Polypharmacology: Challenges and Opportunities in Drug Discovery


Polypharmacology: Challenges and Opportunities in Drug Discovery...

1 downloads 129 Views 2MB Size

Subscriber access provided by Baylor University Libraries

Perspective

Polypharmacology: Challenges and Opportunities in Drug Discovery Andrew Anighoro, Jürgen Bajorath, and Giulio Rastelli J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 19 Jun 2014 Downloaded from http://pubs.acs.org on June 20, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Medicinal Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Perspective

Polypharmacology: Challenges and Opportunities in Drug Discovery

Andrew Anighoro1, Jürgen Bajorath2, and Giulio Rastelli1*

1

Life Sciences Department, University of Modena and Reggio Emilia. Via Campi 183, 41125 Modena (Italy)

2

Department of Life Science Informatics, B-IT, LIMES Program Unit Chemical Biology and Medicinal Chemistry, Rheinische Friedrich-Wilhelms-Universität, Dahlmannstr. 2, D-53113 Bonn, Germany.

Abstract

At present, the legendary magic bullet, i.e. a drug with high potency and selectivity towards a specific biological target, shares the spotlight with an emerging and alternative polypharmacology approach. Polypharmacology suggests that more effective drugs can be developed by specifically modulating multiple targets. It is generally thought that complex diseases such as cancer and central nervous system diseases may require complex therapeutic approaches. In this respect, a drug that “hits” multiple sensitive nodes belonging to a network of interacting targets offers the potential for higher efficacy, and may limit drawbacks generally arising from the use of a single-target drug or a combination of multiple drugs. In this perspective, we will compare advantages and disadvantages of multi-target versus combination therapies, discuss potential drug promiscuity arising from off-target effects, comment on drug repurposing, and introduce approaches to the computational design of multi-target drugs.

ACS Paragon Plus Environment

1

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 42

Introduction

Modern drug discovery has been strongly focused on the development of drugs intended to act against a specific target with high potency and selectivity. This paradigm is based on a direct cause-effect relationship between the activity of a gene product and a particular phenotype. Consequently, a pharmacological agent able to specifically modulate the activity of a deregulated protein should be able to revert a pathological phenotype. For instance, in the case of an aberrant cellular pathway caused by an augmented activity or over-expression of an enzyme, a potent and selective inhibitor would be able to restore normal cellular functions. Likewise, a multi-target (i.e. promiscuous) drug with a wider and sometimes unpredictable spectrum of biological activities could eventually lead to adverse reactions, making its use unsafe. It is now generally recognized that these concepts may be too simplistic to explain the mechanism of action of some drugs as well as to design therapies for complex multifactorial diseases. Nowadays, designing a single drug molecule able to simultaneously and specifically interact with multiple targets is gaining major consideration in drug discovery.1-4 In this Perspective, we will refer to this as “polypharmacology”, to distinguish the approach from combination therapy in which two or more drugs are typically used in combination. Furthermore, it should also be noted that polypharmacology is not synonymous with “compound promiscuity”. The latter term refers to the presence of specific interactions of a compound with multiple targets and hence represents the molecular basis for polypharmacology including associated functional effects.5 Importantly, compound promiscuity as the basis of polypharmacology does not include non-specific binding events due to compound liabilities. The number of articles and reviews published in the last ten years containing the term polypharmacology in the title, abstract or keywords (data retrieved from Scopus6) has steadily increased. Out of 194 papers, 100 were published in the last two years and more than 60 last year. These data witness an increasing interest in this approach. In a recent review by Jalencas and Mestres, ACS Paragon Plus Environment

2

Page 3 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

polypharmacology was analyzed from an evolutionary point of view, suggesting that the ability of early biological organisms to adapt to environmental changes might have favored the selection of biomolecular systems intrinsically prone to chemical promiscuity.7 The authors speculated that the level of polypharmacology observed in current drugs may well just be a latent signature of evolution. In general, understanding polypharmacology and its origins is an important condition for improving our ability to rationally exploit polypharmacological effects for therapeutic purposes. A drug active on multiple targets may be characterized by an improved efficacy when compared to a highly selective pharmacological agent. In fact, multi-target activities may potentiate efficacy – either additively or synergistically - and be less prone to the insurgence of drug resistance mutations. Furthermore, the inherent redundancy of biological networks, especially within the same protein family, might impede the efforts of shutting down a cellular pathway if a single switch is turned off. Several lines of evidence indicate that complex pathologies are often polygenic in nature and tend to involve the deregulation of complex and extended networks of proteins. Such diseases are unlikely to be successfully treated by pharmacological interventions based on a single target, while the modulation of an optimal array of targets may provide a more efficient strategy.8 Notably, modulation of multiple targets may result in a synergistic effect. While this goal is commonly pursued in combination therapies by administrating two or more drugs, multi-target drugs aim at the same goal but using a single drug molecule. Both approaches have strengths and limitations, as it will be further described later on. The analysis of known drug-target associations results in complex networks that were explored in several works.9-13 It is by now generally recognized that several approved drugs elicit their therapeutic effect through complex polypharmacology.14 Importantly, in most cases, such behavior was discovered only retrospectively2 and the complete set of current drug-target associations is still far from being completely characterized. For example, addressing the issue of drugs with multiple activities, a recent analysis15 identified through mining of public molecular databases a set of recurrent molecular scaffolds15 describing small molecules with a tendency for promiscuous activity across different target ACS Paragon Plus Environment

3

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 42

families (Figure 1). Interestingly, such chemotypes (defined here as sets of topologically equivalent scaffolds yielding the same carbon skeleton) were frequently found in approved drugs, suggesting the possibility for unpredicted off-target activities. Specifically, promiscuous scaffolds represented ~6% of all compounds active against current targets. Thirty nine of these scaffolds were found in ~17% of approved drugs and the seven most promiscuous chemotypes alone covered 190 drugs.15 Hence, these drugs represent excellent candidates for experimental polypharmacological profiling. Similar analyses of compound promiscuity through data mining have revealed that there is a clear enrichment of promiscuity along the path from bioactive compounds to drugs,16 for reasons that are currently unknown. For example, while current active compounds from medicinal chemistry sources with available high-confidence activity data are known to bind, on average, to only one to two targets, approved drugs have, on average, close to six known targets.16 Of course, approved drugs are typically more extensively characterized than other active compounds. However, one might also speculate that there could be a selection and enrichment process for promiscuous drug candidates during clinical trials, consistent with the idea that clinical efficacy might often correlate with polypharmacology. The success of multi-target approaches will depend on several challenges, most of which still have to be faced. Rational design of multi-target compounds is still in its infancy and surely will need further implementation and methodological development. Areas of high interest include drug repositioning in different therapeutic areas, prediction of off-target toxicities, and rational design of multi-target drugs. From a medicinal chemistry point of view, both the rational identification of multi-target hit or lead compounds as well as the optimization of complex structure-activity relationship (SAR) profiles are challenging tasks. To this end, several approaches have recently been proposed, and more efforts in this direction will be necessary in the near future. In this perspective we will concisely review and discuss major challenges and opportunities of polypharmacology in drug discovery. In particular, we will focus on pros and cons of multi-target vs. combination therapies, potential drug promiscuity arising from off-target effects, drug repurposing, and ACS Paragon Plus Environment

4

Page 5 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

rational design of multi-target drugs.

Combination therapy vs. multi-target drugs

Single-target therapeutic agents may be affected by side effects and tissue toxicity, resulting in reduced efficacy, drug resistance, and a generally decreased quality of life for patients. In contrast, specific drug combinations are designed to achieve more durable disease control, resulting from simultaneous blockade of disease-relevant targets in properly selected patients. In general, combination drugs should be individually active, yet elicit synergistic effects, have different mechanisms of action, nonoverlapping mechanisms of resistance, and distinct toxicities. There are at least three ways to select an agent for testing in combination with another drug: the drug may act on the same target, an additional target of the same pathway, or a different pathway or cellular process that is involved in the pathogenesis or drug resistance of a specific disease type. For example, the molecular and genetic complexity of advanced-stage diseases such as cancer suggests that targeting a single oncogenic pathway may not be sufficient to achieve durable remissions in patients.17 Accordingly, novel drug discovery and development strategies are focusing on targeting multiple signaling pathways, either with drug combinations or through the design and development of a single compound able to target multiple oncoproteins. A few examples of currently investigated combination therapies in different disease areas are reported in the following. For example, combinations of targeted agents with the proteasome inhibitor bortezomib are currently evaluated in cancer clinical trials. Drug combinations include kinase, heat shock protein 90 (Hsp90), and farnesyltransferase inhibitors, as well as other targeted drugs. Preliminary clinical data suggest that some of these combinations have promising clinical efficacy.18 Hsp90 inhibitors19 act additively or synergistically with many other drugs in the treatment of both solid tumors and leukemias. For this reason, several clinical trials in which targeted drugs are combined with ACS Paragon Plus Environment

