Smallmolecule proteinprotein interaction inhibitors - Wiley Online Library

4 downloads 0 Views 285KB Size Report
Small-Molecule Protein–Protein Interaction Inhibitors: Therapeutic. Potential in Light .... CCR5 receptor and selectively blocks the binding of viral gp-. 120 to this ...
IUBMB

Life, 62(10): 724–731, October 2010

Critical Review Small-Molecule Protein–Protein Interaction Inhibitors: Therapeutic Potential in Light of Molecular Size, Chemical Space, and Ligand Binding Efficiency Considerations Peter Buchwald Diabetes Research Institute and the Department of Molecular and Cellular Pharmacology, Miller School of Medicine, University of Miami, Miami, FL, USA

Summary As the ultimate function of proteins depends to a great extent on their binding partners, protein–protein interactions (PPIs) represent a treasure trove of possible new therapeutic targets. Unfortunately, interfaces involved in PPIs are not wellsuited for effective small molecule binding. Nevertheless, successful examples of small-molecule PPI inhibitors (PPIIs) are beginning to accumulate, and the sheer number of PPIs that form the human interactome implies that, despite the relative unsuitability of PPIs to serve as ‘‘druggable’’ targets, small-molecule PPIIs can still provide novel pharmacological tools and new innovative drugs in at least some areas. Here, after some illustrative examples, accumulating information on the binding efficiency, molecular size, and chemical space requirements will be briefly reviewed. Therapeutic success can only be achieved if these considerations are incorporated into the search process and if careful medicinal chemistry approaches are used to address the absorption, distribution, metabolism, and excretion requirements of larger molecules that are often needed for this target class due to the lower efficiency of binding. Ó 2010 IUBMB IUBMB

Life, 62(10): 724–731, 2010

Keywords

drug discovery; druggability; ligand efficiency; molecular size; protein binding. Abbreviations ADME, absorption, distribution, metabolism, and excretion; HTS, high-throughput screening; JNK, c-Jun N-terminal kinase; GPCR, G-protein coupled receptor; LE, ligand efficiency; PPI, protein–protein interaction; PPII, protein–protein interaction inhibitor; RGS, regulators of G-protein signaling; TNF, tumor necrosis factor.

Received 17 August 2010; accepted 9 September 2010 Address correspondence to: Peter Buchwald, Diabetes Research Institute, Miller School of Medicine, University of Miami, 1450 NW 10 Ave (R-134), Miami, FL 33136, USA. Tel.: 11 305 243 9657. E-mail: [email protected] ISSN 1521-6543 print/ISSN 1521-6551 online DOI: 10.1002/iub.383

INTRODUCTION Drug research and discovery, fields that undeniably were the main driving force behind the considerable medical progress of the last century, are facing increasing challenges and are in clear need of innovative new approaches (1). Despite exponentially increasing R&D expenditures, the number of new drug approvals in the United States has been essentially stagnating around 15–25 per year since the mid-1960s. Furthermore, the majority of launched drugs is in fact derived by modifications of known drug structures or published lead structures, and the chemical structures of the final drugs are very closely related to those of their original leads (2). Hence, potential new therapeutic targets are of considerable interest, and protein–protein interactions (PPIs) could offer many attractive opportunities as they play critical roles in many biological processes both under normal physiological as well as during pathological conditions. It is now well recognized that most biological processes are not performed by stand-alone single proteins, but by complexes consisting of different, interacting proteins. Accordingly, as the ultimate function of a protein will depend on its binding partners in such complexes, PPI inhibitors (PPIIs) could achieve more specific modulation of protein functions, which could result in novel therapeutic applications. The current interest in the therapeutic potential of biotechnology products such as fusion proteins or antibodies that can act as potent and specific modulators of PPIs is a clear illustration of this trend. PPIIs approved for clinical applications are usually based on humanized monoclonal antibodies. Although antibodies have the advantage of being highly specific for their targets and quite stable in human serum, they, as all other protein therapies, are often seriously hindered by solubility, route of administration, distribution, and stability problems as well as by the possibility of a strong immune response that the body can mount against foreign proteins (3). A recent disastrous drug trial with

