Epigenetic polypharmacology: from combination therapy to multitargeted drugs
- 2.9k Downloads
The modern drug discovery process has largely focused its attention in the so-called magic bullets, single chemical entities that exhibit high selectivity and potency for a particular target. This approach was based on the assumption that the deregulation of a protein was causally linked to a disease state, and the pharmacological intervention through inhibition of the deregulated target was able to restore normal cell function. However, the use of cocktails or multicomponent drugs to address several targets simultaneously is also popular to treat multifactorial diseases such as cancer and neurological disorders. We review the state of the art with such combinations that have an epigenetic target as one of their mechanisms of action. Epigenetic drug discovery is a rapidly advancing field, and drugs targeting epigenetic enzymes are in the clinic for the treatment of hematological cancers. Approved and experimental epigenetic drugs are undergoing clinical trials in combination with other therapeutic agents via fused or linked pharmacophores in order to benefit from synergistic effects of polypharmacology. In addition, ligands are being discovered which, as single chemical entities, are able to modulate multiple epigenetic targets simultaneously (multitarget epigenetic drugs). These multiple ligands should in principle have a lower risk of drug-drug interactions and drug resistance compared to cocktails or multicomponent drugs. This new generation may rival the so-called magic bullets in the treatment of diseases that arise as a consequence of the deregulation of multiple signaling pathways provided the challenge of optimization of the activities shown by the pharmacophores with the different targets is addressed.
KeywordsPolypharmacology Epigenetic drugs Combination therapies
- AKT (PKB RAC)
Protein kinases B
Acute lymphoblastic leukemia
Acute myeloid leukemia
Acute promyelocytic leukemia
Anatomical therapeutic chemical
Breakpoint cluster region—Abelson
Bromodomain and extra terminal domain
B cell chronic lymphocytic leukemia-lymphoma-like 11 gene
ATP-dependent helicase SMARCA4, a transcription activator
Chronic myelogeneous leukemia
Disruptor of telomeric silencing 1
Epidermal growth factor receptor
Enhancer of zeste homologous 2
Flavin adenine dinucleotide
G protein-coupled receptor
Human epidermal growth factor receptor
Histone lysine methyltransferase
3-hydroxy-3-methylglutaryl coenzyme A reductase
Inosine monophosphate dehydrogenase
JmjC domain-containing demethylases
Kinase insert domain receptor
Lysine specific demethylase 1
Mixed lineage leukemia
Maximum tolerated dose
Natural killer cell large glanural lymphocytes
NUT midline carcinoma
Non-small cell lung cancer
Phospho-protein kinase B
Peripheral blood mononuclear cells
Phospho-protein kinase RNA-like endoplasmatic reticulum kinase
Protein arginine methyltransferase
Suberoylanilide hydroxamic acid
Selective ER downregulator
Selective estrogen receptor modulators
Su(var)3-9: suppressor of position-effect variegation; E(z): enhancer of zeste; Trx: trithorax
Silent information regulator 2
Tyrosine kinase inhibitors
Tumor suppressor gene
Vascular endothelial growth factor
Vascular endothelial growth factor receptor
Principles of polypharmacology
Notwithstanding the success of combination therapy, the use of a single drug that modulates several targets might be therapeutically advantageous over the use of drugs in combination. In cancer, the design and synthesis of new molecules that simultaneously modulate multiple biochemically distinct oncogenic targets is of current interest. Polypharmacology refers to the ability of drugs to interact simultaneously and specifically with multiple targets (multitarget drugs). Although polypharmacology might be associated with compound promiscuity, it should be considered a different category since the so-called promiscuous drugs are instead a class of compound that show a wide spectrum of biological activities and adverse reactions .
Among the advantages of multitarget drugs vs drug combinations are the more predictable pharmacokinetic (PK) and pharmacodynamic (PD) relationship of the components of a single medicine, the possibility that one motif might improve the bioavailability of the second entity, the greater efficacy against advanced-stage diseases, the lower toxicities, the simultaneous presence of the chemical entities in multiple tissues, and the improved patient compliance . To benefit from those effects, it is required that the multitarget drug exhibit balanced in vitro and in vivo activities to match potency for the corresponding targets, as well as optimized PK and safety profiles. A combination of drugs faces the problem of the different solubilities that may modify the bloodstream uptake, which requires fine-tuning the formulation in order to ensure the required blood level of each drug. In addition, the regulatory requirements are more complex when the agents are used in combination, since the safety profile of each drug needs to be demonstrated before clinical trials, and this can be further delayed due to regulatory and IP issues, in particular if the two drugs are being developed by different companies .
Efforts are underway to use chemoinformatics to help understand drug effects from a signal transduction network perspective , to confidently predict new molecular targets for known drugs, and to explain polypharmacology. Another current trend in therapy is drug repurposing or the re-discovery of a new therapeutic area for a drug used traditionally to treat a given pathology, either through the ability to modulate an additional target or by the involvement of the primary target in multiple pathologies. Examples include the use of the anti-angina drug sildenafil to treat sexual disfunction or the infamous sedative thalidomide as therapy for multiple myeloma. Indeed, the polypharmacology of current drugs has been studied using a statistical ligand-based approach . This study, aimed to discover chemical similarities between drugs and ligand sets, has revealed unanticipated promiscuities but also previously unreported polypharmacologies. The screening study of the 3665 FDA-approved and investigational drugs was conducted using databases containing the chemical structures of hundreds of thousands of biologically active compounds for which the binding characteristics to a panel of 1400 target proteins were known . A massive network of interactions (nearly 7000 of them with high probability) for the studied compounds with off-targets were predicted, which indicates that polypharmacology is, perhaps unintentionally, a feature intrinsic to the therapeutic efficacies of drugs.
