Encyclopedia of Cancer, 2015
Histones are negatively charged proteins which assemble into a multi-subunit protein octamer complex around which the DNA of eukaryotic cells is wrapped. The posttranslational modification of histones, along with DNA methylation, has been commonly termed an epigenetic process, although the heritable nature of histone modifications has yet to be clearly demonstrated. Histones are aberrantly modified in cancers, leading to deregulated expression of tumor suppressors and oncogenes, events important in tumorigenesis and cancer progression. In addition, cellular levels of histone modifications are altered in cancer and are predictive of clinical outcome in a number of cancer types.
Histones and chromatin
The DNA of eukaryotic cells associates with numerous different proteins to formchromatin, the physiologically relevant form of DNA. The most abundant proteins in chromatin are the histone proteins which are small, basic proteins that are highly conserved among eukaryotic organisms. There are four canonical histone proteins, H2A, H2B, H3, and H4 which interact with each other, and DNA to form a nucleosome. Histones H2A and H2B interact with each other to form a heterodimer, while two copies each of H3 and H4 interact to form a heterotetramer. The H3–H4 heterotetramer associates with two H2A–H2B heterodimers to assemble into a histone octamer with a roughly cylindrical core in which the N-terminal tails of each of the histones extend outwards in a relatively unstructured manner.
The intimate association of DNA with histones in the context of nucleosomes has significant functional consequences on DNA-templated processes, including DNA replication, DNA repair, and transcription. The nucleosome presents an efficient barrier to these processes, thereby imposing the need for a dynamic chromatin structure. This principle is reflected partly in the heterogeneous distribution of nucleosomes and histone modifications throughout the human genome giving rise to the nucleosome rich, transcriptionally inactive heterochromatin and the relatively nucleosome depleted and transcriptionally active euchromatin structure.
Chromatin structure is also affected by the covalent posttranslational modification of the histone proteins, particularly at their N-terminal tails. The modification of histones occurs at lysine, arginine, and serine amino acid residues by the covalent attachment of chemical moieties including – but not limited to – acetyl, methyl, and phosphoryl groups. In general, acetylation of lysine residues is associated with an open chromatin conformation that allows for active gene transcription. Methylation of lysines and arginines tends to be more ambiguous in its effects on chromatin structure and transcription, with the specific functional output depending on the particular amino acid residue that is methylated.
Large families of enzymes, collectively termed histone-modifying enzymes, catalyze the modification of histones. Histone acetyltransferases (HATs) utilize acetyl CoA to attach acetyl groups to lysine residues while histone deacetylases (HDACs) remove this mark. Methylation of lysine and arginine residues is catalyzed by a family of histone methyltransferases (HMTs) using S-adenosyl methionine (SAM) as a substrate. Methyl groups are removed from histones by a family of histone demethylases (HDMs). Examples from each class of histone-modifying enzymes have been found to be altered in cancer. Translocation of the HAT p300 has been found in certain hematological malignancies, while inactivating mutations of a closely related HAT, CBP, are associated with the onset of Rubinstein-Taybi syndrome which predisposes to cancer. Elevated levels of the HMT SMYD3 occur in colorectal and hepatocellular carcinomas due to an increase in the number of repeats of an E2F transcription factor binding site, while another HMT, MLL, is found translocated in various forms of leukemia. A number of HDACs and HDMs are also found mis-expressed in multiple cancer types, including GASC1, a gene originally identified as being amplified in oesophageal squamous carcinomas, and which were only later shown to possess demethylase activity. Changes in levels of expression for the histonemodifying enzymes seem to be a common feature of cancer and may explain why expression patterns of select groups of histone-modifying enzymes can be used to distinguish cancer cells from normal cells and also have the ability to specify cancer type.
Histone-modifying enzymes are targeted to specific gene promoters by sequence-specific transcription factors or through specialized domains such as the bromodomain and chromodomain that recognize and bind to existing histone modifications. In this way, histone modifications are regulated at the level of individual genes which contribute to regulation of gene expression at these genes. In cancer, the aberrant recruitment of histone-modifying enzymes establishes altered patterns of histone modifications at specific genes including oncogenes and tumor suppressors. For example, the cyclin E gene – which has an important role in the G1 to S phase transition – is normally silenced by the deacetylation of lysine residues catalyzed by HDAC1. HDAC1 is recruited to this locus by the retinoblastoma (Rb) tumor suppressor protein. Mutations in Rb are common in cancer and can result in aberrant recruitment of HDAC1 to genes such as cyclin E that are involved in cell cycling with subsequent changes in the levels of gene expression. Similarly, an altered pattern of lysine methylation at the CDKN2A locus – which encodes the p16INK4A and p14ARF tumor suppressors – is found concurrent with silencing of this gene. The upregulation of oncogenes by epigenetic means is less well documented than the silencing of tumor suppressors, but examples have been noted. For instance, the MLL HMT is translocated in certain leukemias, leading to the upregulation of the developmentally important Hox genes, a step that may be necessary for leukemic transformation. Other than gene promoters, lower levels of acetylation of lysine 16 and methylation of lysine 20 on histone H4 have been found at DNA repetitive elements in hematological malignancies and colorectal adenocarcinomas.