SMALL-MOLECULE PROTEIN–PROTEIN INTERACTION INHIBITORS

TGN1412, an anti-CD28 superagonist monoclonal antibody that induced a life-threatening cytokine storm in the first human clinical patients pointing to the importance of species relevance (4), is likely to make development of such biotechnology compounds even more complicated (5). Traditional small molecule drugs could provide a convenient alternative; however, small molecules were considered unlikely to be effective PPIIs because these interactions usually involve relatively large pro˚ 2) that also lack the well-defined tein surfaces (1,500–3,000 A binding pockets present on the traditional targets of most existing drugs [G-protein coupled receptors (GPCRs), ion channels, and enzymes]. Hence, for a long time, small molecules were not pursued as possible PPIIs. During the last two decades, it has become increasingly clear that, at least in certain cases, small molecules can act as quite effective PPIIs. This is probably due to the fact that most PPIs involve so-called ‘‘hot spots,’’ relatively small parts of the interface that are essential for high-affinity binding (6, 7). Here, we will focus exclusively on PPIIs (antagonists) and not on agonists that enhance binding or stimulate activity. Small molecules are much less likely to succeed as stimulators as this, in addition to binding, also requires the correct triggering of the activation cascade. As antagonists in general, small molecule PPIIs can be either ‘‘orthosteric’’— directly interfering with critical hot spots on the interface and competing with the original protein ligand—or ‘‘allosteric’’— binding at some different site, but causing conformational changes that are sufficient to interfere with the binding of the protein ligand. Within the last few years, sufficiently effective small-molecule inhibitors have been identified for a few important PPIs (8–12). With accumulating information, even databases of small-molecule PPIIs and related protein structures are being built, for example, http: //www-cryst.bioc.cam.ac.uk/databases/timbal (13) or http: //2p2idb.cnrs-mrs.fr/ (14). Furthermore, PPIIs from Abbott, Genentech, Johnson & Johnson, or Roche are in clinical development, and two drugs already on the market can also be considered as PPIIs (15): tirofiban and maraviroc. However, it has to be noted that for targets such as SH2 domains, proteases, and integrins, where interactions with the protein partner typically involve binding to an isolated peptide loop or strand and not to a broad protein surface, it is considerably easier to identify small-molecule inhibitors. Because these interactions more closely resemble cleft-based protein–small molecule binding than true broadsurface PPIs, the corresponding inhibitors sometimes are not considered as PPIIs in a stricter sense (12). On the otherhand, protein aggregation inhibitors, which are of special interest, for example, for amyloid-b or a-synuclein, are sometimes also considered as PPIIs in a wider sense (16). Here, after a brief review of a few illustrative examples (for detailed reviews, see refs. 8–12 and references therein), we will focus on molecular size, chemical space, and ligand binding efficiency aspects.

725

ILLUSTRATIVE EXAMPLES Extracellular Targets Integrins. Some of the very first small-molecule PPIIs were found for integrins (11). Integrins are adhesion receptors on the cell surface composed of noncovalently attached a and b chains and are involved in various cell–cell interaction functions. Some of the earliest drug discovery efforts targeted the aIIbb3 receptor, a platelet specific integrin that plays important roles in stopping the bleeding at vascular injury sites and in pathological thrombosis leading to heart attacks and stroke. These efforts ultimately resulted in two molecules that achieved clinical approval for use in preventing platelet aggregation: epifibatide, which retains substantial peptide character and is not discussed here, and tirofiban (1, Fig. 1) (11). Starting from the recognition that binding of the protein ligands (e.g., fibronectin and vitronectin) to this receptor involves a so-called RGD motif (Arg– Gly–Asp), various research group succeeded in designing potent small-molecule inhibitors that reproduced the two charged functional groups important for binding, setting the stage for the mimicry of natural protein binding partners as a strategy for PPI inhibition (17). CCR5–gp-120. The HIV-1 coreceptors, CXCR4 and CCR5, which belong to the GPCR superfamily, are good targets for therapeutic interventions as their binding to a portion of gp-120 is necessary for HIV-1 entry into cells expressing CD4 receptors (e.g., T cells and macrophages). Maraviroc (2), a CCR5-receptor antagonist, binds reversibly to the human chemokine CCR5 receptor and selectively blocks the binding of viral gp120 to this receptor, thereby impeding the conformational changes required for the cell-entry of CCR5-tropic HIV-1 (18); hence, it can be considered as a PPII (15). IL-2–IL-2R. Interleukin-2 (IL-2) is a cytokine that binds to a heterotrimeric receptor (IL-2R) on the surface of T cells and induces their activation. Therefore, it is an important therapeutic target for immunosuppression, for example, in organ transplant recipients. Antibodies inhibiting this interaction (basiliximab and daclizumab) have shown clinical success, and several quite potent small-molecule inhibitors have also been identified, for example, SP4206 (3, Ki 5 60 nM) (10, 19). A relatively large molecular size combined with good charge- and shape-complementarity at the binding site (Fig. 3A) make this stronger binding possible resulting in nanomolar affinity. TNF–TNF-R. Tumor necrosis factor (TNF) is a cytokine involved in systemic inflammations and in immune regulation; hence, it is an obvious therapeutic target for many diseases. The development of antibodies (infliximab and adalimumab) and fusion proteins (etanercept) inhibiting the binding of TNF to its receptors (TNF-R) is one of the few recent immunopharmacology success stories (20). Accordingly, there is considerable in-

