N Engl J Med 2012; 367:647-657August 16, 2012
The principal tenet in oncology is that cancer is a disease that is initiated and driven by somatic aberrations of our genome. The challenge is to decipher how these genomic alterations culminate in malignant transformation and how they can be targeted for therapeutic gain. Molecular insights provided by the study of hematopoietic cancers have elucidated many fundamental principles in cancer biology.
Recurrent chromosomal translocations involving transcriptional regulators in several hematopoietic cancers illustrate the importance of transcriptional dysregulation in cancer.1 The earliest detected translocations provided examples of abnormalities in both transcriptional activation and transcriptional repression.1,2 The introduction of more refined techniques in molecular biology has led to the identification of further recurrent translocations and somatic mutations that result in pathognomonic transcriptional alterations.3 These gene-expression signatures have diagnostic utility as well as prognostic significance, but as yet few new treatments have emerged. The task of cataloguing all the genomic aberrations in cancer has now begun in earnest.4 Using next-generation sequencing platforms, the International Cancer Genome Consortium has already provided an unparalleled annotation of recurrent somatic mutations in protein-coding genes for a large variety of cancers.5 These efforts have brought into focus a new central theme: recurrent mutations in epigenetic regulators, which are especially prevalent in the hematopoietic cancers (Table 1Table 1
Epigenetic Regulators with Reader Domains Recurrently Mutated in Cancer.).
Epigenetics and Chromatin Biology
The term “epigenetics” (see the Glossary) remains the object of contention and ambiguity.6,7 It was originally coined by Waddington to describe heritable changes in gene expression and cellular phenotype that were independent of alterations in the DNA sequence.7 Epigenetics is most frequently used to describe the study of chromatin biology, and this will be the definition used in this review. Chromatin is the macromolecular complex of DNA and histone proteins. It provides the scaffold for the packaging of our entire genome and contains the heritable material of eukaryotic cells.
The basic functional unit of chromatin is the nucleosome, which consists of an octamer containing two each of the histones H2A, H2B, H3, and H4, around which 147 bp of DNA are wrapped6 (Figure 1AFigure 1The Nucleosome.).
Nucleosomes compact and package DNA in a dynamic and highly controlled fashion that caters to the multitude of DNA-based processes.
Consecutive nucleosomes are separated by unwrapped linker DNA, typically between 20 and 50 bp in length.8 Wrapped nucleosomal DNA is inherently less accessible than linker DNA, thus the genomic positioning and compaction of nucleosomes strongly influences the ability of proteins to bind target sequences within DNA and to carry out their function. Broadly speaking, chromatin can be divided into two disparate states: heterochromatin, which is tightly packaged and contains primarily inactive genes, and euchromatin, which has a more relaxed conformation that provides a more permissive environment for active transcription. Several factors influence both local and global chromatin architecture, but perhaps the most influential elements that coordinate this process are the covalent modifications of either DNA or histones (Figure 1B).8 These chromatin modifications play an instructive role in regulating DNA-templated processes, including transcription, repair, and replication.6,9
Chromatin modifications serve two main functions: first, they physically enhance or weaken the noncovalent interactions between histones or between histones and DNA, determining accessibility to specific DNA loci, and second, they provide an informative platform for the recruitment of epigenetic regulators. At least 4 different DNA modifications10,11 and at least 16 distinct classes of histone modification have been described (Figure 1B).9,12 All of these chemical modifications are dynamic; they are laid down by “chromatin writers” and removed by “chromatin erasers” in a highly regulated manner (Figure 2AFigure 2Epigenetic Regulation.).
Chromatin immunoprecipitation coupled with next-generation sequencing has made it possible to annotate comprehensive global maps in specific cellular contexts for many of these modifications.13,14 The vast array of modifications are not random and often adopt a predictable genomic distribution that can be used to define specific cellular processes, such as active transcription.14 A fundamental advance in our understanding of chromatin regulation was the realization that many chromatin regulators possess specialized domains that allow these proteins to survey the epigenetic landscape and dock at specific regions within the genome. The binding modules within these “chromatin readers” recognize different covalent modifications of the nucleosome and assemble functional complexes onto specific loci to facilitate DNA-templated processes (Figure 2B).15,16
In contrast with the static somatic alterations in DNA, the dynamic plasticity of the epigenome lends itself well to therapeutic manipulation. Epigenetic therapies targeting the catalytic activity of chromatin regulators have been developed and approved by the Food and Drug Administration for use in a small number of hematopoietic cancers.17-19 These therapies include small molecules that target epigenetic writers (DNA methyltransferase inhibitors)17,18 and epigenetic erasers (histone deacetylase inhibitors).19 Despite these clinical successes, the pleiotropic effects of these compounds have made it difficult to decipher their exact molecular mechanism of action and have hampered their broader application in oncology. More recently, it has been possible to develop and deploy small molecules that specifically target the protein–protein interaction modules of certain epigenetic readers. This new therapeutic approach will be the primary subject of this review.