5

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 42

Hsp90 inhibitors are currently underway.20 For example, drug combinations involving the Her2/Erb2B inhibitor trastuzumab (Herceptin™) and the Hsp90 inhibitor 17-AAG have been explored in clinical trials.21 The implementation of artemisinin-based combination therapies resulted in a great success for treating chloroquine- and sulphadoxine-pyrimethamine-resistant malaria, such that the WHO recommends these drug combinations as the preferred first-line antimalarials against P. falciparum malaria.22 Interestingly, the mechanism by which artemisinin derivatives exert antimalarial activity is still far from being fully elucidated, and several lines of evidence suggest that these molecules modulate the activity of multiple targets. Combination therapy is also explored in multiple sclerosis (MS), a complex central nervous system (CNS) inflammatory disease process that comprises multiple potential therapeutic targets.23 For example, one drug can be used to reduce inflammation and another to augment neuroprotection or repair, or to reduce side effects. However, the advantages of combination therapy in MS are still a matter of debate, as adverse effects of individual drugs can be additive or even synergistic, and drugs might interfere with each other, thereby reducing efficacy. Despite the highly significant therapeutic relevance of combination therapies, potential advantages of a targeted therapy based on a single drug that modulates the activity of multiple targets over singletargeted or combination therapy can be prospected as follows: i) a molecule that acts simultaneously on multiple targets often shows superior efficacy against advanced-stage diseases, compared to compounds with high specificity for a single target; ii) a single molecule with dual activity may have a more predictable pharmacokinetic profile compared to multiple molecules administered in combination. Therefore, it may display a superior pharmacokinetic (PK) and safety profile. This is particularly true if adverse effects (AEs) are moleculebased, because a single molecule would be administered instead of two. On the other hand, no clear advantage may be obtained if AEs are target-based, because having two mechanisms of action may ACS Paragon Plus Environment

6

Page 7 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

create distinct AEs; iii) experience is showing that acute and delayed toxicities may be higher in drug combinations, especially when combining drugs that are not particularly selective. This suggests that modifications in the doses/schedule that are applied when the agents are used individually are generally necessary, and that phase 1 and 2 trial designs may underestimate the toxicity of a combination; iv) the combination of multiple active principles, as typically performed in combination therapies, may result in positive but also negative synergistic effects, thus limiting the number of useful combinations; v) the probability of developing target-based resistance to multi-target drugs is statistically lower than is the probability of developing resistance against single-target drugs; vi) administering a single compound having multiple biological targets guarantees the simultaneous presence of the molecule in tissues where the active principle is required to work such that the compounds interacts with its multiple targets; vii) in some cases, the risks of drug-drug interactions associated with combination therapies could be mitigated, resulting in a simplified therapeutic regimen; viii) the use of drug cocktails often complicates dosing schedules and negatively impacts patient compliance; ix) a single agent binding to multiple targets might be easier to develop given that the regulatory requirements for demonstrating activity and safety of a combination are more arduous than for a single agent. Regulatory agencies generally require the demonstration of safety of the individual drugs prior to human testing of their combination, which delays the assessment of novel combinations until after the two agents have been assessed as single drugs and in combination with standard therapies. These combination trials may also highlight novel toxicities and adverse effects. In addition to the principal considerations above, problems may also arise when two drugs that are candidates for combination therapy are being developed by different pharma companies, because regulatory and intellectual property issues may delay clinical trials of these investigational drug ACS Paragon Plus Environment

7

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 42

combinations. To facilitate such trials the Cancer Therapy Evaluation Program (CTEP)24 at the National Cancer Institute (NCI) has developed data-sharing and patent rights language, which has prompted collaboration between cancer research centers, hospitals and companies to enter into agreements to develop and test novel combinations.25,26 For all these reasons, in the future, the development of single-molecule drugs with a desired multitarget profile may offer an appealing and cost-effective complement or alternative to drug combinations. At present, combination therapy is more extensively explored in the clinic than the use of multi-target drugs. Both approaches are potentially able to yield drugs or drug combinations with improved safety and efficacy profiles. In either case, key to success will likely be the choice of suitable biological targets and molecular pathways that need to be concurrently modulated with drug molecule(s).

Drug promiscuity and off-target effects

The balance between a “beneficial polypharmacology” and a “harmful promiscuity” of drugs is an issue that needs to be carefully evaluated in drug discovery. Could multi-target ligands be inherently more prone to side effects and adverse reactions? Adverse reactions are a major concern for drug safety and frequently determine attrition,27 ultimately resulting in the withdrawal of a drug from late-stage clinical trials or the market. Adverse reactions can be mediated by the primary target of the drug or by promiscuous interactions of the drug with detrimental off-targets (also called anti-targets).28 Challenging situations arise when primary and anti-targets share significant homology, because it will be harder to achieve the desired selectivity in cases where primary and anti-targets are closely related. For example, while carbonic anhydrase (CA) IX is considered a relevant target for cancer therapy, CA II inhibition may cause a series of relevant adverse reactions such as constipation, diarrhea, eye irritation, watering, blurred vision and taste changes.29 These two CA isoforms share 31% sequence identity and their three-dimensional structure is very similar. In this case, a difference of only three ACS Paragon Plus Environment

8

Page 9 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

residues in the active site has been identified as a key feature exploitable for selectivity. Usually, the risk of activity against dangerous off-targets is explored and partially managed by screening drug candidates against a panel of relevant proteins that are known to mediate side effects in humans. However, it is important to note that the size of the human proteome and the only partial understanding of its functional spectrum do currently not permit complete, or even nearly complete, off-target screening. Moreover, many post-translational modifications of proteins have profound and often tissue-specific functional consequences that can hardly be accounted for in typical screening settings. Taking these difficulties into consideration, screening performed against a small panel of representative targets has been suggested as an affordable way to identify dangerous promiscuity of early stage drug candidates.30 To address this issue, different predictive models have also been proposed. For example, Kuhn et al. integrated known drug-target and drug-side effect data to predict new target-side effects relations emerging from statistically significant correlations.31 They were able to predict and validate in vivo that zolmitriptan (Figure 2), a drug used to treat migraine targeting serotonin receptors 5-HT-1B and 5-HT-1D, is responsible for increased pain sensitivity (hyperesthesia) as a consequence of binding to the off-target 5-HT-7 (26% sequence identity with 5-HT-1B and 5-HT1D). Lounkine et al. applied a computational strategy based on the evaluation of the similarity of several drugs to sets of reference ligands32 of targets with previously known associations to adverse reactions.33,34 They were able to predict drug-target interactions confirmed by database annotations or binding assays. Furthermore, they associated in vitro activities with adverse reactions in patients by obtaining drug-target-adverse reaction links. For example, the histamine H1 receptor was identified as an off-target of prenylamine (Figure 2), a calcium channel blocker.35 This activity was associated with sedation, a reported side effect of this drug. Interestingly, this side effect was not associated with any of the previously known targets of prenylamine, illustrating that ligand-based approaches can be used to discover off-target activities of known drugs and rationalize observed adverse reactions, at least in cases where substantial small molecule ligand information is available for therapeutic targets. ACS Paragon Plus Environment

9

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 42

Based on these considerations, one would conclude that beneficial polypharmacology approaches should be directed towards the most suitable and promising target combinations, avoiding promiscuous effects likely arising from interaction with harmful anti-targets. Thus, prior knowledge inevitably play an important role in devising meaningful polypharmacology approaches. Regardless, the ability to predict off-target interactions and their effect on patients would have a strong impact on drug development, saving time and resources and navigating more efficacious and potentially safer drugs through the discovery pipeline. It is likely that the development of methods for predicting off-target activities will be one of the future key tasks for polypharmacology.

Drug repurposing

It has been recognized that drugs used in the treatment of a given pathology may sometimes be “rediscovered” in different therapeutic areas.36,37 This event may take place if the primary target of a drug is involved in more than one pathology, or if the drug is able to modulate the activity of additional targets that are relevant for the new application. Of course, one would hope that predictive approaches might also be applicable to help narrow down potential secondary applications. A perspective on unpredicted off-target effects was given by Keiser et al., who proposed a statistical ligand-based approach to predict new off-targets for approved drugs and relevant pharmaceutical compounds.36 With their approach they were able to capture similarities between ligands of unrelated proteins and to identify previously unreported polypharmacologies. In recent years, pharmaceutical companies in particular were attracted by drug repositioning, because the possibility to use an approved or investigational drug in a new therapeutic area avoids the expensive and time-consuming pharmacokinetic and toxicological profiling typically required for new candidates. Furthermore, drugs not approved for use in a given therapeutic area because of adverse reactions may be “rescued” if the benefits obtained in a new therapeutic area justify their administration. Therefore, drug repurposing and ACS Paragon Plus Environment