726

BUCHWALD

Figure 1. Chemical structures of selected PPIIs for extracellular targets.

terest in corresponding small-molecule PPIIs, but only limited success has been achieved so far. SP304 (4) is one of the best inhibitors identified until now, but it still has a relatively low affinity (Kd 5 15 lM) (21). Interestingly, it seems to act by causing the dissociation of a monomeric unit from the trimeric TNF, a mechanism that might be relevant for other members of the TNF ligand–receptor superfamily. A set of compounds identified as possible PPIIs in an earlier work (22) provided an important warning. At first, they seemed to be promising nanomolar inhibitors in vitro, but later it turned out that they modified the receptor covalently by a photochemical reaction due to exposure to light so that their activity was considerably diminished in dark, which, of course, is the common physiological condition. CD154–CD40. This ligand–receptor pair of the TNF superfamily provides an important costimulatory interaction regulating T cell activation and, hence, is a promising therapeutic target in autoimmune diseases and in transplant recipients. We have recently identified the first small-molecule PPIIs for this interaction. Starting from suramin, we found a number of organic dyes such as direct red 80 or mordant brown 1 (5) as sufficiently active (low-micromolar) and specific inhibitors (23, 24). Because of their good protein binding ability, organic dyes could provide a valuable chemical space to search for smallmolecule PPIIs assuming that sufficient specificity can be achieved, which could be a challenge in many cases, but was possible for CD154–CD40.

Intracellular Targets Inhibition of intracellular PPIs is even more difficult because the inhibitor needs to have adequate membrane permeability to

reach its target. For example, compound 6 (Fig. 2) was identified as a sufficiently potent inhibitor of the c-Jun N-terminal kinase-1 (JNK-1)–JNK-interacting protein 1 (JIP-1) interaction [median inhibitory concentration (IC50) 5 5 lM]; however, it was not cell-permeable, so its biological activity could not be tested (25). Nevertheless, quite a few examples are available even with submicromolar potency, and there are even compounds in clinical development that are intended for oral administration. HDM2–p53. The transcription factor p53 regulates the cell cycle in response to cellular stress leading to DNA damage. In normal cells, p53 is usually inactive being bound to the human protein double minute 2 (HDM2; or its mouse analog known as MDM2). HDM2 is a ubiquitin E3 ligase that binds to p53 concealing its N-terminal transactivation domain and eventually leading to its degradation via the proteasome (26). In cancerous cells, tumor suppression by p53 seems reduced; hence, the inhibition of its interaction with HDM2 could restore p53 function resulting in apoptosis of the abnormal, cancerous cell. Small molecule HDM2–p53 inhibitors have been reviewed recently in detail (27). Nutlin-3 (7; Fig. 2), an imidazoline derivative identified by high-throughput screening (HTS), is probably one of the best characterized p53–MDM2 antagonists (Kd 5 0.1 lM) and one of its analogs (R7112, RO5045337) is in clinical trials (28). JNJ-26854165 (8) is another one of the PPIIs that reached clinical development and is a novel HDM2 ubiquitin ligase antagonist intended for oral administration (27). For this compound, some recent results also indicate the possibility of p53-independent activities (28). Bcl-2/Bcl-XL–BAK/BAD. Members of the B cell lymphoma 2 (Bcl-2) family play important roles in mediating apoptosis. They bind to a-helical portions of the proapoptotic molecules

SMALL-MOLECULE PROTEIN–PROTEIN INTERACTION INHIBITORS

727

Figure 2. Chemical structures of selected PPIIs for intracellular targets.