Epigenetic Readers
Many chromatin regulators, including several catalytic enzymes, act as “chromatin readers,” possessing specialized domains that bind to distinct covalent modifications of the nucleosome and respond to information conveyed by upstream signaling cascades.15 Critical residues within the binding pocket of the reader domain confer a preference for specific modification states, whereas residues outside the binding pocket contribute to determining the histone-sequence specificity. This modular combination allows proteins with similar binding domains to dock at different modified residues, or at the same amino acid displaying a different modification state.
For example, lysine residues have been shown to contain at least eight different covalent modifications, including acetylation, methylation, ubiquitylation, and sumoylation. Further complexity exists by virtue of the fact that each lysine residue may be unmethylated, monomethylated, dimethylated, or trimethylated. For instance, the plant homeodomain (PHD) finger is a region within a protein that is capable of detecting methylated histones. The PHD fingers of the proteins BHC80 and DNMT3L preferentially bind unmethylated lysine residues,20,21 whereas the PHD finger of ING2 binds most avidly to dimethylated and trimethylated lysines.22,23 Other, similar methyl–lysine-recognition motifs exist, including chromodomains, tudor domains, and the PWWP domain, named for its characteristic proline–tryptophan–tryptophan–proline motif (Figure 2B).24-26 Notably, when the same lysine residue undergoes another modification, such as acetylation, it may then provide docking sites for other proteins containing acetyl–lysine binding domains, such as bromodomains.27,28 Finally, to add to the complexity, many chromatin regulators have more than one type of reader domain, and their chromatin binding can be further influenced by neighboring histone modifications, so-called multivalent engagement-of-histone modifications.29 These examples highlight the multifaceted mechanisms that chromatin readers use to decipher the intricate epigenetic landscape. (The best-characterized protein-binding pockets contained within chromatin-associated proteins are summarized in Figure 2B.)
The importance of chromatin reader domains in maintaining homeostasis was initially revealed in the seminal observation that mutations in the PHD finger of RAG2 abrogate the protein's ability to bind trimethylated-H3K4, reduce V(D)J recombination (in which variable, diverse, and joining gene segments are randomly combined), and result in immunodeficiency syndromes.30 Mutations that abrogate the chromatin-reading capacity of many epigenetic regulators play an influential role in a variety of diseases, including cancer.31 Using these chromatin reader modules as therapeutic targets may offer a unique opportunity to tailor therapies to specific diseases. An exemplar of this process has recently been provided with small molecules that specifically and avidly inhibit the tandem bromodomains of the bromodomain and extraterminal (BET) family of proteins.32-34
BET Bromodomain Inhibitors
Bromodomains are highly conserved motifs present in a number of proteins throughout phylogeny.27 More than 40 different human proteins contain a bromodomain, and some have multiple bromodomains. They can be clustered into nine major families according to sequence identity. Although the acetyl–lysine binding pocket for all bromodomains is hydrophobic, there can be considerable variation in the electrostatic interactions at the opening of the pocket among bromodomain families. This variation determines the specificity of individual bromodomains and provides the opportunity to develop specific small molecules that are targeted against certain families of bromodomains, including the BET proteins.35,36
The BET family has four members, including bromodomain-containing proteins 2, 3, and 4 (BRD2, BRD3, and BRD4), whose expression is ubiquitous, and BRDT, whose expression is confined to the germ cells. BET proteins share a common structural design, featuring a tandem bromodomain at the N-terminal end of the protein, and play an integral role in transcription and cell growth.37 Early studies characterizing the BET proteins showed that BRD4 is associated with the mediator complex,38 an important multicomponent protein complex facilitating the initiation of transcription.39 BRD4 also associates with the active form of positive transcription elongation factor b (P-TEFb), a complex that regulates RNA polymerase II activity at the onset of transcriptional elongation.40,41 BET proteins also maintain an association with chromatin throughout mitosis, and this feature facilitates “gene bookmarking,” a process of rapid transcriptional reactivation of critical genes after mitosis.42 Underscoring the physiological importance of the BET proteins in cellular homeostasis is the observation that the experimental knockout of either Brd2 43 or Brd4 44 in mice results in early embryonic lethality.