Multikinase inhibitors, for example, initially considered to be highly specific for one of the 518 kinases of the kinome, have proven successful in treating previously refractory cancers, perhaps as a result of simultaneous inhibition of multiple kinases. As an example, sunitinib, a promising drug for the treatment of anaplastic thyroid cancer, inhibits 79 kinases with K D < 10 μM. Therefore, the success of (multi)kinase inhibitors in treating cancer is a consequence of the modulation of multiple signaling pathways that support cancer cell proliferation, apoptosis, angiogenesis and recruitment of surrounding tissues.
Also, in infectious diseases, current drugs show off-target effects. This is the case with the HIV protease inhibitor nelfinavir, which has been found to also inhibit the proliferation of cancer cells due to a weak modulation of multiple kinases.
Polypharmacology is prevalent in the area of CNS diseases. The activity of drugs acting on the CNS is often mediated by G protein-coupled receptors (GPCRs), a group of receptors that are also implicated in multiple therapeutic areas and share structural and functional similarities that makes selectivity a very difficult issue. For example, the use of atypical antipsychotic clozapine is associated with undesired side effects, such as diabetes and seizures, which may be due to its broad range of targets, among them different isoforms of the serotonin, dopamine, muscarinic, and adrenergic receptors, members of the GPCR superfamily. On the other hand, a single drug exhibiting polypharmacology for more than one target of the same disease could exhibit synergistic effects. This is the case of ladostigil, an inhibitor of acetylcholine esterase (AChE) and the brain monoamine oxidases (MAO) A and B, which has shown efficacy in models of Alzheimer’s disease.
Hybrid molecules , in contrast to pro-drugs, contain two (or more than two) domains with different biological functions and dual activities that ideally act as distinct pharmacophores, although not necessarily on the same biological target. Thus, multiple ligands usually consist of the combination of pharmacophores of selective ligands (either already known drugs or candidates). From the point of view of the medicinal chemist, pharmacophores that are similar and share common substructures, usually hydrophobic or basic ring systems, can be synthetically fused or merged (see examples in next section). Alternatively, if pharmacophores are dissimilar, they can be joined as conjugates with cleavable or non-cleavable linkers, although this strategy often leads to structures of high molecular weight (MW) and lipophilicity .
When both pharmacophores are connected by a linker that is labile or can be easily cleaved in vivo, they are called dual or mutual pro-drugs since each of them uses another pharmacologically active compound instead of some inert molecule as carrier. Being released simultaneously inside the cancer cells, they might act synergistically and affect distinctive cellular targets, in contrast to the simultaneous administration of two individual synergistic agents, which are usually transported to the site of action with different efficiencies.
Examples of hybrid anticancer molecules containing an epi-drug and another antitumor agent connected via a linker are shown in Scheme 1. The scaffold of the HDACi dacinostat (LAQ-824, compound 7; an early candidate that was further improved as panobinostat 8) and a tubulin binder (thiocolchicine 9 and paclitaxel 10) were connected via a disulfide bond as in 5 and 6 . Glutathione 3 would release the thiolates via disulfide exchange reactions, which in turn would produce the thiolactones to free the second component.
Mutual pro-drugs of all-trans-retinoic acid (ATRA) and several HDACis (butyric acid 13, tacedinaline 15 and entinostat 17) have been engineered via glycine acyloxyalkyl carbamate linker (which would be presumably cleaved by esterases, compounds 11a-12 and 11a-14) or through a benzyl ester linker (which would be presumably released through a 1,6-elimination reaction, compounds 11b-12 and 11b-16 and 11b-14) . The last series of mutual pro-drugs showed potent inhibition of the growth of several hormone-insensitive/drug resistant breast cancer cell lines and the hormone-insensitive PC-3 prostate cancer cell line .
Still, the development of multitarget drugs from leads is more complex than that of single drugs. Drug-like molecular properties for multiple pharmacological activities must be optimized and unintended interactions with additional targets minimized. Moreover, balancing the pharmacological activities is another complication, as often the optimal ratio is not 1:1. For example, although the hybrid compounds 5 and 6 were able to retain antimitotic and proapoptotic activity, the potency of the construct was lower than anticipated .
Finally, according to the mechanism of action, hybrid molecules can be classified in three different categories: (a) both entities interact with the same target (“double sword” molecules); (b) both entities independently interact with two different and nonrelated targets; (c) both entities interact simultaneously with two related targets at the same time .
Challenges for rational epigenetic drug polypharmacology
The new paradigm of single chemical entities that antagonize multiple biochemically distinct targets to overcome conventional single-target therapeutics is being pursued in the epigenetic field, in particular for the treatment of cancer [11, 12]. The challenge in this field is the design of small molecules that have the property to modulate at the same time several of the epigenetic targets with contrasting or totally unrelated mechanism of action. Promiscuity, traditionally considered an undesired property of drugs, might turn out to be advantageous also in epigenetics and the polypharmacology of these epi-drugs a feature intrinsic to their therapeutic efficacies.