10

Page 11 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

rescuing can be regarded as “shortcuts” in the drug discovery process that can potentially lead to major savings in terms of time and economical investments. However, it has proven difficult in pharmaceutical R&D to implement widely applicable and robust strategies for drug repurposing. Given that many drugs are active against multiple targets, repurposing might frequently be feasible. However, how to best discern polypharmacology patterns and differentiate between target profiles that are relevant for complementary or alternative therapeutic applications are as of yet unsolved problems. Hence, successful drug repositioning campaigns currently are exceptions, rather than the rule, despite the conceptual attractiveness of the approach. Famous examples of successfully repositioned drugs are represented by sildenafil from angina to sexual dysfunctions in men,38 thalidomide, a drug rescued from morning sickness to leprosy and multiple myeloma,40 and methotrexate from cancer to psoriasis and rheumatoid arthritis (Figure 2).40 In the literature, several studies have put forward new candidates for repositioning. Nygren et al. proposed mebendazole (MBZ, Figure 2), an antihelmintic drug, as a candidate for repositioning in advanced colon cancer treatment.41 MBZ was identified as a hit in a phenotypic screening of 1,600 drugs against two colon cancer cell lines. The drug proved active against relevant kinases such as ABL and BRAF. This compound was also used to treat a patient with refractory metastatic colon cancer, who showed signs of complete or good partial remission of metastases in lungs, lymph nodes, and liver. In a recent review, Corbett et al. described compounds collected from different studies that are currently investigated for a potential repositioning in Alzheimer's disease (AD).42 Interestingly, these compounds originated from very different pathologies and therapeutic applications such as type 2 diabetes mellitus, hypertension, antibiotics and retinoid therapy. From a methodological point of view, computational methods are now being actively investigated as tools for identifying candidates for drug repurposing. The speed and scalability of these approaches may have a strong impact on driving experiments on new targets and further reducing the costs involved in this process. Due to the potential advantages offered by drug repositioning, computational ACS Paragon Plus Environment

11

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 42

methods able to prioritize drug repurposing experiments will surely play an increasingly important role, despite the inherent complexity of the underlying polypharmacology. In addition, experimental approaches such phenotypic screening, which currently experiences a renaissance in drug discovery, are being investigated for repositioning. However, similar to computational approaches, these efforts are also still in their infancy. One would hope that integration of experimental and computational target annotation approaches might ultimately lead to more general drug repositioning strategies.

Polypharmacology in cancer therapy

Polypharmacology is mostly relevant for diseases involving wide target networks and cellular pathways. A prime example is provided by cancer. Typically, cancer cells are characterized by a transformed phenotype with excessive proliferation and survival. This abnormal cellular activity can be sustained by deregulation in the expression or activity of different proteins. For example, protein kinases belong to a family of over 500 members in humans and are involved in multiple cellular pathways and networks. Traditionally, the medicinal chemistry gold standard in kinase research was represented by the (putative) development of highly selective inhibitors. For example, many currently known (highly) promiscuous ATP site-directed kinase inhibitors, several of which are marketed as anticancer drugs, were originally thought to be rather specific.2.43 Today, however, it is becoming increasingly clear that more effective targeted therapies based on kinase inhibitors used for cancer treatment should hit the disease on multiple sensitive target nodes, which explains the observed efficacy of marketed kinase inhibitors in the treatment of several forms of cancer. The architecture of the ATP binding site is largely conserved, hence achieving the desired selectivity profile with type I inhibitors has always been a major issue.44-46 In some types of cancers, the deregulation of the activity of a single kinase seems to represent the driving force for the disease, and a selective inhibition can prove successful. Nevertheless, appearance of mutations and redundancy in biological networks can ACS Paragon Plus Environment

12

Page 13 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

compensate the effects of inhibiting one target by over-activating another target or cellular pathway, eventually making a single-target approach ineffective. Again, it is emphasized that kinase inhibitors initially thought to elicit their therapeutic effect by inhibiting a particular kinase actually owe their success to a more complex pattern of activities.2,43 Simultaneous inhibition of multiple kinases is now established as a successful therapeutic strategy. Sunitinib (Figure 2), a promising compound for the therapy of anaplastic thyroid cancer, was tested in vitro and in vivo proving its ability to inhibit Akt and ERK1/2 phosphorylation and to down-regulate cyclin-D1.47 The size and complexity of the human kinome with 518 currently known members makes the identification of kinases well-suited for multitarget approaches rather challenging. Dar et al. addressed this issue with an integrated approach including whole-animal testing, chemistry and genetics.48 They put forward an optimized multiple inhibition profile of Ret, Raf, Src, Tor and S6K leading to lower toxicity and improved survival in animal models. In another study, Xie et al. proposed that the known inhibition of the proliferation of cancer cells displayed by the HIV-protease inhibitor nelfinavir (Figure 2) is due to a weak modulation of multiple kinases upstream of the PI3K/Akt pathway.49 The authors suggested nelfinavir as a promising starting point for the design of new kinase inhibitors. Recently, Hearn et al. revealed potential polypharmacology of organometallic complexes containing Iridium(III), although specific targets have not yet been identified. Iridium complexes displayed selectivity for different cell lines when different chelating ligands were used, suggesting modulation of different targets.50 Going beyond cancer, it should be noted that kinase inhibitor polypharmacology might not always be desirable. For example, for the treatment of a number chronic inflammatory diseases (as well as other chronic diseases) therapeutic requirements fundamentally change and achieving kinase specificity using inhibitors with different mechanisms of action becomes a major issue.51 Another family of relevant drug targets for cancer therapy is represented by poly(ADP-ribose) polymerases (PARPs).52 The activity of these enzymes consists of ADP-ribosylation of target proteins, finally resulting in the regulation of several cellular mechanisms such as DNA repair, protein ACS Paragon Plus Environment

13

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 42

degradation and apoptosis. PARP inhibitors have been in clinical trials for several years, both as single drugs and in combination with cytotoxic agents. Most PARP inhibitors bind to the nicotinamide binding pocket, causing polypharmacology throughout this family of targets. The advantages of selective versus a wider activity profile within PARPs is still a matter of investigation and debate, and further work will be necessary to draw more definitive conclusions.53 Hsp90 has been extensively investigated as a cancer drug target, but none of the ATP-competitive inhibitors19 has so far reached the market, mainly because of toxicity issues. In cancer cells, Hsp90 refolds, stabilizes and regulates the trafficking of more than two hundred client proteins (the Hsp90 interactome), some of which are responsible for uncontrolled proliferation and apoptotic resistance. As a matter of fact, the interest in Hsp90 inhibitors resides mainly in the destabilizing effect of these compounds on oncogenic client proteins relevant for cancer. In addition to the exploration of Hsp90 inhibitors, both alone and in combination with other targeted agents, recent evidences suggest that having a single molecule that simultaneously inhibits Hsp90 and one or more of its client proteins,54 such as kinases, could improve efficacy. Moulick et al. described an affinity-based proteomics approach combined with bioinformatics for the characterization of Hsp90 complexes interacting with specific small molecules in chronic myeloid leukemia.55 They provided evidence that PU-H7155 (Figure 2), a known Hsp90 inhibitor, preferentially targets tumor-enriched Hsp90 complexes and affinity captures Hsp90-dependent oncogenic client proteins. This strategy may be useful for identifying targets for both combination therapies and multi-target inhibitor design. In another study, the discovery of Hsp90/tubulin dual inhibitors has been reported.56 Tubulin is an Hsp90 client protein53 and also a prime cancer drug target.57 A small molecule (MDG89256, Figure 2) able to interact with both targets was discovered through combined ligand- and structure-based virtual screening including docking and pharmacophore modeling.56 Another compound (CDBT58, Figure 2) targeting Hsp90 and tubulin was discovered through phenotypic screening by observing its activity in non-small-cell lung cancer (NSCLC) cells.58 Recently, an ad hoc computational protocol for selecting promising target ACS Paragon Plus Environment

14

Page 15 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

combinations and building focused compound libraries for polypharmacology approaches within the Hsp90 interactome has also been put forward.59 These investigations indicate that the selection of optimal target combinations is a key aspect in polypharmacology research.