Figure 3. Three-dimensional structures of the protein–small-molecule PPII complex for IL-2–SP4206 (A: PDB no. 1PY2) and BclXL–ABT-737 (B: PDB no. 2YXJ). In both cases, the small-molecule PPIIs (shown as ball-and-stick structure) bind on the surface of the protein (shown as a surface colored by the electrostatic potential) in a position that directly interferes with the binding of the protein partner. Both cases nicely illustrate that electrostatic and van der Waals interactions due to charge and shape complementarity, respectively, are responsible for most of the binding energy. BAK or BAD, and these PPIs are valuable targets for inhibition as possible cancer treatments. Several research groups have identified high-affinity small-molecule inhibitors, and one of them (10) is in clinical development. In fact as shown in Fig. 2, navitoclax (10, ABT-263) is a modification of the original lead

ABT-737 (9) specifically designed to improve oral bioavailability (29). GPCR Signaling Components. GPCRs being the target of many existing, successful drugs, the therapeutic efficacy of tar-

728

BUCHWALD

geting GPCR signaling is well-confirmed. Hence, there has been considerable interest in the modulation of signaling downstream of these receptors via the inhibition of either G-protein–effector or G-protein–regulator PPIs (16). G-protein heterotrimers, which are the first signaling molecules downstream of GPCRs, represent an obvious first target for PPIIs. Within the last few years, a number of small molecules such as 11 (M119) were identified that bind to the Gbg subunit inhibiting its interactions with downstream effectors in sufficiently specific manner (16, 30, 31). Another possibility is to inhibit PPIs involved in regulating G-protein pathways such as those with the ‘‘regulators of G-protein signaling’’ (RGS) that are members of the family of GTPase accelerating proteins and bind directly to the Ga subunit accelerating the intrinsic hydrolysis rate of guanosine triphosphate (GTP) (32). A small-molecule inhibitor of RGS4 binding to Gao with low micromolar activity has been identified (CCG-4986) (33); however, in the end, it turned out to be a covalent binder (34), an issue somewhat similar to that encountered for TNF PPIIs. Hence, promising progress has been made in identifying small-molecule PPIIs to modulate signaling downstream of GPCRs, but the therapeutic applicability of the compounds identified so far is somewhat unclear.

BINDING EFFICIENCY AND CHEMICAL SPACE Size Matters Therapeutic applicability requires sufficient activity toward the intended target. This is usually characterized by the median effective dose (ED50)—median IC (IC50) for an inhibitor—or the closely related dissociation constant (Kd), all expressed on a molar scale. It is desirable to have potency somewhere in the submicromolar (\1 lM) range; for example, the average of all binding-affinity–related endpoints of all existing small molecule drugs is quite low being around 20 nM (35). For effective PPII, the small molecule should have an affinity in a range comparable to that of the protein partner to not have a serious disadvantage in displacing it. This requires sufficiently strong interactions (K 5 e–DG/RT, DG being the binding free energy) that, however, are more difficult to achieve on protein surfaces than at well-defined ligand-binding domains. It is not entirely accidental that traditional drug targets (e.g., GPCRs, ion channels, and enzymes) are typically those that have a preformed cavity or cleft for binding their natural (small molecule) ligand(s) with good affinity and specificity. The binding site allows the focusing of multiple binding interactions (hydrogen bonds, ionic- or polar-interactions) in a relatively small volume, and then this can also be exploited for drug design purposes to obtain efficient ligand binding (i.e., a large binding energy to ligand volume ratio) and adequate specificity for the target of interest. Targetable binding pockets on protein–protein interfaces (i.e., those that are suitable to bind small molecules) were indeed found to be considerably smaller than the binding pockets of traditional protein–ligand interactions (15). For example,

although most marketed drugs were found to target a single ˚ 3, binding pocket with an average occupied volume of 271 A PPIIs were found to typically target between three and five pockets with an average occupied volume that is much smaller ˚ 3) (15). As pocket size limits the achievable maximum (100 A energy (36), in most cases, PPIIs can achieve adequate affinity only by being large enough to simultaneously reach a sufficient number of smaller pockets. For many PPIs, inhibition with nanomolar potency will require larger molecular sizes, typically larger than desired for drug-like structures and especially for good absorption properties (Mw \ 500). A comparison of identified PPII structures and FDA approved drugs revealed that PPIIs indeed tend to be larger than enzyme inhibitors, ion channel modulators, or receptor ligands (37). A relatively large molecular size tends to become a problem as such compounds are likely to be difficult to formulate and to have poor membrane permeability as well as poor oral bioavailability. Until recently, it had been standard practice to filter drug candidates and not pursue those that do not meet Lipinski’s so-called ‘‘rule-of-five’’ criteria (38), that is, those that have high molecular weight (Mw [ 500), high number of hydrogen bond donors ([5) or acceptors ([10), and high lipophilicity (CLOGP [ 5). Nevertheless, even if many PPIIs might be larger than commonly considered desirable for ‘‘drug-like’’ compounds, in a number of cases, it has been possible to optimize absorption, distribution, metabolism, and excretion (ADME) and pharmacokinetic properties and even develop compounds intended for oral administration. For example, navitoclax (10) was designed to improve the bioavailability of 9 (29), and it indeed has improved oral bioavailability despite violating three of the four ‘‘rules-of-five.’’