In principle, since some of the epigenetic enzymes such as sirtuins (SIRTs), protein arginine methyltransferases (PRMTs), DNA methyltransferases (DNMTs), and lysine methyltransferases (KMTs) use the same cofactor or cofactors containing adenosine, modulators of several of these enzymes that bind to the corresponding adenosine pockets can be designed, and moreover, these might also cross-react with related receptors such as kinases. Likewise, the metalloenzymes HDACs and Jumonji lysine demethylases (KDMs) can be subjected to simultaneous inhibition with metal-chelating containing compounds. However, these simple assumptions cannot be extrapolated to the different protein families. For example, the S-adenosyl methionine (SAM) binding site of lysine methyltransferases is more extended than that of other enzymes using the same cofactor such as DNMTs. Both the SAM cofactor and the substrate of KMTs access the protein from opposite faces in domains linked through a narrow hydrophobic channel. In addition, the SAM cofactor adopts different conformations in the domains of KMTs compared to PRMTs although both enzymes transfer a methyl group to protein side-chains.
Even more challenging is the inhibition of epigenetic enzymes with unrelated mechanistic principles. To get a glimpse of the difficulties expected in the rational design of multiple epigenetic ligands, a brief description of the reaction mechanisms for the most common epigenetic enzymes follows.
Writers/erasers of acetyl groups
The acetylation status of lysine ε-amino residues in histones is under the control of the opposing activities of histone deacetylases and histone acetyltransferases (HDACs and HATs, respectively). In addition to the regulation of chromatin function and structure, acetylation has a broad regulatory role in many biological processes (cell cycle, splicing, nuclear transport, actin nucleation , cellular metabolism [14, 15], etc.) beyond chromatin remodeling. These effects might be due to the modulation by the acetylation/deacetylation mechanistic switch of the activities of a large number (more than 1700) of histones and non-histone proteins, among them tubulin, p53, Hsp90, and NFYA (nuclear transcription factor Y subunit alpha) .
Mechanism(s) of acetyl transfer to lysine residues
Mechanism(s) of acetyl-lysine hydrolysis by Zn2+-dependent deacetylases
The histone deacetylase family is composed of 18 members , which are divided into two groups depending on their mechanism of action: the classical Zn2+-dependent enzymes (HDAC1-11) and the NAD+ cofactor-dependent enzymes (SIRT1-7). The metalloprotein HDACs can be further classified into three groups: class I (HDAC1–3 and 8); class II (HDAC4–7 and 9–10), which may be divided into two subclasses, class IIa (HDACs 4, 5, 7, and 9) and class IIb (HDACs 6 and 10) on the basis of evolutionary relationships; and class IV, composed of HDAC11. Differences between class I and II HDACs are primarily noted in their size (with class II being from two to three times larger), their cellular localization, the conservation of sequence motifs in their catalytic domains, the identity of the protein-protein interaction complexes, and their tissue distribution.
HDAC inhibitors  emulate the native acetylated lysine using a Zn2+-chelating “head group” attached via a connector of variable length and functionality to a cap region. The Zn2+-chelating “head groups” reported in HDACis includes virtually all functionalities known to bind transition metal ions (hydroxamic acids, thiols, mercaptoamides, trifluoromethylketones…), which most likely compete with the natural substrate after binding site occupancy . For general inhibition by hydroxamic acids, a spontaneous proton transfer to an active site histidine upon binding of the inhibitor to the zinc was supported by recent computations; accordingly, for thiol-containing inhibitors (or precursors such as disufides or thioesters), the thiolate appears to be the active species .
Mechanism(s) of acetylated lysine deacetylation by sirtuins
Writers/erasers of methyl groups
Mechanism of methyl transfer catalyzed by DNMTs
Mechanism of methyl transfer catalyzed by HMTs
Similarly, the addition of methyl groups to arginine residues catalyzed by PRMTs uses SAM as cofactor but can produce mono- and/or dimethylarginine derivatives, the latter as the symmetric or non-symmetric isomers (Scheme 6) .
Nature uses two unrelated mechanisms for the removal of methyl groups from the methylated lysine and arginine residues [38, 39], each catalyzed by different demethylase enzymes : (a) lysine specific demethylase 1 (LSD1/KDM1) and (b) Jumonji JmjC domain-containing demethylases (JHDMs).
The bromodomain and extra terminal domain (BET) family of tandem bromodomain-containing proteins (BRD2, BRD3, BRD4, and BRDT) exhibit a wide variety of biological effects and are the first readers to be targeted in epigenetic drug discovery. They are promising agents for the treatment of a spectrum of human diseases, ranging from cancer and inflammation to viral infections [47, 48, 49]. The binding of small molecules to the acetyl-lysine pocket (KAc) can block the recognition of their acetylated partner proteins via protein-protein interactions. Despite their overall structural similarity [47, 48, 49], subtle differences exist between the bromodomain structures and their functions that can account for their specificity.
Epi-drugs in the clinic
The first drugs targeting epigenetics in fact predated a clear understanding of such mechanisms at the molecular level and the identification of the proteins responsible. Instead, the early compounds were advanced through the drug discovery process on the basis of their phenotypic effects in cancer models without knowledge of the precise targets.