Polypharmacology in CNS diseases

Another important area of application of polypharmacology is in the field of CNS diseases. The activity of these drugs is often mediated by the interaction with G-protein coupled receptors (GPCRs). Importantly, progress in the structural biology of GPCRs is providing unprecedented opportunities for determining crystal structures of novel receptors in complex with different ligands,60 thus opening up the way to structure-based drug design investigations.61 It is known that several small molecules bind multiple GPCRs.62 Moreover, taking into consideration that some GPCRs are linked to multiple therapeutic areas, drugs targeting GPCRs constitute privileged starting points for drug repurposing. In addition to kinases, GPCRs are another prominent target family whose functional spectrum is substantially influenced by compound promiscuity and drug polypharmacology. CNS drugs used in the therapy of psychosis or depression are usually characterized by very complex mechanisms of action. Both the antipsychotic activity and the side effects elicited by these drugs can be attributed to a complex pattern of biological activities on multiple receptors. A classic example is clozapine (Figure 2), an atypical antipsychotic with improved efficacy compared to its predecessors. Clozapine is successful in avoiding extra-pyramidal side effects but it is also associated with some dangerous side effects such as seizures and diabetes.62 Many different isoforms of the serotonin, dopamine, muscarinic, and adrenergic receptors are among known targets of clozapine. In particular, histamine H1 receptor is linked to weight gain, and other side effects are caused by activity on adrenergic and muscarinic receptor. On the other hand, dopamine D2 and serotonin 5-HT-2A seem responsible for antipsychotic therapeutic effect. It is worth noting that, besides GPCRs, proteins ACS Paragon Plus Environment

15

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 42

belonging to other families are also involved in the mechanism of CNS disorders and may represent potential drug targets. For instance, cognitive impairment in schizophrenia is hardly treated by currently available antipsychotics. An approach to address the lack of efficacy in treating cognitive impairment has been recently proposed by Lipina et al.,63 who described the synergistic effect associated to schizophrenia obtainable by dual inhibition of phosphodiesterase 4 (PDE4) and glycogen synthase kinase 3 (GSK-3).64 Moreover, they described antipsychotic activity and promising effects on cognitive impairment obtained by administrating such dual-inhibitor in vivo. Neurodegenerative diseases such as Parkinson (PD) and Alzheimer (AD) are actively investigated using multi-target approaches. For instance, ladostigil (Figure 2) showed good efficacy in models of AD. This effect is due to increased levels of acetylcholine, dopamine, noradrenaline, and 5-HT that in turn are determined by the ability of the compound to inhibit acetylcholine esterase (AChE) and the brain monoamine oxidases (MAO) A and B.65 As a matter of fact, choline esterases and MAOs are actively pursued as targets for polypharmacology in neurodegenerative diseases. Bolea et al. synthesized and tested a series of novel compounds targeting AChE, butyrylcholinesterase (BuChE), MAO-A and MAO-B,66 and molecular modeling studies were performed to hypothesize their binding mode. Inhibitory effects on β−amyloid aggregation were observed, suggesting that these compounds are promising candidates for polypharmacology approaches to AD. Additional proteins have been proposed as drug targets in this field. Tahtouh et al. suggested that the neuroprotective effects of compounds based on leucettines such as leucettine L41 (Figure 2), i.e. natural compounds derived from a marine sponge, are responsible for the inhibition of a set of proteins and lipid kinases involved in neurodegeneration.67 Recently, a comprehensive review on multi-target therapies in AD was published by León et al..68 They presented diverse approaches targeting AchE in combination with different enzymes, ion channels, and receptors such as the aspartyl protease BACE-1, MAOs, calcium channels, and the N-methyl-D-aspartate receptor, and described non-cholinergic based drug discovery efforts performed so far. ACS Paragon Plus Environment

16

Page 17 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

The application of polypharmacology in CNS and neurodegenerative diseases is challenging due to the complex mechanisms underlying these pathologies. Relevant targets such as GPCRs are difficult to address using multi-target activity approaches due to the cross-reactivity displayed by several of their ligands against different receptors.62 Moreover, as mentioned above, other proteins besides GPCRs also represent promising CNS drug targets and are currently investigated in connection with GPCRs. Despite this complexity, polypharmacology approaches will surely be increasingly utilized for finding novel medicines in this field.

Rational design of multi-target ligands

Taking into consideration all aspects highlighted so far, the ability to rationally design drugs with a desired multi-target activity profile is becoming increasingly important.69,70 This task, which is challenging without doubt, involves considering structure-activity relationship profiles of molecules interacting with different biological targets, and becomes most challenging when these targets are only distantly related or unrelated, i.e., when they belong to different protein families. At present this remains a critical point for the development of multi-targeted drugs, and explains why such activity profiles are generally sought within families of related proteins such as kinases and GPCRs. Activity landscapes are representations of the relation between structural similarity and biological activity of sets of active compounds.71 This approach can be extended from the analysis of the activity against a single target to multiple targets. Modelling of multi-target activity landscapes may provide useful insights into SAR of ligands active on proteins spanning different families.72 The exploration of such multi-target SARs is only a first step in the rationalization of multi-target profiles. However, multi-target (promiscuity) profiles of compounds can be systematically extracted from available compound activity data,73,74 their distribution across different targets and families can be assessed,73 and relationship between promiscuity profiles and molecular similarity can be explored,73 hence ACS Paragon Plus Environment

17

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 42

providing a knowledge-base to aid in the design of compounds with desired multi-target activity. A comprehensive review focused on rationally designed multi-target ligands was published by Morphy and Rankovic.75 In this article the main strategies and applications in this field as reported in the literature between 1990 and 2004 are described. Over the last years, more methods have been proposed to rationalize the design of multi-target ligands. For example, Apsel et al.76 identified compounds that inhibit both tyrosine kinases and phosphatidylinositol-3-OH kinases. Proteins belonging to these kinase subfamilies are characterized by low overall sequence identity (e.g. 9% between SRC and PI3K p110γ isoform) but share a similar fold (two-lobed structure) and some conserved residue motifs in the ATP site. Both kinase subfamilies include relevant cancer drug targets. On the basis of iterative chemical synthesis, X-ray crystallography and biochemical profiling, they discovered compounds with promising dual-target activity.76 Structural features were also examined in protein/ligand complexes to rationalize the selectivity toward their targets. This work represents one of the first successful rational design attempts of multi-target ligands within the human kinome. More recently, Besnard et al.77 described an automated approach for the design of ligands with desirable polypharmacological profiles. Their method is based on the design of small focused libraries of analogues of an initial compound through automated machine learning of medicinal chemistry design strategies derived from the literature. Bayesian models were built to prioritize the generated compounds according to a set of multiple objectives, such as activity towards different targets or even ADME properties. Eight hundred ligandtarget predictions were tested experimentally, of which 75% were confirmed correct. In particular, starting from the approved drug donepezil, an acetylcholine esterase inhibitor used in the therapy of AD, they predicted and confirmed weak polypharmacological activity on dopamine receptors D4 and D2. Importantly, since D2 antagonism may be beneficial for AD therapy, the authors applied their algorithm to optimize the activity toward this target. In this way, they were able to generate compounds with optimized activity towards D2, enhanced blood brain barrier permeability, and reduced activity towards

ACS Paragon Plus Environment

18

Page 19 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

the α1-adrenoceptor anti-target that can cause low blood pressure as a side effect. In another recent review, computational approaches focused on fragment-based methods have been proposed for the design of multi-target ligands.78 The use of fragments has the advantage of leaving room for optimization of hit compounds and associated increases in molecular weight. Ligand- and structure-based virtual screening continue to be proposed as attractive alternatives to more expensive high-throughput screening. In particular, the generation of libraries enriched with molecular structures focused on intended targets is thought to improve the likelihood of identifying multi-target compounds. In this context, an analysis of activity annotations deposited in public compound databases79 identified compounds active against a single target, multiple targets belonging to the same protein family, and multiple targets belonging to different families (Figure 3).80 These findings highlighted that molecular scaffolds associated with different levels of target selectivity or promiscuity are found in known active compounds, and that database annotations reporting promiscuous compounds are steadily growing over time. Such analyses can provide insights relevant for the rational design of multi-target compounds by exploiting information already available in public databases. A ligand-based approach named Similarity Ensemble Approach (SEA) was introduced by Keiser et al. and successfully applied to identify proteins sharing related ligand sets and discover unknown offtarget activities.81 This approach quantitatively evaluates chemical similarity of two sets of ligands by measuring the Tanimoto coefficient (Tc) of ligand pairs and by applying a statistical model reminiscent the BLAST82 algorithm to normalize similarity scores. Applying SEA the authors were able to identify off-target activities for the drugs emetine and loperamide acting on α2 adrenergic and neurokinin NK2 receptors, respectively.81 This approach was also used for other off-target activity predictions,37,83 for the discovery of new active molecules,84 and for insightful chemoinformatics analyses.85 Recently, Reker et al. put forward an approach, named self-organizing map-based prediction of drug equivalence relationships (SPiDER), designed to profile the targets of de novo designed molecules.86 SPiDER is based on self-organizing maps (SOM)87 combined with a consensus scoring and a statistical analysis ACS Paragon Plus Environment

19

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 42

that evaluates prediction confidence. Two computationally designed HIV-1 protease inhibitors were predicted by SPiDER and experimentally confirmed to be active on three different GPCRs, namely bradykinin B1, neurokinin 1, and vanilloid 1 receptors. 86 Other computational structure-based protocols were proposed for the characterization of polypharmacology of antipsichotic drugs within bioaminergic GPCRs.88,89 In general, molecular docking into crystal structures or homology models of target proteins can provide a valuable strategy for screening libraries of potential ligands active against multiple receptors, despite the approximations inherent in this approach. For example, virtual screening90-92 may be independently performed on two or more biological targets of interest, and multi-target hits may be identified from compounds located at the top of all ranked lists. Having favorable scores for all proteins, these compounds constitute suitable candidates for potent multi-target inhibition. The potential that resides in the automation and transferability of computational methods to different case studies involving diverse targets will be key in the future for the rational design of drugs with an a priori polypharmacological profile. The examples reported in this and previous sections represent a scenario in which rational polypharmacology approaches aiming at the design of multi-target drugs can benefit from the integration of diverse studies as schematically represented in Figure 4.