Ligand Binding Efficiency The binding energy/molecular size problem can be even more clearly highlighted with the help of a more quantitative metric characterizing ligand binding efficiency: binding energy normalized to size (39). Ligand efficiency (LE) is defined as the binding energy per unit size, typically the binding free energy per nonhydrogen atom: LE 5 DG0/NnHa (39). For a given target, comparing the LE of all identified ligands can give a wide range as many might not yet have an optimized structure. Along these lines, LE was indeed found to cover a wide range for all compounds in a PPII database (0.6–1.7 kJ/NnHa and even higher) (13). Comparing LE for optimized ligands is much more informative as it can better highlight differences among classes of binding sites and/or ligands. The average LE of highest-affinity fragments and small molecules that target PPIs were found to be 1.0 kJ/atom (10), similar but somewhat less than the average LE of 1.5 kJ/atom seen for usual protein–ligand interactions (36, 39–41). As a useful guide, this latter value corresponds to an approximate twofold increase in affinity (halving of Kd or IC50) with the addition of each (nonhydrogen) atom. As most PPIIs in their bound positions are not entirely sur-

SMALL-MOLECULE PROTEIN–PROTEIN INTERACTION INHIBITORS

rounded by their protein target, they can only interact with part of their surface (Fig. 3); hence, the binding will be less efficient than at traditional binding pockets. Accordingly for PPIIs, affinities in the nanomolar range are more difficult to achieve than for typical drugs. It is still possible, but, in general, will require larger structures. For example, achieving a Kd 5 20 nM, the average of existing small molecule drugs (35), with an LE 5 1 kJ/atom requires a structure with [45 nonhydrogen atoms, which is around the limit of ‘‘druggability’’ highlighting the importance for the need of structures that can show good protein binding. Interestingly, binding of the small molecule ligand was often found to cause large conformational rearrangements of the PPI sites (usually resulting in better burying of the inhibitor), for example, in the case of Bcl-XL, IL-2, MDM-2, or survivin (42).

Chemical Space There are still relatively few successful and/or sufficiently potent PPIIs identified to clearly delineate their chemical space and compare it with that of existing drugs. Although good inhibitory activity can be achieved without very close protein mimicry in many cases, it is also becoming increasingly clear that the chemical space of PPIIs does not fully overlap with that of existing drugs. Notably, out of close to 5,000 compounds in the DrugBank small set, only seven showed good structural similarity (Tanimoto index [ 0.8) with one of 66 PPIIs analyzed in a very recent study (43). Compared with known drugs, PPIIs tend to have a larger molecular size, a more hydrophobic nature, and structures with more rigid, aromatic scaffolds in combination with charged or polar functionalities. A number of recent studies confirmed a poor correspondence between the chemical spaces of existing screening libraries and known PPIIs as evidenced especially by descriptors related to molecular shape and the number of aromatic rings (37, 43, 44). Hence, available chemical libraries commonly used for HTS that are focused on a ‘‘drug-like’’ chemical space may not necessarily be a good place to search for PPIIs. This can limit the success rate of HTS approaches and might at least partially account for the initial failures of many screening efforts to identify sufficiently promising small-molecule PPIIs. Nevertheless, certain structures clearly show good proteinbinding ability, and a chemical space with privileged structures for protein binding (45) might very well exist. For example, biphenyl scaffolds or COOH moieties seem to be present with high frequency among good protein binders (45). We have also found certain organic dyes such as erythrosine B to act as promiscuous PPIIs (23). An analysis of protein druggability found that a substantial number of proteins involved in binding to other proteins contained druggable binding sites (29%); hence, despite all these difficulties, a fair number of targets involved in PPIs could still be suitable for small molecule modulation (46). There are also computational, even web-based methods to predict protein interaction hotspots, that is, sites where a mutation or a small mole-

729

cule docking is most likely to disrupt binding, for example, see ref. 7. Compared to antibodies, small-molecule PPIIs might have a kinetic advantage as they can more easily intercalate between the interacting protein partners in which the interface has a dynamic nature (10). In fact, with recent successes, the tide has turned enough that it has even become feasible to contemplate the possibility of modulating signal transduction enzymes by inhibiting their PPIs—that is, the possibility that there might be cases in which targeting the more difficult PPI site is advantageous compared with targeting the typically more druggable active site of the enzyme (47). Small-molecule inhibitions with micromolar activities have been achieved for PPIs for important signal transduction pathways involving, for example, JNK, polo-like kinase 1, phosphosoinositide-dependent protein kinase 1, calcineurin, or heat shock protein 90 (ref. 47 and references therein).