The recognition that analogues of purine and pyrimidine nucleosides might act as anti-metabolites that dirsupt nucleic acid biosynthesis or function led to many such molecules being investigated. At the Czech Academy of Sciences, Piskala and Sorm synthesized 5-azacytidine (18) and the corresponding deoxyribose analogue 19 and demonstrated antileukemic activity in cells and AKR mice. Clinical trials with 18 began in 1967 in Europe and in 1971 an Investigational New Drug application was filed with the American National Cancer Institute. This was however rejected due to an unacceptably high level of toxicity. Interest in these nucleosides was rejuvenated due to the 1980 publication by Jones identifying DNMTs as their molecular target. Both 18 and 19 went back into clinical trials for the treatment of myelodysplastic syndrome, a hematological stem cell disorder that frequently progresses to acute myeloid leukemia, and were approved by the FDA in 2004 and 2006, respectively . The nucleosides 18 and 19 are pro-drugs that enter the cell through transporters and are metabolically converted to the 5′-triphosphate of 19. The triphosphate is incorporated into DNA strands, recognized as a cytosine substrate by DNMTs, and forms a covalent adduct with the enzyme via addition of the active site Cys residue to C-6 of the azapyrimidine heterocycle (see Scheme 5). The drugs are thus irreversible DNMT inhibitors, but their lack of selectivity between DNMT isoforms may be one reason for the high toxicity observed in settings other than myelodysplastic syndrome.
Combination therapies with epi-drugs towards polypharmacology
Current clinical practice uses drug combination therapies rather than single drugs  to treat patients with complex diseases . The first clinical success with combination chemotherapy for childhood acute lymphoblastic leukemia (ALL) consisted of the co-administration of the anti-folate methothrexate, the tubulin-targeting vincristine (a Vinca alkaloid), the antimetabolite 6-mercaptopurine and the steroid prednisone. Either such a drug cocktail containing two or more individual tablets to combine therapeutic mechanisms or the co-formulation of two or more agents in a single tablet are the traditional modalities of drug combinations. The design of a drug combination aims to simultaneously block disease-related targets and is expected to ensure a more durable control of the disease progression compared to single agents. Therefore, the individual drugs should be active against their own target and ideally elicit synergistic effects when used in combination without increasing the toxicity and reducing drug resistance. Mathematical models have been recently developed that analyse the dynamics of pairs of drugs in a weighted linear superposition in order to obtain predictive drug effects (synergy, independence, antagonism…) from their use as multidrug and multidose combinations .
Following the trait mentioned above for the treatment of ALL, numerous combination therapies have been investigated for treating complex pathologies such as cancer, parasitic diseases, and multiple sclerosis that are polygenic in nature and result from the deregulation of complex protein networks. New drugs in the market, in particular those with a defined mechanism of action or target, are studied in combination even before they are launched. For example, in cancer, there are a large number of clinical studies that combine the proteasome inhibitor bortezomib with other drugs targeting not only the epigenome but also Hsp90, kinases, farnesyltransferases, etc., for both solid tumors and leukemias.
Post-genomic research over the last decade is shifting the focus of rational combination modalities to what is called “personalized medicine.” In the case of cancer, it involves targeting pathogenic oncogene and non-oncogene addictions, synthetic lethalities, and other vulnerabilities, attacking complementary cancer hallmarks or distinct cell populations with molecular targeted agents and using in addition other therapeutic options such as cytotoxic chemotherapy .
Despite the success of HDACis as single agents in the treatment of hematological maligancies, the treatment of patients with solid tumors has demonstrated limited clinical benefit . For example, vorinostat 1 failed as monotherapy for the treatment of metastatic breast cancer in clinical trials . This failure has prompted the investigation of novel treatment combinations with other cancer therapeutics, including kinase inhibitors, DNA-damaging chemotherapeutic agents, radiotherapy, hormonal therapies, and other epi-drugs (primarily DNA methyltransferase inhibitors), for which a rationale has been described .
In the case of tyrosine kinase inhibitors (TKIs), combination and multitarget therapies, including epigenetic drugs, are being developed since a large number of patients do not respond to single therapy or develop resistance. The results are encouraging. Vorinostat 1 and sorafenib 25 appear to interact in a synergistic fashion to kill carcinoma cells by activating CD95 through generation of ROS due to induction of cytosolic Ca2+ that elevates dihydroceramide levels . Vorinostat 1 and other antagonists of receptor tyrosine kinase induced a synergistic induction of growth inhibition and apoptosis in the treatment of non-small cell lung cancer (NSCLC) (NCT00251589) (NCT00503971). The HDACi MPT0E028 45 (shown in Fig. 4 below) enhances erlotinib (26)-induced cell death in epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKI)-resistant NSCLC cells . Combination of EGFR-TKIs with vorinostat 1 resulted in significantly decreased cell viability through the activation of the apoptotic pathway and caspase-independent autophagic cell death .
Combination of vorinostat 1 with second-generation TKIs such as afatinib 27 or third-generation TKIs including WZ4002 28 enhanced anti-tumor effect on xenografts of H1975 cells in vivo. The combination of new generation EGFR-TKIs and vorinostat 1 may be a new strategy to overcome the acquired resistance to EGFR-TKIs in T790M mutant lung cancer .
Synergistic effects of vorinostat 1 or sodium butyrate 23 with imatinib 29, an ABL kinase inhibitor that can kill Breakpoint cluster region—Abelson (BCR-ABL) positive chronic myeloid leukemia (CML) cells, were observed and shown to enhance apoptosis in BCR-ABL expressing CML cells. The combination treatment was also effective against imatinib-refractory CML. Both wild-type BCR-ABL and the T315I mutant form of BCR-ABL, which is resistant to imatinib, were equivalently degraded following that combinatorial treatment [62, 63].