Summary and outlook

Polypharmacology is emerging as a new paradigm in drug discovery. Approved drugs for which a beneficial polypharmacological profile was discovered only retrospectively provide strong support for the concepts of polypharmacology and multi-target compound design. The next challenges will be how to rationally design ligands with a desired polypharmacology profile and transform them into drug candidates. Molecules with optimal multi-target activities have the potential of improving efficacy and safety of drugs in the therapy of complex diseases. In this perspective we have highlighted that multiACS Paragon Plus Environment

20

Page 21 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

targeted drugs may represent a valuable complement or even alternative to therapeutic regimens based on drug combinations. We have also described relevant examples of polypharmacology-based therapeutic approaches in cancer and CNS diseases. In cancer a major example is represented by kinase inhibitors, while in CNS disorders such as depression and psychosis, GPCRs ligands’ cross-reactivity is the main source of polypharmacology. A general concern is that excessive promiscuity could lead to adverse reactions caused by interactions with anti-targets. Hence, multi-target drug candidates should be designed by trying to optimize activity profiles towards the desired targets, while minimizing the risk of anti-target activity. These requirements emphasize the need to explore multi-target SARs in the practice of medicinal chemistry, which is still far from being routine and will require the development of advanced SAR analysis tools to enable simultaneous SAR exploration and compound design directed at multiple targets. For multi-target approaches, the identification of the most suited combination of targets is challenging but extremely important. Screening of candidate drugs against panels of closely and/or distantly related targets, often referred to as compound profiling, is already performed on a routine basis pharmaceutical research to identify the most promising and safe candidate compounds for given target(s) and eliminate compounds with anti-target liabilities. It is likely that compound profiling efforts will need to be further expanded and standardized to substantially aid in the design of compounds with multi-target activities. Moreover, screening directed towards a more extended characterization of the mechanisms of action of known drugs may have a large impact on the identification of optimal target combinations. Genetic and biochemical characterization of pathology mechanisms will also be of great importance for uncovering the most relevant target combinations. Moreover, screening and analyses of network activities might be increasingly directed toward exploring possibilities of repositioning drugs that modulate targets involved in different pathologies. This approach is likely to be further explored because re-using old drugs is economically convenient (yet more difficult to accomplish than often thought). A major challenge from a medicinal chemistry point of view is the ability to rationally design multi-target ligands, especially when the targets of interest are ACS Paragon Plus Environment

21

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 42

not phylogenetically or structurally related. Here, further methodological advancements will play a major role. To address this issue, dedicated efforts are made through integrated approaches involving medicinal chemistry, genetics, chemical biology, and computational chemistry. Promising computational approaches in this field include data mining and ligand-based analyses for the identification of target combinations, and virtual screening for the design of multi-target ligands. It is conceivable that in the near future rational polypharmacology will play an increasingly important role in drug discovery. The combination of different disciplines and expertise (experimental and computational) will likely be key to success.

Corresponding author *Tel +39 059 2055145, email [email protected]

Notes The authors declare no competing financial interest.

Abbreviations Used: SAR, structure-activity relationship; Hsp90, heat shock protein 90; WHO, world health organization; MS, multiple sclerosis; CNS, central nervous system; PK, pharmacokinetic; AE, adverse effect; NCI, National Cancer Institute; CTEP, Cancer Therapy Evaluation Program; CA, carbonic anhydrase; MBZ, mebendazole; AD, Alzheimer’s disease; PARP, poly(ADP-ribose) polymerase; NSCLC, non-small-cell lung cancer; GPCR, G-protein coupled receptor; PDE4, phosphodiesterase 4; GSK-3, glycogen synthase kinase 3; PD, Parkinson’s disease; AChE, acetylcholine esterase; MAO, monoamine oxidases; BuChE, butyrylcholinesterase; SEA, Similarity Ensemble Approach; Tc, Tanimoto coefficient; SPiDER, self-organizing map-based prediction of drug equivalence relationships; SOM, self-organizing maps;

ACS Paragon Plus Environment

22

Page 23 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Biographies Andrew Anighoro is postdoctoral fellow at the Molecular Modeling & Drug Design laboratory (www.mmddlab.unimore.it) of the University of Modena and Reggio Emilia. He received his Ph.D. in Science and Technologies for Health Products at the University of Modena and Reggio Emilia. During his Ph.D., he was visiting student at the University of Bonn in the laboratory of Prof. Jürgen Bajorath. His research is focused on molecular modelling, structure- and ligand-based methods with a particular interest in multi-target drug design.

Jürgen Bajorath is Professor and Chair of Life Science Informatics at the University of Bonn. He is also an Affiliate Professor in the Department of Biological Structure at the University of Washington, Seattle. His research interests include drug discovery, computer-aided medicinal chemistry and chemical

biology,

and

chemoinformatics.

For

further

details,

please

see:

http://www.lifescienceinformatics.uni-bonn.de

Giulio Rastelli is Professor in medicinal chemistry and head of the Molecular Modeling & Drug Design laboratory (www.mmddlab.unimore.it) of the University of Modena and Reggio Emilia. He received his Ph.D. in medicinal chemistry at the University of Modena and Reggio Emilia and has been research fellow at the University of California San Francisco under the supervision of Proff. Daniel Santi and Peter Kollman. His research interests focus on the development and application of computational drug design methodologies to address problems that lie at the interface between chemistry, biology and medicine. His lab recently developed BEAR, an innovative tool for virtual screening. He collaborates with academic and private institutions for the discovery and development of small-molecule inhibitors of relevant drug targets, with a special focus on cancer.

ACS Paragon Plus Environment

23

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 42

References

(1) Hopkins, A. L. Network Pharmacology: The next Paradigm in Drug Discovery. Nat. Chem. Biol. 2008, 4, 682–690. (2) Boran, A. D. W.; Iyengar, R. Systems Approaches to Polypharmacology and Drug Discovery. Curr. Opin. Drug Discov. Devel. 2010, 13, 297–309. (3) Mestres, J.; Gregori-Puigjané, E. Conciliating Binding Efficiency and Polypharmacology. Trends Pharmacol. Sci. 2009, 30, 470–474. (4) Peters, J.-U. Polypharmacology - Foe or Friend? J. Med. Chem. 2013, 56, 8955–8971. (5) Hu, Y.; Bajorath, J. Compound Promiscuity: What Can We Learn from Current Data? Drug Discov. Today 2013, 18, 644–650. (6) http://www.scopus.com; accessed Dec 2013. (7) Jalencas, X.; Mestres, J. On the Origins of Drug Polypharmacology. Med. Chem. Commun. 2012, 4, 80–87. (8) Lu, J.-J.; Pan, W.; Hu, Y.-J.; Wang, Y.-T. Multi-Target Drugs: The Trend of Drug Research and Development. PLoS ONE 2012, 7, e40262. (9) Adams, J. C.; Keiser, M. J.; Basuino, L.; Chambers, H. F.; Lee, D.-S.; Wiest, O. G.; Babbitt, P. C. A Mapping of Drug Space from the Viewpoint of Small Molecule Metabolism. PLoS Comput. Biol. 2009, 5, e1000474. (10) Tian, L.; Zhang, S. Mapping Drug-Target Interaction Networks. In Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2009. EMBC 2009; 2009; pp. 2336– 2339. (11) Simon, Z.; Peragovics, Á.; Vigh-Smeller, M.; Csukly, G.; Tombor, L.; Yang, Z.; ZahoránszkyKőhalmi, G.; Végner, L.; Jelinek, B.; Hári, P.; Hetényi, C.; Bitter, I.; Czobor, P.; Málnási-Csizmadia,