CONCLUSIONS Most PPIs are difficult to inhibit with small molecules; nevertheless, the sheer number of PPIs, estimated to be in the range of 650,000 for humans (human interactome) (48), implies that ultimately a significant number will be druggable. Existing small molecule drugs target only about 1% of the human proteome, which may have around 30,000 unique proteins (12); hence, even with a 10-fold lower success rate (0.1%), PPI targeting still could provide therapeutic value for a treasure trove of hundreds of protein interactions. Successful examples of small-molecule PPIIs, including clinical successes, are beginning to accumulate, but it is also becoming increasingly clear that because of the lower efficiency of binding seen here, therapeutic success can only be achieved by using careful medicinal chemistry approaches to search within the correct chemical space and to address the ADME requirements of the larger molecules that are likely to be needed for PPI inhibition. ACKNOWLEDGEMENTS The author is grateful to the Diabetes Research Institute Foundation (www.diabetesresearch.org) for financial support.

REFERENCES 1. Paul, S. M., Mytelka, D. S., Dunwiddie, C. T., Persinger, C. C., Munos, B. H., Lindborg, S. R., and Schacht, A. L. (2010) How to improve R&D productivity: the pharmaceutical industry’s grand challenge. Nat. Rev. Drug Discov. 9, 203–214. 2. Proudfoot, J. R. (2002) Drugs, leads, and drug-likeness: an analysis of some recently launched drugs. Bioorg. Med. Chem. Lett. 12, 1647–1650. 3. Leader, B., Baca, Q. J., and Golan, D. E. (2008) Protein therapeutics: a summary and pharmacological classification. Nat. Rev. Drug Discov. 7, 21–39. 4. Suntharalingam, G., Perry, M. R., Ward, S., Brett, S. J., Castello-Cortes, A., Brunner, M. D., and Panoskaltsis, N. (2006) Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N. Engl. J. Med. 355, 1018–1028.

730

BUCHWALD

5. Wadman, M. (2006) London’s disastrous drug trial has serious side effects for research. Nature 440, 388–389. 6. Bogan, A. A. and Thorn, K. S. (1998) Anatomy of hot spots in protein interfaces. J. Mol. Biol. 280, 1–9. 7. Darnell, S. J., LeGault, L., and Mitchell, J. C. (2008) KFC Server: interactive forecasting of protein interaction hot spots. Nucleic Acids Res. 36, W265–W269. 8. Berg, T. (2003) Modulation of protein–protein interactions with small organic molecules. Angew. Chem. Int. Ed. Engl. 42, 2462–2481. 9. Arkin, M. R. and Wells, J. A. (2004) Small-molecule inhibitors of protein–protein interactions: progressing towards the dream. Nat. Rev. Drug Discov. 3, 301–317. 10. Wells, J. A. and McClendon, C. L. (2007) Reaching for high-hanging fruit in drug discovery at protein–protein interfaces. Nature 450, 1001– 1009. 11. Fry, D. C. (2006) Protein–protein interactions as targets for small molecule drug discovery. Biopolymers 84, 535–552. 12. Whitty, A. and Kumaravel, G. (2006) Between a rock and a hard place? Nat. Chem. Biol. 2, 112–118. 13. Higueruelo, A. P., Schreyer, A., Bickerton, G. R., Pitt, W. R., Groom, C. R., and Blundell, T. L. (2009) Atomic interactions and profile of small molecules disrupting protein–protein interfaces: the TIMBAL database. Chem. Biol. Drug Des. 74, 457–467. 14. Bourgeas, R., Basse, M. J., Morelli, X., and Roche, P. (2010) Atomic analysis of protein–protein interfaces with known inhibitors: the 2P2I database. PLoS ONE 5, e9598. 15. Fuller, J. C., Burgoyne, N. J., and Jackson, R. M. (2009) Predicting druggable binding sites at the protein–protein interface. Drug Discov. Today 14, 155–161. 16. Blazer, L. L. and Neubig, R. R. (2009) Small molecule protein–protein interaction inhibitors as CNS therapeutic agents: current progress and future hurdles. Neuropsychopharmacology 34, 126–141. 17. Fry, D. C. (2008) Drug-like inhibitors of protein–protein interactions: a structural examination of effective protein mimicry. Curr. Protein Pept. Sci. 9, 240–247. 18. Dorr, P., Westby, M., Dobbs, S., Griffin, P., Irvine, B., Macartney, M., Mori, J., Rickett, G., Smith-Burchnell, C., Napier, C., Webster, R., Armour, D., Price, D., Stammen, B., Wood, A., and Perros, M. (2005) Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti-human immunodeficiency virus type 1 activity. Antimicrob. Agents. Chemother. 49, 4721–4732. 19. Raimundo, B. C., Oslob, J. D., Braisted, A. C., Hyde, J., McDowell, R. S., Randal, M., Waal, N. D., Wilkinson, J., Yu, C. H., and Arkin, M. R. (2004) Integrating fragment assembly and biophysical methods in the chemical advancement of small-molecule antagonists of IL-2: an approach for inhibiting protein–protein interactions. J. Med. Chem. 47, 3111–3130. 20. Palladino, M. A., Bahjat, F. R., Theodorakis, E. A., and Moldawer, L. L. (2003) Anti-TNF-alpha therapies: the next generation. Nat. Rev. Drug Discov. 2, 736–746. 21. He, M. M., Smith, A. S., Oslob, J. D., Flanagan, W. M., Braisted, A. C., Whitty, A., Cancilla, M. T., Wang, J., Lugovskoy, A. A., Yoburn, J. C., Fung, A. D., Farrington, G., Eldredge, J. K., Day, E. S., Cruz, L. A., Cachero, T. G., Miller, S. K., Friedman, J. E., Choong, I. C., and Cunningham, B. C. (2005) Small-molecule inhibition of TNF-a. Science 310, 1022–1025. 22. Carter, P. H., Scherle, P. A., Muckelbauer, J. K., Voss, M. E., Liu, R. Q., Thompson, L. A., Tebben, A. J., Solomon, K. A., Lo, Y. C., Li, Z., Strzemienski, P., Yang, G., Falahatpisheh, N., Xu, M., Wu, Z., Farrow, N. A., Ramnarayan, K., Wang, J., Rideout, D., Yalamoori, V., Domaille, P., Underwood, D. J., Trzaskos, J. M., Friedman, S. M., Newton, R. C., and Decicco, C. P. (2001) Photochemically enhanced binding of small molecules to the tumor necrosis factor receptor-1 inhibits the