The essential nature of the BET proteins and their fundamental importance in the regulation of transcription has increased interest in elucidating the molecular mechanisms of their action. These efforts will be aided considerably by the recent identification of the complete nuclear BET-protein interactome.32 In this study, three complementary global proteomic strategies were used to isolate the nuclear complexes containing the BET proteins. These data show that BET proteins are integral components of a large number of nuclear protein complexes that play a role in DNA replication, chromatin remodeling, DNA damage, and transcriptional regulation. Recurrent translocations of both BRD3 and BRD4 underpin the pathogenesis of NUT (or nuclear protein in testis) midline carcinoma,33 and more recently, RNA interference (RNAi) screening strategies have been used to identify a central role for BRD4 in acute myeloid leukemia.45 These findings have provided the impetus for investigating the potential of BET bromodomain inhibitors as novel anticancer agents.BET Inhibitors in Cancer
Three different BET inhibitors (I-BET762,34 JQ1,33 and I-BET15132) show highly specific binding to both of the tandem domains of BRD2, BRD3, and BRD4 and inhibit their ability to engage with acetyl–lysine residues (Figure 3AFigure 3The BET Inhibitors.). These agents inhibit growth in a range of hematopoietic malignant cell lines32,45-47 and in human NUT midline carcinoma cell lines.33 Although the therapeutic efficacy of BET inhibition has not yet been tested in patients, its efficacy has been shown in vivo in murine models.
NUT Midline Carcinoma
NUT midline carcinoma is a rare but aggressive epithelial malignant disease that is invariably fatal, with a mean survival of less than 1 year.48 It is most often manifested in the midline of the upper aerodigestive tract and mediastinum. Histologic diagnosis of NUT midline carcinoma is difficult, and confirmation relies primarily on the demonstration of rearrangement of the NUT gene with the use of fluorescence in situ hybridization.48 BRD4 and BRD3 represent the NUT translocation partners in the majority of cases, and these fusions result in the aberrant localization of NUT to chromatin. Knockdown of the NUT fusions resulted in dramatic differentiation and growth arrest of the malignant cells, providing a therapeutic rationale for the targeting of BET proteins. Remarkably, the BET bromodomain inhibitor JQ1 was able to displace the BRD4–NUT fusion protein from chromatin and induce a rapid differentiation and arrest of proliferation in NUT midline carcinoma cell lines.33 In addition, JQ1 showed excellent efficacy in murine xenograft models of NUT midline carcinoma, resulting in tumor differentiation and regression and increased survival.33
Acute Myeloid Leukemia with MLL Translocations
Recurrent translocations of MLL, the mixed-lineage leukemia gene, result in aggressive leukemias with a poor prognosis that are often refractory to conventional therapies.49 More than 70 different MLL-translocation partners have been identified in leukemia.49 Despite this variation, a central abnormality in transcriptional elongation appears to underpin the molecular pathogenesis of this disease.50 Many MLL fusion partners are members of the superelongation complex (SEC), an important regulator of transcriptional elongation.51-53 Moreover, the N-terminal portion of MLL, a region conserved in all MLL fusion proteins, associates with another critical transcriptional complex called the polymerase-associated factor complex (PAFc).54,55 The functional integrity of these complexes is critical for malignant transformation by MLL fusion proteins.52-55 Two recent studies using complementary approaches have identified an essential role for the BET proteins in a broad range of acute myeloid leukemias, including MLL-translocated leukemias.32,45
Dawson and colleagues used a global proteomic approach that identified BRD3 and BRD4 as key components of both PAFc and SEC.32 In contrast, Zuber and colleagues used an RNAi screen to show that the depletion of BRD4 dramatically reduced the viability of MLL-AF9 leukemia in vitro and in vivo.45 Both JQ1 and the novel BET inhibitor I-BET151 showed remarkable efficacy in vitro and in vivo against MLL fusion leukemia, resulting in the rapid induction of cell-cycle arrest and apoptosis. Both studies highlighted a reduction in expression of critical regulators of transformation, including MYC, BCL2, and CDK6 (a cyclin-dependent protein kinase), after treatment with a BET inhibitor. Dawson and colleagues also found that the therapeutic efficacy of I-BET151 could be attributed at least in part to the inhibition of BRD3–BRD4-mediated recruitment of PAFc and SEC to chromatin (Figure 3). These molecular events resulted in reduced recruitment of RNA polymerase II to the promoters of crucial oncogenes such as MYC, BCL2, and CDK6.32 BET inhibition was also effective against primary human acute myeloid leukemia cells in vitro.32,45 Notably, in primary human MLL-translocated leukemia cells, BET inhibition dramatically reduced the clonogenic capacity of the leukemia stem-cell compartment, suggesting that disease eradication may be possible.
Although Zuber and colleagues focused primarily on MLL-translocated leukemias, they also showed the efficacy of BET inhibition in a range of acute myeloid leukemia subtypes in which MLL rearrangement was absent. These findings raise the intriguing possibility that BET inhibition may be a more broadly applicable therapy in this genetically heterogeneous disease.