HDACis have been shown to downregulate estrogen receptor (ER) and androgen receptor (AR) mRNA in receptor-positive breast and prostate cancer cells [64, 65, 66]. Current drug therapies include tamoxifen 32 and raloxifene 33, competitive ER inhibitors that act as selective estrogen receptor modulators (SERMs), and pure anti-estrogens such as fulvestrant 34, which act as a selective ER downregulator (SERDs). Raloxifene 33 is an antagonist in all tissues, whereas tamoxifen 32 displays partial agonistic activity in a tissue and gene specific manner. In ER-negative cells, silenced ERs can be re-expressed using HDACi, restoring sensitivity to tamoxifen 32 . HDACi increase the antitumor effects of tamoxifen 32 in several ER-positive breast cancer cell lines and breast tumors that are resistant to tamoxifen (NCT00365599) (NCT01194427) (NCT02395627). Co-treatment of breast cancer cells with HDACi and tamoxifen 18 produced a synergistic effect with depletion of both ER and progesterone receptor (PR), and this effect was exclusive of HDAC2-selective inhibitors . In phase II clinical studies, the combination of vorinostat 1 and tamoxifen 32 is well tolerated by patients with ER-positive metastatic breast cancer progressing on endocrine therapy and exhibits promising activity in reversing hormone resistance. A 19 % objective response rate and a 40 % clinical benefit rate were noted .
A phase I study of the histone deacetylase inhibitor entinostat 17 in combination with 13-cis-retinoic acid 36 was carried out in patients with solid tumors, but no tumor responses were seen .
A phase I study has been initiated to evaluate the safety and efficacy of oral panobinostat 8 in combination with radiotherapy or chemoradiotherapy in patients with inoperable stage III NSCLC .
The approved DNMTi are likewise undergoing clinical studies in combination with other agents. Promising results have been obtained in the combination of DNMTi decitabine 19 plus TIK dasatinib 31 in phase I/II clinical studies in patients with CML (NCT1498445) . Decitabine 19 combined with the DNA-damaging agents carboplatin 40 is in phase II clinical trials in platinum-resistant ovarian cancer (NCT00477386) .
The LSD1 inhibitor tranylcypromine (TCP, 41) combined with all-trans-retinoic acid 37 (Fig. 3) is now in clinical trials for the treatment of adult patients with acute myeloid leukemia (AML) and myelodisplastic syndrome (MDS) (NCT02273102) and patients with relapsed or refractory AML (NCT02261779) in non-acute promyelocytic leukemia (APL) AMLs when treatment with all-trans-retinoic acid 37 is not effective.
The topoisomerase IIa (TopIIa) inhibitor etoposide 42 combined with the Enhancer of Zeste Homologous 2 (EZH2) inhibitors 7-deazaneplanocin A (DZNep) 43 or GSK126 44 induces cell death in murine and human prostate cancer cell lines and showed therapeutic efficacy in vivo. Thus, the combination of a low dose TopIIa inhibitor with a EZH2 inhibitor is beneficial against aggressive prostate cancer . Likewise, EZH2 inhibition sensitizes transcription activator BRG1 (ATP-dependent helicase SMARCA4) and EGFR mutant lung tumors to TopoII inhibitors, which suggest that combination therapy is a promising approach to this cancer .
Novel epigenetic modulators continue to reach clinical trials. For example, the first-in-man study of the toxicity, pharmacokinetics, and pharmacodynamics of CG200745 46, a pan-HDAC inhibitor, in patients with refractory solid malignancies was initiated in 2015 . CG200745 46 can be safely administered at effective dose levels that inhibit HDAC in peripheral blood mononuclear cells (PBMCs) and tumor tissue, although maximum tolerated dose (MTD) was not reached .
Combinations of epigenetic drugs
In the investigation of novel treatment options, the simultaneous targeting of multiple epigenetic systems, notably when HDACi and DNMTi are administered together, aims to achieve efficient epigenetic gene reactivation (http://clinicaltrials.gov/). For example, results of the phase I/II trial of combined epigenetic therapy with DNMTi azacitidine 18 and HDACi entinostat 17 in extensively pretreated patients with recurrent metastatic NSCL are encouraging . The combination of vorinostat 1 and cladribine (2-chlorodeoxyadenosine) 47 synergistically induced apoptosis in natural killer cell large granular lymphocytes (NK-LGL) leukemia . Cladribine 47 is a drug approved for the treatment of hairy-cell leukemia and acts as indirect DNMTi, since it inhibits SAH hydrolase, increasing competition of SAH for the SAM binding site.
The combination of HDACi (and also other chromatin remodeling enzyme inhibitors such as DNMTi) with the lysine methyltransferase inhibitor DZNep 43 revealed the importance of pharmacological combinatorial approaches in breast cancer cells and in the regulation of cancer immunity . Also encouraging are the results of triple combination using HDACi (TSA 48), DNMTi (5-AZA-CdR, 19), and EZH2 inhibitor (DZNep, 43) on human AML cells . The triple combination (which proved to be more effective than the combination of two agents or a single agent) induced a remarkable synergistic antineoplastic effect as demonstrated by an in vitro colony assay and also showed a potent synergistic activation of several key tumor suppressor geners (TSGs) as determined by real-time PCR.