ACS Paragon Plus Environment

24

Page 25 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

A. Drug Effect Prediction by Polypharmacology-Based Interaction Profiling. J. Chem. Inf. Model. 2012, 52, 134–145. (12) Yildirim, M. A.; Goh, K.-I.; Cusick, M. E.; Barabási, A.-L.; Vidal, M. Drug-Target Network. Nat. Biotechnol. 2007, 25, 1119–1126. (13) Reddy, A. S.; Zhang, S. Polypharmacology: Drug Discovery for the Future. Expert. Rev. Clin. Pharmacol. 2013, 6, 41–47. (14) Rask-Andersen, M.; Almén, M. S.; Schiöth, H. B. Trends in the Exploitation of Novel Drug Targets. Nat. Rev. Drug Discov. 2011, 10, 579–590. (15) Hu, Y.; Bajorath, J. Polypharmacology Directed Compound Data Mining: Identification of Promiscuous Chemotypes with Different Activity Profiles and Comparison to Approved Drugs. J. Chem. Inf. Model. 2010, 50, 2112–2118. (16) Hu, Y.; Bajorath, J. High-Resolution View of Compound Promiscuity. F1000Res 2013, 2, 144. (17) Altieri, D. C. Mitochondrial Hsp90 Chaperones as Novel Molecular Targets in Prostate Cancer. Future Oncol. 2010, 6, 487–489. (18) Wright, J. J. Combination Therapy of Bortezomib with Novel Targeted Agents: An Emerging Treatment Strategy. Clin. Cancer Res. 2010, 16, 4094–4104. (19) Sgobba, M.; Rastelli, G. Structure-Based and in Silico Design of Hsp90 Inhibitors. ChemMedChem 2009, 4, 1399–1409. (20) Lu, X.; Xiao, L.; Wang, L.; Ruden, D. M. Hsp90 Inhibitors and Drug Resistance in Cancer: The Potential Benefits of Combination Therapies of Hsp90 Inhibitors and Other Anti-Cancer Drugs. Biochem. Pharmacol. 2012, 83, 995–1004. (21) Modi, S.; Stopeck, A.; Linden, H.; Solit, D.; Chandarlapaty, S.; Rosen, N.; D’Andrea, G.; Dickler, M.; Moynahan, M. E.; Sugarman, S.; Ma, W.; Patil, S.; Norton, L.; Hannah, A. L.; Hudis, C. HSP90 Inhibition Is Effective in Breast Cancer: A Phase II Trial of Tanespimycin (17-AAG) Plus Trastuzumab

ACS Paragon Plus Environment

25

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 42

in Patients with HER2-Positive Metastatic Breast Cancer Progressing on Trastuzumab. Clin. Cancer Res. 2011, 17, 5132–5139. (22) Eastman, R. T.; Fidock, D. A. Artemisinin-Based Combination Therapies: A Vital Tool in Efforts to Eliminate Malaria. Nat. Rev. Microbiol. 2009, 7, 864–874. (23) Conway, D.; Cohen, J. A. Combination Therapy in Multiple Sclerosis. Lancet Neurol 2010, 9, 299–308. (24) http://ctep.cancer.gov; accessed Dec 2013. (25) Gadgeel, S. M.; Ivy, P.; Chen, W.; Mauer, J.; Smith, D.; Lorusso, P. Analysis of Lung Cancer Patients Enrolled in CTEP (Cancer Therapy Evaluation Program)-Sponsored Phase I Trials. Clin. Lung Cancer 2011, 12, 218–223. (26) Luke, J. J.; Rubinstein, L. V.; Smith, G. L.; Ivy, S. P.; Harris, P. J. Similar Efficacy for Phase I Trials in Comparison with DTIC for Advanced Malignant Melanoma: An Analysis of Melanoma Outcomes in CTEP-Sponsored Phase I Trials 1995-2011. Melanoma Res. 2013, 23, 152–158. (27) Giacomini, K. M.; Krauss, R. M.; Roden, D. M.; Eichelbaum, M.; Hayden, M. R.; Nakamura, Y. When Good Drugs Go Bad. Nature 2007, 446, 975–977. (28) Overall, C. M.; Kleifeld, O. Tumour Microenvironment - Opinion: Validating Matrix Metalloproteinases as Drug Targets and Anti-Targets for Cancer Therapy. Nat. Rev. Cancer 2006, 6, 227–239. (29) Aggarwal, M.; Kondeti, B.; McKenna, R. Insights towards Sulfonamide Drug Specificity in ΑCarbonic Anhydrases. Bioorg. Med. Chem. 2013, 21, 1526–1533. (30) Peters, J.-U.; Hert, J.; Bissantz, C.; Hillebrecht, A.; Gerebtzoff, G.; Bendels, S.; Tillier, F.; Migeon, J.; Fischer, H.; Guba, W.; Kansy, M. Can We Discover Pharmacological Promiscuity Early in the Drug Discovery Process? Drug Discov. Today 2012, 17, 325–335. (31) Kuhn, M.; Al Banchaabouchi, M.; Campillos, M.; Jensen, L. J.; Gross, C.; Gavin, A.-C.; Bork, P. Systematic Identification of Proteins That Elicit Drug Side Effects. Mol. Syst. Biol. 2013, 9, 663.

ACS Paragon Plus Environment

26

Page 27 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

(32) Lounkine, E.; Keiser, M. J.; Whitebread, S.; Mikhailov, D.; Hamon, J.; Jenkins, J. L.; Lavan, P.; Weber, E.; Doak, A. K.; Côté, S.; Shoichet, B. K.; Urban, L. Large-Scale Prediction and Testing of Drug Activity on Side-Effect Targets. Nature 2012, 486, 361–367. (33) Bender, A.; Scheiber, J.; Glick, M.; Davies, J. W.; Azzaoui, K.; Hamon, J.; Urban, L.; Whitebread, S.; Jenkins, J. L. Analysis of Pharmacology Data and the Prediction of Adverse Drug Reactions and off-Target Effects from Chemical Structure. ChemMedChem 2007, 2, 861–873. (34) Azzaoui, K.; Hamon, J.; Faller, B.; Whitebread, S.; Jacoby, E.; Bender, A.; Jenkins, J. L.; Urban, L. Modeling Promiscuity Based on in Vitro Safety Pharmacology Profiling Data. ChemMedChem 2007, 2, 874–880. (35) Lamers, J. M.; Cysouw, K. J.; Verdouw, P. D. Slow Calcium Channel Blockers and Calmodulin. Effect of Felodipine, Nifedipine, Prenylamine and Bepridil on Cardiac Sarcolemmal Calcium Pumping ATPase. Biochem. Pharmacol. 1985, 34, 3837–3843. (36) Chong, C. R.; Sullivan, D. J. New Uses for Old Drugs. Nature 2007, 448, 645–646. (37) Keiser, M. J.; Setola, V.; Irwin, J. J.; Laggner, C.; Abbas, A. I.; Hufeisen, S. J.; Jensen, N. H.; Kuijer, M. B.; Matos, R. C.; Tran, T. B.; Whaley, R.; Glennon, R. A.; Hert, J.; Thomas, K. L. H.; Edwards, D. D.; Shoichet, B. K.; Roth, B. L. Predicting New Molecular Targets for Known Drugs. Nature 2009, 462, 175–181. (38) Boolell, M.; Allen, M. J.; Ballard, S. A.; Gepi-Attee, S.; Muirhead, G. J.; Naylor, A. M.; Osterloh, I. H.; Gingell, C. Sildenafil: An Orally Active Type 5 Cyclic GMP-Specific Phosphodiesterase Inhibitor for the Treatment of Penile Erectile Dysfunction. Int. J. Impot. Res. 1996, 8, 47–52. (39) Silverman, W. A. The Schizophrenic Career of a “Monster Drug.”Pediatrics 2002, 110, 404–406. (40) Cronstein, B. N. Low-Dose Methotrexate: A Mainstay in the Treatment of Rheumatoid Arthritis. Pharmacol. Rev. 2005, 57, 163–172. (41) Nygren, P.; Fryknäs, M.; Agerup, B.; Larsson, R. Repositioning of the Anthelmintic Drug Mebendazole for the Treatment for Colon Cancer. J. Cancer Res. Clin. Oncol. 2013, 139, 2133–2140. ACS Paragon Plus Environment

27

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 42

(42) Corbett, A.; Williams, G.; Ballard, C. Drug Repositioning: An Opportunity to Develop Novel Treatments for Alzheimer’s Disease. Pharmaceuticals (Basel) 2013, 6, 1304–1321. (43) Knight, Z. A.; Lin, H.; Shokat, K. M. Targeting the Cancer Kinome through Polypharmacology. Nat. Rev. Cancer 2010, 10, 130–137. (44) Zhang, J.; Yang, P. L.; Gray, N. S. Targeting Cancer with Small Molecule Kinase Inhibitors. Nat. Rev. Cancer 2009, 9, 28–39. (45) Blanc, J.; Geney, R.; Menet, C. Type II Kinase Inhibitors: An Opportunity in Cancer for Rational Design. Anticancer Agents Med. Chem. 2013, 13, 731–747. (46) Palmieri, L.; Rastelli, G. αC Helix Displacement as a General Approach for Allosteric Modulation of Protein Kinases. Drug Discov. Today 2013, 18, 407–414. (47) Di Desidero, T.; Fioravanti, A.; Orlandi, P.; Canu, B.; Giannini, R.; Borrelli, N.; Man, S.; Xu, P.; Fontanini, G.; Basolo, F.; Kerbel, R. S.; Francia, G.; Danesi, R.; Bocci, G. Antiproliferative and Proapoptotic Activity of Sunitinib on Endothelial and Anaplastic Thyroid Cancer Cells via Inhibition of Akt and ERK1/2 Phosphorylation and by down-Regulation of Cyclin-D1. J. Clin. Endocrinol. Metab. 2013, 98, E1465–1473. (48) Dar, A. C.; Das, T. K.; Shokat, K. M.; Cagan, R. L. Chemical Genetic Discovery of Targets and Anti-Targets for Cancer Polypharmacology. Nature 2012, 486, 80–84. (49) Xie, L.; Evangelidis, T.; Xie, L.; Bourne, P. E. Drug Discovery Using Chemical Systems Biology: Weak Inhibition of Multiple Kinases May Contribute to the Anti-Cancer Effect of Nelfinavir. PLoS Comput. Biol. 2011, 7, e1002037. (50) Hearn, J. M.; Romero-Canelón, I.; Qamar, B.; Liu, Z.; Hands-Portman, I.; Sadler, P. J. Organometallic Iridium(III) Anticancer Complexes with New Mechanisms of Action: NCI-60 Screening, Mitochondrial Targeting, and Apoptosis. ACS Chem. Biol. 2013, 8, 1335–1343. (51) Laufer, S.; Bajorath, J. New Frontiers in Kinases: Second Generation Inhibitors. J. Med. Chem. 2014, 57, 2167–2168. ACS Paragon Plus Environment