binding of TNF-alpha. Proc. Natl. Acad. Sci. USA 98, 11879–11884. 23. Margolles-Clark, E., Umland, O., Kenyon, N. S., Ricordi, C., and Buchwald, P. (2009) Small molecule costimulatory blockade: organic dye inhibitors of the CD40–CD154 interaction. J. Mol. Med. 87, 1133–1143. 24. Margolles-Clark, E., Kenyon, N. S., Ricordi, C., and Buchwald, P. (2010) Effective and specific inhibition of the CD40–CD154 costimulatory interaction by a naphthalenesulphonic acid derivative. Chem. Biol. Drug Des. 76, 305–313. 25. Chen, T., Kablaoui, N., Little, J., Timofeevski, S., Tschantz, W. R., Chen, P., Feng, J., Charlton, M., Stanton, R., and Bauer, P. (2009) Identification of small-molecule inhibitors of the JIP-JNK interaction. Biochem. J. 420, 283–294. 26. Fischer, P. M. and Lane, D. P. (2004) Small-molecule inhibitors of the p53 suppressor HDM2: have protein–protein interactions come of age as drug targets? Trends Pharmacol. Sci. 25, 343–346. 27. Do¨mling, A. (2008) Small molecular weight protein–protein interaction antagonists: an insurmountable challenge? Curr. Opin. Chem. Biol. 12, 281–291. 28. Kojima, K., Burks, J. K., Arts, J., and Andreeff, M. (2010) The novel tryptamine derivative JNJ-26854165 induces wild-type p53- and E2F1mediated apoptosis in acute myeloid and lymphoid leukemias. Mol. Cancer Ther. 9, 2545–2557. 29. Park, C. M., Bruncko, M., Adickes, J., Bauch, J., Ding, H., Kunzer, A., Marsh, K. C., Nimmer, P., Shoemaker, A. R., Song, X., Tahir, S. K., Tse, C., Wang, X., Wendt, M. D., Yang, X., Zhang, H., Fesik, S. W., Rosenberg, S. H., and Elmore, S. W. (2008) Discovery of an orally bioavailable small molecule inhibitor of prosurvival B-cell lymphoma 2 proteins. J. Med. Chem. 51, 6902–6915. 30. Bonacci, T. M., Mathews, J. L., Yuan, C., Lehmann, D. M., Malik, S., Wu, D., Font, J. L., Bidlack, J. M., and Smrcka, A. V. (2006) Differential targeting of Gbc-subunit signaling with small molecules. Science 312, 443–446. 31. Lehmann, D. M., Seneviratne, A. M., and Smrcka, A. V. (2008) Small molecule disruption of G protein beta gamma subunit signaling inhibits neutrophil chemotaxis and inflammation. Mol. Pharmacol. 73, 410– 418. 32. Sandiford, S. L., Wang, Q., Levay, K., Buchwald, P., and Slepak, V. Z. (2010) Molecular organization of the complex between muscarinic M3 receptor and regulator of G protein signaling Gb5-RGS7. Biochemistry 49, 4998–5006. 33. Roman, D. L., Talbot, J. N., Roof, R. A., Sunahara, R. K., Traynor, J. R., and Neubig, R. R. (2007) Identification of small-molecule inhibitors of RGS4 using a high-throughput flow cytometry protein interaction assay. Mol. Pharmacol. 71, 169–175. 34. Kimple, A. J., Willard, F. S., Giguere, P. M., Johnston, C. A., Mocanu, V., and Siderovski, D. P. (2007) The RGS protein inhibitor CCG-4986 is a covalent modifier of the RGS4 Ga-interaction face. Biochim. Biophys. Acta 1774, 1213–1220. 35. Overington, J. P., Al-Lazikani, B., and Hopkins, A. L. (2006) How many drug targets are there? Nat. Rev. Drug Discov. 5, 993–996. 36. Buchwald, P. (2008) Glucocorticoid receptor binding: a biphasic dependence on molecular size as revealed by the bilinear LinBiExp model. Steroids 73, 193–208. 37. Neugebauer, A., Hartmann, R. W., and Klein, C. D. (2007) Prediction of protein–protein interaction inhibitors by chemoinformatics and machine learning methods. J. Med. Chem. 50, 4665–4668. 38. Lipinski, C. A., Lombardo, F., Dominy, B. W., and Feeney, P. J. (1997) Experimental and computational approaches to estimate solubility and permeability in drug discovery and development setting. Adv. Drug Deliv. Rev. 23, 3–25. 39. Hopkins, A. L., Groom, C. R., and Alex, A. (2004) Ligand efficiency: a useful metric for lead selection. Drug Discov. Today 9, 430–431. 40. Hajduk, P. J. (2006) Fragment-based drug design: how big is too big? J. Med. Chem. 49, 6972–6976.