The combination of vorinostat 1 and the LSD1 inhibitor tranylcypromine 40 was able to reduce glioblastoma stem cell viability and displayed efficacy in a U87 xenograft model .
BET inhibitors are also promising therapeutic agents [47, 49, 83], although resistance has been documented [84, 85]. Their efficacy might be explained by the chromosomal translocations involving bromodomains BRD3 and BRD4 occurring in NUT midline carcinoma (NMC) and in AML [86, 87]. BET inhibition led to promising results in mouse models of sepsis , autoimmunity (in combination with a Myc inhibitor) , and inflammation of the lung . As an example, JQ-1 49  prevented tumor progression by promoting differentiation in murine NMC  and also cardiac hypertrophy in mice .
The combined inhibition of BET family proteins and HDAC has been considered as a potential epigenetics-based therapy for the treatment of pancreatic ductal adenocarcinoma . Clinical trials have also been initiated for the treatment of relapsed or refractory neuroblastoma (NCT02337309) and dose escalation studies for intravenous infusions in patients with other solid cancers are underway (NCT00907205).
A combination of BET inhibitors (JQ-1 49) and SIRT activators (SRT1720, 51) was found to alleviate inflammatory response due to the upregulation of SIRT1 by the BETi JQ-1 49, thus reversing the pro-inflammatory response to SIRT1 inhibition in a cellular lung disease model . On the other hand, the combination of JQ-1 49 with gamma-secretase inhibitors was shown to be effective against primary human leukemias in vivo .
Dual acting hybrids with an epigenetic and a second mechanism of action
Epigenetic therapies are coming of age, and seven drugs have been approved for cancer, with many more undergoing clinical trials. Advances in genome-wide analyses and bioinformatics are providing information on the disease-supportive and disease-irrelevant gene networks that are deregulated by aberrant epigenetic modifications. Features important for epigenetic therapies are well recognized: (a) epigenetic deregulation causes both gene specific and systemic effects; (b) crosstalk and complex formation occur between epigenetic modifiers, which implies that multiple epigenetic systems are likely to be affected [97, 98]; (c) genetic instability of cancer cells has the most likely altered multiple epigenetic systems at the time a patient is diagnosed; (d) the observations that existing epigenetic drugs affect normal cells less than cancer cells indicated either a higher epigenetic plasticity of normal cells or a particular sensitivity of tumor cells to certain epigenetic drug activities; (e) epigenetic drugs are in principle non-genotoxic and their action can be made reversible upon discontinuation of the treatment; (f) as for all drugs, the development of resistance to a single agent is a concern; (g) certain epigenetic drugs can “override” not only their cognate target but also the silencing caused by other epigenetic enzymes .
Together, the above aspects provide a rationale for the combination of pharmacophores, one of them targeting the epigenetic enzymatic machinery, and also for the simultaneous targeting of multiple epigenetic systems. Most of the epigenetic drugs developed following the first strategy are hybrid molecules containing the scaffold of an HDACi fused/linked to another anticancer drug, cytotoxic agent, anti-angiogenesis drug, etc., acting at a related target. There are two principal reasons for the popularity of HDACs in the design of dual targeting agents. Firstly, HDACs are the epigenetic targets that have received the most attention for drug discovery, and consequently, there are a multitude of high affinity inhibitors known with diverse chemical scaffolds. Secondly, the HDAC pharmacophore, backed up by X-ray cocrystal structures of enzyme-inhibitor complexes, includes a “cap” region that is protruding from the active site channel and engaged in binding interactions with the enzyme surface. While these are important, they are less dominant in driving potency compared to the coordination to the active site zinc cation. As a result, the surface-binding cap is tolerant of a high degree of structural variation without compromising HDAC binding. It is hence possible to incorporate a cap that contains the pharmacophore for a second non-HDAC target and the resulting chimeric molecule is capable of binding to both these targets.
The most popular strategy for a dual action HDAC inhibitor consists of taking a known pharmacophore for a second target and grafting a side-chain containing a spacer and a zinc-binding group. This has been extensively studied with heterocyclic scaffolds that are protein kinase inhibitors. Part of the rationale comes from the synergy observed with kinase and HDAC inhibition in vitro and in vivo models that has spurred clinical trials featuring combination therapy as discussed above (“Combinations of epigenetic drugs”). Furthermore, just like HDAC inhibitors, kinase inhibitors often contain regions that are involved in improving pharmacokinetics rather than bonding interactions with the enzyme active site and are amenable to modification. Since resistance is a major issue with kinase inhibitors in vivo, the addition of an independent mechanism of action may help circumvent this problem.
The above examples illustrate the dual action against HDACs and a non-epigenetic enzyme. In the same way, it is possible to design HDAC inhibitors that are ligands for a non-epigenetic receptor. For example, the membrane protein smoothened is part of the Wnt/β-catenin signaling pathway and is targeted by the recently approved antagonist vismodegib. The hybrid molecule 63 is nanomolar in binding to the protein and inhibiting HDAC1, 2, 3 and 6 . In cell-based assays, levels of acetylated histones and tubulin was increased whereas Gli-2 and Hedgehog signaling was decreased, supporting dual target engagement. A number of ligands for the nuclear hormone superfamily have been successfully modified to be dual HDAC inhibitory agents. For example, 64 is a submicromolar inhibitor of HDAC1 and the estrogen receptor and inhibited the MCF7 cell line with an IC50 of 5 μM . The level of activity against HDACs is rather surprising as the compound contains a carboxylic acid rather than the usual hydroxamic acid as the zinc-binding group. Other groups have reported ligands for the vitamin D  retinoid X  and androgen  receptor that also inhibit HDACs.