Histone modifications and cancer prognosis
Differences in levels of histone modifications at the DNA repetitive elements, and at individual genes, have not proven to be predictive of clinical outcome. However, in addition to alterations at specific gene promoters, histone modifications also show large-scale deregulations in cancer. When examined at the level of individual nuclei (i.e., cellular level), cells within a tumor tissue display dissimilar levels of histone modifications, where cells have relatively high or low levels, revealing an epigenetic heterogeneity at the cellular level within cancers (Fig. 1). This cellular epigenetic heterogeneity exists not only within individual tumors but also between patients who have varying levels of cells with relatively high or low cellular levels of histone modifications. In fact, patterns of cellular levels of various histone modifications can be used to group patients into separate classes that exhibit significantly different clinical outcomes. Generally, patients with increased prevalence of cells with lower levels of histone modifications have poorer outcome. Although the biological significance and the underlying molecular mechanism of the cell-tocell variability in histone modifications are poorly understood, the clinical data suggest that cellular levels of modification may serve as a valuable biomarker in the near future.
Epigenetic therapy: HDAC inhibitors
The reversible nature of histone modifications has raised the possibility that changes in gene expression due to altered patterns of modifications can be effectively treated. An altered pattern of histone acetylation as a result of mis-targeting of HDACs is a common phenomenon seen in hematological malignancies. This type of epigenetic aberration has received much attention due to the discovery of a group of natural and synthetic molecules known to inhibit HDACs, termed HDAC inhibitors (HDACi). These molecules have been shown to inhibit HDACs by occluding access of the substrate to the active site of the enzyme. HDACi are currently utilized in chemotherapeutic regimens and can be grouped into several classes based on their molecular structure. Some HDACi inhibit specific classes of HDACs (which are grouped into three different classes based on regions of homology), while others are known to inhibit all three classes. However, the specificity of a given HDACi for a specific HDAC is poorly defined and continues to be a topic of investigation.
The treatment of transformed cells with HDACi promotes differentiation of tumor cells and leads to growth arrest andapoptosis in both cultured cell lines and animal models. HDACi exert their function selectively on cancer cells which display a tenfold higher level of sensitivity to these drugs than do their non-transformed counterparts. The use of HDACi in cancer treatment relies on the ability of these molecules to affect the gene expression program, silencing the expression of oncogenes while promoting the re-expression of silenced tumor suppressors. Perhaps surprisingly, the use of HDACi affects the expression of only a small subset of the genome, with some genes being upregulated and some being downregulated. While the exact mechanism of HDACi has not been elucidated, it is thought that they work in part by affecting chromatin structure. This has been shown to occur at tumor suppressor genes, such as the p21 gene where treatment with HDACi has been shown to increase levels of acetylation on both histones H3 and H4. In addition to increasing acetylation, the p21WAF1 locus demonstrated a more open chromatin conformation upon treatment with HDACi. This observation has raised the idea that HDACi may also work by inducing changes in chromatin structure leading to genomic or chromosomal instability. Increased genomic instability may activate the cell cycle checkpoint, leading to cell cycle arrest.
Fig. 1. Cancer tissues exhibit cellular epigenetic heterogeneity. Immunohistochemical (IHC) analysis of histone modifications in malignant prostate glandular epithelial cells from tumors of similar grade and stage reveals heterogeneity in cellular levels of specific modifications. In a given patient’s tissue, some cells have higher levels of certain histone modifications (brown nuclei indicated by the brown arrow) than other cells (blue nuclei indicated by the blue arrow). Increased prevalence of cells with low levels of histone modifications (i.e., cells with blue nuclei) indicates poorer prognosis. In the case shown here, patient 2 has increased number of cells with low levels of histone acetylation at histone H3 (lysine 18) and therefore poorer prognosis compared to patient
Esteller M (2006) Epigenetics provides a new generation of oncogenes and tumour-suppressor genes. Br J Cancer 94:179–183
Jones PA, Baylin SB (2007) The epigenomics of cancer. Cell 128:683–692
Kouzarides T (2007) Chromatin modifications and their function. Cell 128:693–705
Seligson DB, Horvath S, Shi T et al (2005) Global histone modification patterns predict risk of prostate cancer recurrence. Nature 435(7046):1262–1266
Yoo CB, Jones PA (2006) Epigenetic therapy of cancer: past, present and future. Nat Rev Drug Discov 5:37–50