SMALL-MOLECULE PROTEIN–PROTEIN INTERACTION INHIBITORS

41. Reynolds, C. H., Bembenek, S. D., and Tounge, B. A. (2007) The role of molecular size in ligand efficiency. Bioorg. Med. Chem. Lett. 17, 4258–4261. 42. Wendt, M. D., Sun, C., Kunzer, A., Sauer, D., Sarris, K., Hoff, E., Yu, L., Nettesheim, D. G., Chen, J., Jin, S., Comess, K. M., Fan, Y., Anderson, S. N., Isaac, B., Olejniczak, E. T., Hajduk, P. J., Rosenberg, S. H., and Elmore, S. W. (2007) Discovery of a novel small molecule binding site of human survivin. Bioorg. Med. Chem. Lett. 17, 3122– 3129. 43. Reyne`s, C., Host, H., Camproux, A. C., Laconde, G., Leroux, F., Mazars, A., Deprez, B., Fahraeus, R., Villoutreix, B. O., and Sperandio, O. (2010) Designing focused chemical libraries enriched in protein–protein interaction inhibitors using machine-learning methods. PLoS Comput. Biol. 6, e1000695.

731

44. Sperandio, O., Reyne`s, C. H., Camproux, A. C., and Villoutreix, B. O. (2010) Rationalizing the chemical space of protein–protein interaction inhibitors. Drug Discov. Today 15, 220–229. 45. Hajduk, P. J., Bures, M., Praestgaard, J., and Fesik, S. W. (2000) Privileged molecules for protein binding identified from NMR-based screening. J. Med. Chem. 43, 3443–3447. 46. Hajduk, P. J., Huth, J. R., and Tse, C. (2005) Predicting protein druggability. Drug Discov. Today 10, 1675–1682. 47. Arkin, M. R. and Whitty, A. (2009) The road less traveled: modulating signal transduction enzymes by inhibiting their protein–protein interactions. Curr. Opin. Chem. Biol. 13, 284–290. 48. Stumpf, M. P., Thorne, T., de Silva, E., Stewart, R., An, H. J., Lappe, M., and Wiuf, C. (2008) Estimating the size of the human interactome. Proc. Natl. Acad. Sci. USA. 105, 6959–6964.