Bortezomib (20) is a highly potent proteasome inhibitor that protects tumor suppressor proteins from degradation. It is approved and marketed for the treatment of multiple myeloma, but it also promises to be effective in combination with other targeted therapies and conventional chemotherapy in leukemia (Niewerth et al., 2013). Combination of bortezomib with the HDAC inhibitor valproic acid (27) was effective against B-cell precursor ALL in cell culture and mouse xenograft with or without simultaneous treatment with chemotherapeutic agents. Evidence suggests that the activity of the drug combination is mediated through modulation of NF-κB activity, thereby promoting apoptosis (Bastian et al., 2013). A Phase I trial in pediatric ALL and AML demonstrated the effectiveness of bortezomib as a chemosensitizer in conjunction with chemotherapy, although not as a single agent (Horton et al., 2007). Bortezomib is currently being tested in a Phase III trial in AML, as well as in ALL in two Phase II combination trials, one with the HDAC inhibitor vorinostat (28, see also “Histone deacetylases”) together with dexamethasone, and the other with combination chemotherapy. (See Figures 2G and 3B for structures of the compounds discussed in this section.)

Inhibition of the anti-apoptotic protein BCL-2 sensitizes cancer cells to chemotherapy induced apoptosis (Fulda, 2009). In cell lines and mouse xenografts derived from ALL patients, the BCL-2 antagonist ABT-737 (21) demonstrated synergistic cytotoxicity with standard chemotherapy (Kang et al., 2007). Furthermore, in the pediatric pre-clinical testing program the orally bioavailable BCL-2 antagonist navitoclax (22) (ABT-263) showed the greatest efficacy against ALL among all the cancer models tested (Lock et al., 2008). A recent study demonstrated cytotoxicity of the BCL-2 antagonist obatoclax (23) in a broad range of patient-derived infant ALL cell lines, and synergistic cell killing in combination with chemotherapeutic agents (Urtishak et al., 2013). The clinical effectiveness of obatoclax combined with standard chemotherapy is being tested in a Phase I trial in ALL and AML. (See Figure 2H for structures of the compounds discussed in this section.)

Elevated levels of DNA methylation lead to impaired expression of tumor suppressor genes and other epigenetic changes that promote precancerous and cancerous conditions (Lyko and Brown, 2005). Two DNA methyltransferase inhibitors have been approved for the treatment of myelodysplastic syndrome: 5-azacytidine (24) and decitabine (25) (Fandy, 2009). These drugs have been evaluated in clinical trials in AML, and decitabine received European but not US approval in 2012 for treatment of adult AML. Two trials in pediatric AML are pending: a Phase II trial focused on maintenance therapy employing 5-azacytidine in combination with the immunomodulatory biologic sargramostim, and a Phase I trial of decitabine in combination with the HDAC inhibitor AR-42 (30, see 2.2.2). (See Figures 3A,B for structures of the compounds discussed in this section.)

Acetylation of histones promotes transcription by increasing accessibility of chromatin and exposing acetylated lysine residues as key epigenetic markers that lead to recruitment of RNA polymerase. In the context of cancer, HDACs remove these acetyl modifications, blocking expression of key tumor suppressor genes. Following numerous trials of HDAC inhibitors, two compounds have been approved for the treatment of adult cancers (Giannini et al., 2012). Currently, pediatric leukemia trials are underway to test HDAC inhibitors as single agents and also in combination with DNA methyltransferase inhibitors (Mummery et al., 2011). Vorinostat (28) was the first HDAC inhibitor approved for clinical use. It is a relatively simple molecule containing a zinc-binding hydroxamic acid moiety that inhibits a broad spectrum of HDACs. It is in Phase II pediatric testing in ALL and AML. A Phase I trial in ALL and AML is also underway with panobinostat (29), another broad-spectrum HDAC inhibitor containing a hydroxamic acid group under development for various adult cancers. A Phase I trial in AML is due to start soon combining the DNA methyltransferase inhibitor decitabine (25) with the HDAC inhibitor AR-42 (30), which was reported to be superior to other broad-spectrum HDAC inhibitors (Lucas et al., 2010). (See Figures 3A,B for structures of the compounds discussed in this section.)

