Evolutionarily conserved enhancer-associated features within the MYEOV locus suggest a regulatory role for this non-coding DNA region in cancer

Abstract

The myeloma overexpressed gene (MYEOV) has been proposed to be a proto-oncogene due to high RNA transcript levels found in multiple cancers, including myeloma, breast, lung, pancreas and esophageal cancer. The presence of an open reading frame (ORF) in humans and other primates suggests protein-coding potential. Yet, we still lack evidence of a functional MYEOV protein. It remains undetermined how MYEOV overexpression affects cancerous tissues. In this work, we show that MYEOV has likely originated and may still function as an enhancer, regulating CCND1 and LTO1. Firstly, MYEOV 3' enhancer activity was confirmed in humans using publicly available ATAC-STARR-seq data, performed on B-cell-derived GM12878 cells. We detected enhancer histone marks H3K4me1 and H3K27ac overlapping MYEOV in multiple healthy human tissues, which include B cells, liver and lung tissue. The analysis of 3D genome datasets revealed chromatin interactions between a MYEOV-3'-putative enhancer and the proto-oncogene CCND1. BLAST searches and multi-sequence alignment results showed that DNA sequence from this human enhancer element is conserved from the amphibians/amniotes divergence, with a 273 bp conserved region also found in all mammals, and even in chickens, where it is consistently located near the corresponding CCND1 orthologues. Furthermore, we observed conservation of an active enhancer state in the MYEOV orthologues of four non-human primates, dogs, rats, and mice. When studying this homologous region in mice, where the ORF of MYEOV is absent, we not only observed an enhancer chromatin state but also found interactions between the mouse enhancer homolog and Ccnd1 using 3D-genome interaction data. This is similar to the interaction observed in humans and, interestingly, coincides with CTCF binding sites in both species. Taken together, this suggests that MYEOV is a primate-specific gene with a de novo ORF that originated at an evolutionarily older enhancer region. This deeply conserved putative enhancer element could regulate CCND1 in both humans and mice, opening the possibility of studying MYEOV regulatory functions in cancer using non-primate animal models.

Keywords: CCND1 (Cyclin D1); MYEOV; enhancer; evolution; oncogene.

FIGURE 1

Location of the MYEOV-3′-putative enhancer. ChIP-seq, RNAseq and DNase-Seq data from lymphoblastoid cell line GM12878 were accessed from the ENCODE dataset alongside active regions and driver elements identified using the SHARPR-RE method (Wang X. et al., 2018), which predicts sub-regions with high regulatory activity at nucleotide resolution. MYEOV-3′-putative enhancer, Mikulasova (Mikulasova et al., 2022) and Garcia-Perez (García-Pérez et al., 2021) regulatory elements are all shown in orange blocks. ChIP-seq and RNAseq data from B cell, lung, liver and pancreatic tissue together with DNase-seq data in B cells, taken from the ENCODE dataset and visualised using the WashU Epigenome Browser.

FIGURE 2

Interacting partners of the MYEOV-3′-putative enhancer. (A) Top panels: genome tracks showing ChIP-seq data for H3K4me1 (orange), H3K4me3 (green) and H3K27ac (blue) from primary B cells, and CTCF ChIP-seq data from GM12878 cells. Bottom panels: PCHi-C data from naive B cells, where the purple arcs denote interactions between fragments inside the TAD region containing CCND1 and MYEOV. CTCF ChIA-PET data and RAD-21 ChIA-PET data in GM12878 cells with replicates combined. The purple arcs denote interactions between fragments. Data is visualised using the WashU Epigenome Browser. (B) Detail of interactions between MYEOV, CCND1 and LTO1 (for the region indicated with a red rectangle in panel A). Interactions where CCND1 promoter was taken as the bait are highlighted in purple. Interactions where the LTO1 promoter was taken as bait are shown in blue. Interactions where MYEOV promoter was taken as bait are highlighted in green. Arrows indicate the baits. Data is visualised using the WashU Epigenome Browser.

FIGURE 3

MYEOV 3D chromatin interactions evidenced by Hi-C data and HiP-HoP polymer simulations. Heatmaps showing real and simulated Hi-C data from GM12878 cells which reveal interactions between a broad region around MYEOV and a smaller region enclosing just CCND1 and LTO1, as indicated. (A) Real Hi-C data are binned at 10 kbp resolution (Rao et al., 2014). Colour scale units are log-normalised interaction counts; the darker the colour, the more frequently two genomic loci were found proximal within a population of cells. An unmappable repetitive sequence region is shown in white. (B) Interactions at higher resolutions can be predicted using the HiP-HoP polymer simulation scheme (Rico et al., 2022). Interactions for the same region as above are shown binned at 3 kbp resolution; colour scale units are log simulated interaction counts. Here, the darker the colour, the more frequently two genomic loci were found proximal within a population of simulated chromatin configurations. (C) The simulated Hi-C is plotted again, but now the number of interaction counts is normalised to account for the effect of genomic separation, giving an “interaction enrichment” plot (darker colours indicate that interactions are enriched compared to what would be expected for a random pair of loci at the same genomic separation; see main text for details). Data are binned at 3 kbp resolution, and colour scale units are log₂(observed simulated reads/expected simulated reads).

FIGURE 4

Homologous DNA regions in non-human species show enhancer-associated features. Top five tracks: Regulatory regions defined by ChromHMM in lymphoblastoid cells in human, chimpanzee, gorilla, orangutan, and macaque from García-Pérez et al. (2021). Abbreviation sE–Strong Enhancer, P/E−promoter/enhancer (where one replicate had an enhancer state and the other had a promoter state). Tracks below: ChIP-seq data for H3K4me1 (orange), H3K4me3 (green), H3K27ac (blue) derived from macaques, mice, rats, dogs and chicken (Kern et al., 2021; Roller et al., 2021). Individual sample replicates are combined into one track. The conserved region (the region with homologous sequence) is highlighted in the red box. MYEOV-3′-putative enhancer shaded in blue. Data is visualised on IGV.

FIGURE 5

Multiple sequence alignment highlights conservation of MYEOV-3′-putative enhancer across multiple species. (A) Multiple sequence alignment illustrates the comparative analysis of the human MYEOV-3′-putative enhancer sequence with homologous sequences from 17 distinct species, identified through a BLAST search. The homologous sequences include 16 mammals, including four primates, and chicken. Sequence conservation is visualised using a colour-coded identity matrix, where darker colours signify higher sequence similarity. The consensus identity score is depicted, with yellow-green colours indicating conservation ranging from 50% to 100%. (B) Extraction and realignment of conserved regions, representing the largest contiguous block of identity scores between 50% and 100% present in all species. The realigned conserved region spans 635 bp for mammals and 273 bp for chicken. Identical residues are shaded in grey for clarity.

FIGURE 6

Interactions observed between conserved enhancer sequence and Ccnd1. (A) CH12 mouse lymphoma cell line ChIP-seq data for H3K4me1 (orange), H3K27ac (blue), and CTCF (black). (B) PCHi-C performed in pre-B cells in old mice (Koohy et al., 2018). (C) DNase capture Hi-C data from mESC (Joshi et al., 2015). (D) PCHi-C performed in mESC (Schoenfelder et al., 2015). The conserved enhancer region (the region with homologous sequence) is highlighted in the blue box. The purple arcs denote interactions between fragments. Data is visualised using the WashU Epigenome Browser.