer cell lines may also display a dependency on KMT2A and DOT1L, which makes their further interrogation an important next step. Importantly, KMT2A has been previously implicated as a dependency in castration-resistant prostate cancer by co-activating androgen receptor signaling (Malik et al.).

Schematic model: ERG translocations in a rare basal population with luminal features (BasalLum) push cells into a highly proliferative, multi-lineage intermediate state that progresses to invasive, luminal-type cancer; by contrast, ERG in mature luminal cells tends not to produce invasive disease. The model emphasizes cooperation between ERG, other TFs and epigenetic enzymes (e.g., MLL1/KMT2A, DOT1L).

Stopping Prostate Cancer in Its Tracks: Take-Home Messages

Overall, this exciting study provides some important take-home messages: i) single-cell approaches combined with lineage tracing can identify cancer vulnerabilities not evident from bulk analysis; ii) cancers can initiate in a cell with a different lineage identity from the resultant tumor; and iii) the identification of highly proliferative, stem-like population with a unique chromatin state may provide an opportunity to reveal exploitable ERG-specific dependencies.

The additional integration of simultaneous profiling of transcriptomics and epigenetics in single cells, afforded by applying Paired-Tag technology from Epigenome Technologies, could provide a more detailed description of the progression of ERG