28

Page 29 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

(52) Curtin, N. PARP Inhibitors for Anticancer Therapy. Biochem. Soc. Trans. 2014, 42, 82–88. (53) Ekblad, T.; Camaioni, E.; Schüler, H.; Macchiarulo, A. PARP Inhibitors: Polypharmacology versus Selective Inhibition. FEBS J. 2013, 280, 3563–3575. (54) Echeverría, P. C.; Bernthaler, A.; Dupuis, P.; Mayer, B.; Picard, D. An Interaction Network Predicted from Public Data as a Discovery Tool: Application to the Hsp90 Molecular Chaperone Machine. PLoS ONE 2011, 6, e26044. (55) Moulick, K.; Ahn, J. H.; Zong, H.; Rodina, A.; Cerchietti, L.; Gomes DaGama, E. M.; CaldasLopes, E.; Beebe, K.; Perna, F.; Hatzi, K.; Vu, L. P.; Zhao, X.; Zatorska, D.; Taldone, T.; Smith-Jones, P.; Alpaugh, M.; Gross, S. S.; Pillarsetty, N.; Ku, T.; Lewis, J. S.; Larson, S. M.; Levine, R.; Erdjument-Bromage, H.; Guzman, M. L.; Nimer, S. D.; Melnick, A.; Neckers, L.; Chiosis, G. Affinity-Based Proteomics Reveal Cancer-Specific Networks Coordinated by Hsp90. Nat. Chem. Biol. 2011, 7, 818– 826. (56) Knox, A. J. S.; Price, T.; Pawlak, M.; Golfis, G.; Flood, C. T.; Fayne, D.; Williams, D. C.; Meegan, M. J.; Lloyd, D. G. Integration of Ligand and Structure-Based Virtual Screening for the Identification of the First Dual Targeting Agent for Heat Shock Protein 90 (Hsp90) and Tubulin. J. Med. Chem. 2009, 52, 2177–2180. (57) Owellen, R. J.; Donigian, D. W.; Hartke, C. A.; Dickerson, R. M.; Kuhar, M. J. The Binding of Vinblastine to Tubulin and to Particulate Fractions of Mammalian Brain. Cancer Res. 1974, 34, 3180– 3186. (58) Zhang, Q.; Zhai, S.; Li, L.; Li, X.; Zhou, H.; Liu, A.; Su, G.; Mu, Q.; Du, Y.; Yan, B. Anti-Tumor Selectivity of a Novel Tubulin and HSP90 Dual-Targeting Inhibitor in Non-Small Cell Lung Cancer Models. Biochem. Pharmacol. 2013, 86, 351–360. (59) Anighoro, A.; Stumpfe, D.; Heikamp, K.; Bajorath, J.; Rastelli, G. Targeting Hsp90 Interactome Using in silico Polypharmacology Approaches. La Chimica e L’Industria 2013, 4, 105–106.

ACS Paragon Plus Environment

29

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 42

(60) Stevens, R. C.; Cherezov, V.; Katritch, V.; Abagyan, R.; Kuhn, P.; Rosen, H.; Wüthrich, K. The GPCR Network: A Large-Scale Collaboration to Determine Human GPCR Structure and Function. Nat. Rev. Drug Discov. 2013, 12, 25–34. (61) Shoichet, B. K.; Kobilka, B. K. Structure-Based Drug Screening for G-Protein-Coupled Receptors. Trends Pharmacol. Sci. 2012, 33, 268–272. (62) Roth, B. L.; Sheffler, D. J.; Kroeze, W. K. Magic Shotguns versus Magic Bullets: Selectively Non-Selective Drugs for Mood Disorders and Schizophrenia. Nat. Rev. Drug Discov. 2004, 3, 353– 359. (63) Lipina, T. V.; Palomo, V.; Gil, C.; Martinez, A.; Roder, J. C. Dual Inhibitor of PDE7 and GSK-3VP1.15 Acts as Antipsychotic and Cognitive Enhancer in C57BL/6J Mice. Neuropharmacology 2013, 64, 205–214. (64) Lipina, T. V.; Wang, M.; Liu, F.; Roder, J. C. Synergistic Interactions between PDE4B and GSK3: DISC1 Mutant Mice. Neuropharmacology 2012, 62, 1252–1262. (65) Youdim, M. B. H.; Buccafusco, J. J. Multi-Functional Drugs for Various CNS Targets in the Treatment of Neurodegenerative Disorders. Trends Pharmacol. Sci. 2005, 26, 27–35. (66) Bolea, I.; Juárez-Jiménez, J.; de los Rı́os, C.; Chioua, M.; Pouplana, R.; Luque, F. J.; Unzeta, M.; Marco-Contelles, J.; Samadi, A. Synthesis, Biological Evaluation, and Molecular Modeling of Donepezil

and

N-[(5-(Benzyloxy)-1-Methyl-1H-Indol-2-Yl)methyl]-N-Methylprop-2-Yn-1-Amine

Hybrids as New Multipotent Cholinesterase/Monoamine Oxidase Inhibitors for the Treatment of Alzheimer’s Disease. J. Med. Chem. 2011, 54, 8251–8270. (67) Tahtouh, T.; Elkins, J. M.; Filippakopoulos, P.; Soundararajan, M.; Burgy, G.; Durieu, E.; Cochet, C.; Schmid, R. S.; Lo, D. C.; Delhommel, F.; Oberholzer, A. E.; Pearl, L. H.; Carreaux, F.; Bazureau, J.-P.; Knapp, S.; Meijer, L. Selectivity, Cocrystal Structures, and Neuroprotective Properties of Leucettines, a Family of Protein Kinase Inhibitors Derived from the Marine Sponge Alkaloid Leucettamine B. J. Med. Chem. 2012, 55, 9312–9330. ACS Paragon Plus Environment

30

Page 31 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

(68) León, R.; Garcia, A. G.; Marco-Contelles, J. Recent Advances in the Multitarget-Directed Ligands Approach for the Treatment of Alzheimer’s Disease. Med Res Rev 2013, 33, 139–189. (69) Hopkins, A. L.; Mason, J. S.; Overington, J. P. Can We Rationally Design Promiscuous Drugs? Curr. Opin. Struct. Biol. 2006, 16, 127–136. (70) Morphy, J. R.; Harris, C. J. Designing Multi-Target Drugs; Royal Society of Chemistry: Cambridge, 2012. (71) Bajorath, J.; Peltason, L.; Wawer, M.; Guha, R.; Lajiness, M. S.; Van Drie, J. H. Navigating Structure-Activity Landscapes. Drug Discov. Today 2009, 14, 698–705. (72) Dimova, D.; Wawer, M.; Wassermann, A. M.; Bajorath, J. Design of Multitarget Activity Landscapes That Capture Hierarchical Activity Cliff Distributions. J. Chem. Inf. Model. 2011, 51, 258–266. (73) Hu, Y.; Bajorath, J. Promiscuity Profiles of Bioactive Compounds: Potency Range and Difference Distributions and the Relation to Target Numbers and Families. Med. Chem. Commun. 2013, 4, 1196– 1201. (74) Hu, Y.; Bajorath, J. Activity Profile Relationships between Structurally Similar Promiscuous Compounds. Eur. J. Med. Chem. 2013, 69, 393–398. (75) Morphy, R.; Rankovic, Z. Designed Multiple Ligands. An Emerging Drug Discovery Paradigm. J. Med. Chem. 2005, 48, 6523–6543. (76) Apsel, B.; Blair, J. A.; Gonzalez, B.; Nazif, T. M.; Feldman, M. E.; Aizenstein, B.; Hoffman, R.; Williams, R. L.; Shokat, K. M.; Knight, Z. A. Targeted Polypharmacology: Discovery of Dual Inhibitors of Tyrosine and Phosphoinositide Kinases. Nat. Chem. Biol. 2008, 4, 691–699. (77) Besnard, J.; Ruda, G. F.; Setola, V.; Abecassis, K.; Rodriguiz, R. M.; Huang, X.-P.; Norval, S.; Sassano, M. F.; Shin, A. I.; Webster, L. A.; Simeons, F. R. C.; Stojanovski, L.; Prat, A.; Seidah, N. G.; Constam, D. B.; Bickerton, G. R.; Read, K. D.; Wetsel, W. C.; Gilbert, I. H.; Roth, B. L.; Hopkins, A. L. Automated Design of Ligands to Polypharmacological Profiles. Nature 2012, 492, 215–220.