Multiple Myeloma and Burkitt's Lymphoma
Multiple myeloma is an incurable plasma-cell dyscrasia. Although the disease is genetically heterogeneous,56 disordered expression of the myelocytomatosis viral oncogene homolog MYC is a prominent feature, providing the rationale to test for BET inhibition.57 Similarly, Burkitt's lymphoma, an aggressive lymphoproliferative disorder, is characterized by recurrent translocations of MYC, most frequently involving the IGH locus that results in constitutive high-level expression of MYC. Several multiple myeloma cell lines show a dramatic response to BET inhibition with JQ1.46,47 Treatment with JQ1 leads to a profound cell-cycle arrest and apoptosis of the cell lines; this cellular phenotype is associated with a reduction in MYC transcription and protein expression. This transcriptional repression of MYC is linked to a reduction in the binding of chromatin by BRD4, upstream of the MYC promoter. Delmore and colleagues further translated these findings into work with primary human multiple myeloma cells and murine models of multiple myeloma, further indicating the therapeutic potential of BET inhibition in this disease.46 In addition, Mertz and colleagues extended their findings to reveal a survival benefit in murine xenograft models of Burkitt's lymphoma and acute myeloid leukemia.47
Taken together, the studies involving BET bromodomain inhibitors further validate the use of epigenetic targets in cancer therapy. They show that it is possible to target epigenetic readers as well as catalytic writers and erasers. Moreover, the studies establish that the specific targeting of protein–protein interactions can have antitumor effects in vitro and in animal models. Several questions warrant further investigation. It is unclear why these drugs have effects in hematopoietic cancers but not in the majority of common solid tumors, including malignant tumors of the breast and cervix.47 It is also unclear why a compound that inhibits the localization of proteins germane to transcription in general alters the expression of only a few hundred specific and reproducible genes. Although we have some insights into the molecular mechanism determining efficacy in the MLL fusion leukemias32 and in NUT midline carcinoma,33 it is largely unclear what molecular events dictate efficacy in the other sensitive cancers. A central theme in all the studies cited is the down-regulation of MYC. However, even though MYC has a prominent role in cancer,58 it is unlikely that the profound effects observed with the introduction of BET inhibition are mediated solely by MYC inhibition. There are many malignant cell lines that overexpress MYC yet fail to respond to BET inhibition47; MYC expression is not always affected by BET inhibition,47 and MYC down-regulation is not predictive of a response to BET inhibition.32,45 Furthermore, MYC overexpression fails to prevent the apoptosis induced by BET inhibition.45 Identification of the complete BET protein interactome has shown that the BET proteins are integral components of a large number of nuclear protein complexes.32 This finding suggests that BET proteins are involved with a variety of molecular mechanisms.
Conclusions
Cancer epigenetics is an area of ongoing research that continues to inform our understanding of the molecular pathogenesis of cancer and to identify novel therapeutic targets. Advances in medicinal chemistry have now made it possible to specifically target not just the catalytic activity of epigenetic regulators but also the protein–protein interaction modules that localize many of these proteins to chromatin. However, as is the case for the majority of tumors sensitive to BET inhibition, epigenetic therapeutic targets are not necessarily mutated in sensitive tumor types. Therefore, simple mutational screening may not provide a predictor of response. Potential means of identifying sensitive tumor types include combining drug sensitivity with mutational screening, as has recently been reported in two large screening studies,59,60 and using transcriptional or epigenetic biomarkers to predict response. The successful use of the bromodomain inhibitors in cancer therapy is likely to further energize basic scientists, clinicians, and the pharmaceutical industry to search for compounds that may similarly target and disrupt other chromatin reader motifs, including chromodomains and PHD fingers, which are already known to play key roles in certain cancers.31
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.
We thank Andy Bannister for reviewing an earlier version of this article and for helpful comments; and Rab Prinjha and Chun Wa Chung for providing the chemical structures and surface structure used in Figure 3A.
Source Information
From the Department of Haematology, Cambridge Institute for Medical Research (M.A.D., B.J.P.H.), Cambridge University Hospitals National Health Service Foundation Trust (M.A.D., B.J.P.H.), the Gurdon Institute and Department of Pathology (M.A.D., T.K.), and the Wellcome Trust and Medical Research Council Cambridge Stem Cell Institute (B.J.P.H.), University of Cambridge — all in Cambridge, United Kingdom.Address reprint requests to Dr. Huntly at the Department of Haematology, Cambridge Institute for Medical Research, University of Cambridge, Cambridge CB2 0XY, United Kingdom, or at bjph2@cam.ac.uk.
No hay comentarios:
Publicar un comentario