Multitarget epigenetic modulators
Polypharmacology, rather than a highly specific “magic bullet,” is the norm for small molecule drugs. A recent survey, for example, found that over 40 % of drugs according to the Anatomical Therapeutic Chemical (ATC) classification had a reported IC50 < 10 μM for six or more targets . While such promiscuity is usually discovered serendipitiously, it can also be deliberately incorporated. In this review, we have described the two major ways in which this has been achieved within the relatively new area of epigenetic drug discovery. The first is through combination therapy using two independent and relatively selective drugs. At the present time, there are ongoing clinical trials that are combining either an epigenetic and a non-epigenetic drug or two epigenetic drugs with distinct mechanisms of action. In many cases, there is in vitro and in vivo evidence from animal models that such combinations have a synergistic effect. Furthermore, they may help widen the scope of epigenetic drugs beyond the narrow spectrum of hematological cancers for which they are currently approved. The second approach, which is more radical, involves the rational design of a new entity that exerts its biological activity through two or more pathways. In epigenetics, this has been highly successful with HDAC inhibitors due to their simple and tolerant pharmacophore. The literature abounds with examples of multitarget HDAC inhibitors, and in two cases from the company Curis, both linked with dual kinase inhibition, the compounds have completed phase I clinical trials. As our understanding of epigenetic targets and their biological relevance deepens, further progress with epigenetic polypharmacology will certainly be accomplished that directly benefits patients in the clinic.
This work is the result of collaboration within COST Action CM1406 “Epigenetic Chemical Biology” (EPICHEMBIO; www.epichembio.eu). We also thank MINECO (SAF2013-48397-R-FEDER), Xunta de Galicia (Grant 08CSA052383PR from DXI+D+i; Consolidación 2013/007 from DXPCTSUG; INBIOMED-FEDER “Unha maneira de facer Europa”).
Availability of data and materials
Both authors have discussed the outline and content and wrote and approved the manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
- 10.Gediya LK, Khandelwal A, Patel J, Belosay A, Sabnis G, Mehta J, Purushottamachar P, Njar VCO. Design, synthesis, and evaluation of novel mutual prodrugs (hybrid drugs) of all-trans-retinoic acid and histone deacetylase inhibitors with enhanced anticancer activities in breast and prostate cancer cells in vitro. J Med Chem. 2008;51(13):3895–904.PubMedCrossRefGoogle Scholar
- 37.Smith BC, Denu JM. Chemical mechanisms of histone lysine and arginine modifications. Biochim Biophys Acta - Gene Regulatory Mechanisms. 2009;1789:45–57.Google Scholar
- 52.Zwergel C, Stazi G, Valente S, Mai A. Histone deacetylase inhibitors: updated studies in various epigenetic-related diseases. J Clin Epigenetics. 2016;2:1–15.Google Scholar
- 57.Luu TH, Morgan RJ, Leong L, Lim D, McNamara M, Portnow J, Frankel P, Smith DD, Doroshow JH, Gandara DR, et al. A phase II trial of vorinostat (suberoylanilide hydroxamic acid) in metastatic breast cancer: a California cancer consortium study. Clin Cancer Res. 2008;14(21):7138–42.PubMedPubMedCentralCrossRefGoogle Scholar
- 59.Park MA, Mitchell C, Zhang G, Yacoub A, Allegood J, Häussinger D, Reinehr R, Larner A, Spiegel S, Fisher PB, et al. Vorinostat and sorafenib increase CD95 activation in gastrointestinal tumor cells through a Ca2+-de novo ceramide-PP2A-reactive oxygen species‚ dependent signaling pathway. Cancer Res. 2010;70(15):6313–24.PubMedPubMedCentralCrossRefGoogle Scholar
- 68.Munster PN, Thurn KT, Thomas S, Raha P, Lacevic M, Miller A, Melisko M, Ismail-Khan R, Rugo H, Moasser M, et al. A phase II study of the histone deacetylase inhibitor vorinostat combined with tamoxifen for the treatment of patients with hormone therapy-resistant breast cancer. Br J Cancer. 2011;104(12):1828–35.PubMedPubMedCentralCrossRefGoogle Scholar
- 72.Takhar HS, Singhal N, Gowda R, Penniment M, Takhar P, Brown MP. Phase I study evaluating the safety and efficacy of oral panobinostat in combination with radiotherapy or chemoradiotherapy in patients with inoperable stage III non-small-cell lung cancer. Anti-Cancer Drugs. 2015;26(10):1069–77.PubMedCrossRefGoogle Scholar