ACS Paragon Plus Environment

31

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 42

(78) Bottegoni, G.; Favia, A. D.; Recanatini, M.; Cavalli, A. The Role of Fragment-Based and Computational Methods in Polypharmacology. Drug Discov. Today 2012, 17, 23–34. (79) Gaulton, A.; Bellis, L. J.; Bento, A. P.; Chambers, J.; Davies, M.; Hersey, A.; Light, Y.; McGlinchey, S.; Michalovich, D., Al-Lazikani, B.; Overington, J. P. ChEMBL: A Large-scale Bioactivity Database for Drug Discovery. Nucleic Acids Res. 2012, 40, D1100-D1107. (80) Hu, Y.; Bajorath, J. Systematic Identification of Scaffolds Representing Compounds Active against Individual Targets and Single or Multiple Target Families. J. Chem. Inf. Model. 2013, 53, 312– 326. (81) Keiser, M. J.; Roth, B. L.; Armbruster, B. N.; Ernsberger, P.; Irwin, J. J.; Shoichet, B. K. Relating Protein Pharmacology by Ligand Chemistry. Nat. Biotechnol. 2007, 25, 197–206. (82) Altschul, S. F. Evaluating the Statistical Significance of Multiple Distinct Local Alignments. In Theoretical and Computational Methods in Genome Research; Suhai, S., Ed.; Springer US: New York, 1997; pp. 1–14. (83) DeGraw, A. J.; Keiser, M. J.; Ochocki, J. D.; Shoichet, B. K.; Distefano, M. D. Prediction and Evaluation of Protein Farnesyltransferase Inhibition by Commercial Drugs. J. Med. Chem. 2010, 53, 2464–2471. (84) Carlsson, J.; Yoo, L.; Gao, Z.-G.; Irwin, J. J.; Shoichet, B. K.; Jacobson, K. A. Structure-Based Discovery of A2A Adenosine Receptor Ligands. J. Med. Chem. 2010, 53, 3748–3755. (85) Hert, J.; Keiser, M. J.; Irwin, J. J.; Oprea, T. I.; Shoichet, B. K. Quantifying the Relationships among Drug Classes. J. Chem. Inf. Model. 2008, 48, 755–765. (86) Reker, D.; Rodrigues, T.; Schneider, P.; Schneider, G. Identifying the Macromolecular Targets of de Novo-Designed Chemical Entities through Self-Organizing Map Consensus. Proc. Natl. Acad. Sci. U.S.A. 2014, 111, 4067–4072. (87) Kohonen, T. Self-Organizing Maps; Springer: New York, 2001.

ACS Paragon Plus Environment

32

Page 33 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

(88) Anighoro, A.; Rastelli, G. Enrichment Factor Analyses on G-Protein Coupled Receptors with Known Crystal Structure. J. Chem. Inf. Model. 2013, 53, 739–743. (89) Selvam, B.; Porter, S. L.; Tikhonova, I. G. Addressing Selective Polypharmacology of Antipsychotic Drugs Targeting the Bioaminergic Receptors through Receptor Dynamic Conformational Ensembles. J. Chem. Inf. Model. 2013, 53, 1761–1774. (90) Ripphausen, P.; Nisius, B.; Peltason, L.; Bajorath, J. Quo Vadis, Virtual Screening? A Comprehensive Survey of Prospective Applications. J. Med. Chem. 2010, 53, 8461–8467. (91) Rastelli, G. Emerging Topics in Structure-Based Virtual Screening. Pharm. Res. 2013, 30, 1458– 1463. (92) Rastelli, G.; Degliesposti, G.; Del Rio, A.; Sgobba, M. Binding Estimation after Refinement, a New Automated Procedure for the Refinement and Rescoring of Docked Ligands in Virtual Screening. Chem. Biol. Drug. Des. 2009, 73, 283–286.

ACS Paragon Plus Environment

33

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 42

Figure legends

Figure 1. Scaffolds and promiscuity profiles. Shown is a set of topologically equivalent scaffolds including dipenylmethane, one of the most promiscuous scaffolds found in bioactive compounds, and their promiscuity profiles. Nodes represent scaffolds and an edge connects two scaffolds if they are active against the same target(s). Edge labels report the number of shared targets (and node labels scaffold IDs). The structure of each scaffold is annotated with its promiscuity profile consisting of “target family : number of relevant targets” expressions. For example, the scaffold at the bottom has the activity profile “1:1; 2:3; 10:7”, which means that this scaffold is present in compounds that are active against one target in target family 1, three targets in family 2, and seven targets in family 10. The figure was adopted from reference 15. Scaffolds and target families are numbered according to the original publication.

Figure 2. Chemical structures of representative drugs and pharmaceutically relevant compounds discussed in the text.

Figure 3. Scaffold- and compound-based target networks. For ChEMBL79 release 2 (available at the end of 2009) and release 14 (available in 2012), scaffold- and compound-based target networks are compared on the basis of Ki activity values (equilibrium constants) in (a) and (b), respectively, and on the basis of IC50 measurements in (c) and (d), respectively. Nodes represent targets and edges connect target pairs that share at least five scaffolds or compounds. Scaffolds that represent compounds active against multiple targets from the same family or different protein families are designated single-family (SF) and multi-family (MF) scaffolds, respectively. Intra- and inter-family target pairs are indicated by solid and dashed edges, respectively, and target families are listed and color-coded in (e). Network formation results from the presence of promiscuous bioactive compounds. The comparison of different ACS Paragon Plus Environment

34

Page 35 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

ChEMBL releases illustrates the growth in promiscuous compounds. Furthermore, the comparison of Ki and IC50 value-based networks reveals that the increase in compounds promiscuous across different target families predominantly results from IC50-based compound activity data. The figure was adopted from reference 80.

Figure 4. Different approaches relevant to the generation of multitarget drugs and their interplay are illustrated. The integration of complementary field of studies is thought to lead the way to rational polypharmacology approaches.

ACS Paragon Plus Environment

35

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 42

Figure 1 8:2; 9:1;15:3; 17:4

3

2

7:2; 8:2; 9:1; 15:6; 16:1; 17:1; 18:1; 19:1

1 1:1; 8:6; 19:1 5

4

6 2

2

6 7:5; 8:8; 10:7; 15:1; 17:3; 18:2; 19:3

4

4

5:2; 8:1; 10:4; 15:1; 17:1; 19:1

1:1; 2:3; 10:7

ACS Paragon Plus Environment

36

Page 37 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Figure 2

ACS Paragon Plus Environment

37

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 42

Figure 3

(a)

Ki

ChEMBL 2

(b)

Ki

ChEMBL 2

scaffold-based target pairs 30 targets 36 intra-family pairs 0 inter-family pairs

ChEMBL 14

83 targets 319 intra-family pairs 237 inter-family pairs

compound-based target pairs 97 targets 169 intra-family pairs 0 inter-family pairs

ChEMBL 14

ACS Paragon Plus Environment

176 targets 457 intra-family pairs 86 inter-family pairs

38

Page 39 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(c) IC50 ChEMBL 2

Journal of Medicinal Chemistry

scaffold-based target pairs 24 targets 27 intra-family pairs 2 inter-family pairs

ChEMBL 14

ACS Paragon Plus Environment

157 targets 269 intra-family pairs 717 inter-family pairs

39

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(d) IC50

Page 40 of 42

compound-based target pairs

ChEMBL 2

101 targets 151 intra-family pairs 10 inter-family pairs

ChEMBL 14

269 targets 486 intra-family pairs 178 inter-family pairs

(e) target families Tyr kinase

Glutamate-gated ion channel

Chemokine GPCR family 1

Ser_Thr kinase

Ligand-gated ion channel

Lipid-like GPCR family 1

Ser_Thr_Tyr kinase

Potassium ion channel

Monoamine GPCR family 1

Aspartic protease

Transient receptor ion channel

Nucleotide-like GPCR family 1

Cysteine protease

Sodium ion channel

Short peptide GPCR family 1

Metalloprotease

Carbonic anhydrase

Unspecified GPCR family 1

Serine protease

Aldo/keto reductase

GPCR family 2

Tyr phosphatase

Bcl-2

GPCR family 3

Ser_Thr phosphatase

Cytochrome P450 enzyme

Phosphodiesterase

Ser_Thr_Tyr phosphatase

Glycosyl hydrolase

Phospholipase

Type-B carboxylesterase/lipase

Nuclear receptor Lipoxygenase

Lipase Histone deacetylase

Sodium:dicarboxylate symporter Sodium:neurotransmitter symporter

ACS Paragon Plus Environment

40

Page 41 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Figure 4

ACS Paragon Plus Environment

41

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 42

Table of Contents Graphic

ACS Paragon Plus Environment

42