Compelling evidence links recruitment of the histone methyltransferase DOT1L to transcriptional activation of oncogenes such as HOXA9 in MLL-R leukemia (Bernt and Armstrong, 2011; Zhang et al., 2012). The enzymatic activity of the histone methyltransferase DOT1L has been shown to be essential for MLL-R leukemia in several studies. Conditional knockdown of Dot1l in mice harboring the MLL-AF6 fusion inhibited leukemia progression (Deshpande et al., 2013), and the same group demonstrated that Dot1l knockout prevented initiation of leukemia harboring MLL-AF10 (Chen et al., 2012). An earlier study showed DOT1L to be essential for HOXA9 and MEIS1 expression and the resulting leukemia driven by MLL-AF9 (Nguyen et al., 2011). DOT1L knockdown has also been shown to induce apoptosis in cells transformed with a variety of MLL fusions (Chang et al., 2010). Therefore, efforts are underway to develop inhibitors of the enzymatic activity of DOT1L. DOT1L differs structurally from all other histone lysine methyltransferases (HKMTs) in that it does not possess a SET domain comprising the catalytic site. This structural uniqueness in the architecture of the catalytic site facilitates discovery of selective inhibitors. Furthermore, DOT1L is the only enzyme known to methylate lysine 79 on histone 3 (H3K79), reducing the possibility of activation of an alternative cellular pathway to circumvent a block to DOT1L-mediated methylation.

The most significant progress in the development of DOT1L inhibitors has been made using structure-based and mechanistically guided design based on the published X-ray crystallographic structure of the catalytic domain of the enzyme (Min et al., 2003). This approach led to EPZ004777 (31), a mimic of the methyl donor substrate S-adenosylmethionine (SAM) that inhibits DOT1L with high potency and selectivity. EPZ004777 displayed selective killing of MLL-R leukemia cell lines and suppressed expression of the HOXA9 and MEIS1 oncogenes, consistent with the proposed role of the enzyme target. Moreover, despite poor pharmacokinetic properties, 7-day continuous infusion of EPZ004777 resulted in a modest but significant increase in survival in a mouse model of engrafted MLL-R leukemia. The apparent uniqueness of DOT1L’s normal role had earlier raised concerns that inhibitors would prove to be unacceptably toxic, but evidence with EZ004777 suggested otherwise (Daigle et al., 2011). Subsequently, the X-ray crystal structure of EPZ004777 bound to DOT1L was solved, and the structural insights were used to incorporate a bromine substituent in the adenine ring to give SGC0946 (32). Compared to EPZ004777, SGC0946 showed an improved K_D of 60 pM, increased residence time on the enzyme due to a slower dissociation rate, and improved cell permeability, leading to greater efficacy in reducing intracellular H3K79 methylation (Yu et al., 2012).

Identification of a hydrophobic pocket present in the SAM binding site of DOT1L but absent in all other histone methyltransferases led to the design of a SAM mimic with two key features: a benzyl moiety appended to the adenine ring to impart selective binding, and replacement of the S-methyl group with an N-iodoethyl group for irreversible covalent attachment to the ε-amino group in lysine 79 of histone 3 (33). This compound gave an IC₅₀ of 110 nM and displayed>1000-fold selectivity for DOT1L over other histone methyltransferases. However, the mechanism of action of the compound was not studied to determine whether it reacted with the lysine-containing substrate as predicted (Yao et al., 2011). A follow-up publication reported that 33 and other compounds containing a reactive N-iodoethyl group did not show the desired selective killing of MLL-R leukemia cells, so SAM analogs containing other S-methyl group replacements were synthesized. The most potent of these were analogs of EPZ004777 comprising the native adenine ring appended with N⁶ substituents. Although some cellular activity was observed in MLL-R leukemia cell lines, none of the compounds appeared to offer advantages over EPZ00477 itself (Anglin et al., 2012).

Recently the possibility has arisen that MLL-R leukemia may be targeted by blocking ubiquitinylation of histone 2B, preventing activation of DOT1L. Following a demonstration that the activity of DOT1L is markedly enhanced by interaction of the enzyme with ubiquitinylated histone 2B (McGinty et al., 2008, 2009), it has now been reported that proliferation of MLL-R leukemia cells was suppressed by knockdown of the ubiquitin ligase RNF20 that ubiquitinylates histone 2B (Wang et al., 2013). (See Figure 3C for structures of the compounds discussed in this section.)