- 77.K-p K, Park S, Kim J-E, Hong Y, Lee J-L, Bae K-S, Cha H, Kwon S-K, Ro S, Cho J, et al. First-in-human study of the toxicity, pharmacokinetics, and pharmacodynamics of CG200745, a pan-HDAC inhibitor, in patients with refractory solid malignancies. Invest New Drugs. 2015;33(5):1048–57.CrossRefGoogle Scholar
- 78.Juergens RA, Wrangle J, Vendetti FP, Murphy SC, Zhao M, Coleman B, Sebree R, Rodgers K, Hooker CM, Franco N, et al. Combination epigenetic therapy has efficacy in patients with refractory advanced non‚ small cell lung cancer. Cancer Discovery. 2011;1(7):598–607.PubMedPubMedCentralCrossRefGoogle Scholar
- 89.Bandukwala HS, Gagnon J, Togher S, Greenbaum JA, Lamperti ED, Parr NJ, Molesworth AMH, Smithers N, Lee K, Witherington J, et al. Selective inhibition of CD4+ T-cell cytokine production and autoimmunity by BET protein and c-Myc inhibitors. Proc Natl Acad Sci U S A. 2012;109(36):14532–7.PubMedPubMedCentralCrossRefGoogle Scholar
- 93.Mazur PK, Herner A, Mello SS, Wirth M, Hausmann S, Sanchez-Rivera FJ, Lofgren SM, Kuschma T, Hahn SA, Vangala D, et al. Combined inhibition of BET family proteins and histone deacetylases as a potential epigenetics-based therapy for pancreatic ductal adenocarcinoma. Nat Med. 2015;21(10):1163–71.PubMedPubMedCentralCrossRefGoogle Scholar
- 102.Beckers T, Mahboobi S, Sellmer A, Winkler M, Eichhorn E, Pongratz H, Maier T, Ciossek T, Baer T, Kelter G, et al. Chimerically designed HDAC- and tyrosine kinase inhibitors. A series of erlotinib hybrids as dual-selective inhibitors of EGFR, HER2 and histone deacetylases. Med Chem Commun. 2012;3(7):829–35.CrossRefGoogle Scholar
- 104.Mahboobi S, Dove S, Sellmer A, Winkler M, Eichhorn E, Pongratz H, Ciossek T, Baer T, Maier T, Beckers T. Design of chimeric histone deacetylase- and tyrosine kinase-inhibitors: a series of imatinib hybrides as potent inhibitors of wild-type and mutant BCR-ABL, PDGF-Rbeta, and histone deacetylases. J Med Chem. 2009;52:2265–79.PubMedCrossRefGoogle Scholar
- 105.Mahboobi S, Sellmer A, Winkler M, Eichhorn E, Pongratz H, Ciossek T, Baer T, Maier T, Beckers T. Novel chimeric histone deacetylase inhibitors: a series of lapatinib hybrides as potent inhibitors of epidermal growth factor receptor (EGFR), human epidermal growth factor receptor 2 (HER2), and histone deacetylase activity. J Med Chem. 2010;53(24):8546–55.PubMedCrossRefGoogle Scholar
- 108.Galloway TJ, Wirth LJ, Colevas AD, Gilbert J, Bauman JE, Saba NF, Raben D, Mehra R, Ma AW, Atoyan R, et al. A phase I study of CUDC-101, a multitarget inhibitor of HDACs, EGFR, and HER2, in combination with chemoradiation in patients with head and neck squamous cell carcinoma. Clin Cancer Res. 2015;21(7):1566–73.PubMedCrossRefGoogle Scholar
- 109.Cai X, Zhai H-X, Wang J, Forrester J, Qu H, Yin L, Lai C-J, Bao R, Qian C. Discovery of 7-(4-(3-ethynylphenylamino)-7-methoxyquinazolin-6-yloxy)-N-hydroxyheptanamide (CUDC-101) as a potent multi-acting HDAC, EGFR, and HER2 inhibitor for the treatment of cancer. J Med Chem. 2010;53:2000–9.PubMedCrossRefGoogle Scholar
- 110.Qian C, Lai C-J, Bao R, Wang D-G, Wang J, Xu G-X, Atoyan R, Qu H, Yin L, Samson M, et al. Cancer network disruption by a single molecule inhibitor targeting both histone deacetylase activity and phosphatidylinositol 3-kinase signaling. Clin Cancer Res. 2012;18(15):4104–13.PubMedCrossRefGoogle Scholar
- 111.Cuadrado MDMT, Franco RF, García AMO, Oyarzabal JS, Rabal MOG. World patent 2014, WO131855.Google Scholar
- 121.Mai A, Cheng D, Bedford MT, Valente S, Nebbioso A, Perrone A, Brosch G, Sbardella G, De Bellis F, Miceli M, et al. Epigenetic multiple ligands: mixed histone/protein methyltransferase, acetyltransferase, and class III deacetylase (sirtuin) inhibitors. J Med Chem. 2008;51(7):2279–90.PubMedCrossRefGoogle Scholar
- 122.Pereira R, Benedetti R, Pérez-Rodríguez S, Nebbioso A, García-Rodríguez J, Carafa V, Stuhldreier M, Conte M, Rodríguez-Barrios F, Stunnenberg HG, et al. Indole-derived psammaplin A analogues as epigenetic modulators with multiple inhibitory activities. J Med Chem. 2012;55:9467–91.PubMedCrossRefGoogle Scholar
- 123.Rotili D, Tomassi S, Conte M, Benedetti R, Tortorici M, Ciossani G, Valente S, Marrocco B, Labella D, Novellino E, et al. Pan-histone demethylase inhibitors simultaneously targeting jumonji C and lysine-specific demethylases display high anticancer activities. J Med Chem. 2014;57(1):42–55.PubMedCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.