BRD4 is a member of the BET (bromodomain and extra-terminal domain) family of proteins that binds to acetylated lysine in histones via a structural motif known as a bromodomain. In several cancers including acute leukemia, the resulting activation of transcriptional elongation promotes the expression of growth-promoting oncogenes (Deshpande et al., 2012; Popovic and Licht, 2012). An siRNA screen in AML cells revealed that inhibition of BRD4 led to cell cycle arrest and apoptosis. Structure-based design based on the X-ray crystal structure of the bromodomain in BRD4 led to (+)-JQ1 (34), a potent and selective inhibitor of binding of acetylated lysine (Filippakopoulos et al., 2010). (+)-JQ1 suppressed proliferation of AML cells in culture and also extended survival in a mouse model of AML (Zuber et al., 2011). Subsequently (+)-JQ1 was shown to be cytotoxic to B-cell ALL cell lines, especially those harboring rearrangements of CRLF2. CRLF2 heterodimerizes with the IL7 receptor (IL7R) and drives proliferation and resistance to apoptosis. Strikingly, the genes most strongly downregulated by (+)-JQ1 were MYC and IL7R (Ott et al., 2012). (+)-JQ1 also suppressed growth and induced apoptosis in primary cells from patients with freshly diagnosed and relapsed AML. Based on these promising results (+)-JQ1 will be advanced to clinical trials (Herrmann et al., 2012). Another bromodomain inhibitor, I-BET151 (35), optimized for selective binding to BRD3/4 and good in vivo pharmacokinetics, gave a promising increase in survival in two mouse models of MLL-R leukemia (Dawson et al., 2011). The role of BRD4 in activation of transcription became clearer and potentially more significant with the discovery that it possesses protein kinase activity. BRD4 phosphorylates serine 2 on RNA polymerase II and thereby promotes transcriptional initiation and elongation (Devaiah et al., 2012). (See Figure 3D for structures of the compounds discussed in this section.)

Another approach to disruption of transcriptional activation in MLL-R leukemia is to inhibit the interaction of MLL and menin. Menin appears to be a central hub that interacts with MLL fusions as well as wild type MLL and various other proteins to promote leukemia (Thiel et al., 2012). High-throughput screening for inhibitors of binding of an MLL-derived peptide to menin identified hits that were optimized for potency, leading to MI-2 (36). MI-2 was shown to disrupt the intracellular interaction of menin and MLL-AF9 and arrest cell growth in MLL-R leukemia cells, albeit with rather modest potency in the low micromolar range. Moreover, MI-2 blocked the transforming ability of MLL-AF9, suppressed the expression of the HOXA9 and MEIS1 oncogenes, and induced hematopoietic differentiation. No animal data has been reported so far; the intention is to optimize the inhibitors further to enable assessment of their therapeutic potential (Grembecka et al., 2012).

In addition to menin, various transcription factors and other proteins interact with MLL fusions to form an activation complex that promotes transcriptional elongation (Slany, 2009). In principal, several protein–protein interactions in this complex could be viable targets for disruption. One interaction that has been validated as a target with a peptide, and is being screened for small-molecule inhibitors, is the binding of MLL-AF4 to the transcription factor AF9 (Watson et al., 2013). A direct AF4–AF9 interaction has been demonstrated (Erfurth et al., 2004), and shown to be essential for the survival of leukemia cell lines containing MLL-AF4 and MLL-AF9 fusions (Srinivasan et al., 2004). Yeast two-hybrid assays identified the minimal amino acid sequence in AF4 required for binding to AF9, and a cell-permeable peptide derived from this sequence was shown to disrupt intracellular AF4–AF9 binding. Moreover, this AF4-derived peptide killed cells harboring MLL-AF4 and MLL-AF9 fusions but had no effect on normal hematopoietic progenitor cells (Srinivasan et al., 2004). (See Figure 3E for structure of the compound discussed in this section.)