Small‐Molecule and Peptide Inhibitors of the Pro‐Survival Protein Mcl‐1

Abstract The ability of protein–protein interactions to regulate cellular processes in both beneficial and detrimental ways has made them obvious drug targets. The Bcl‐2 family of proteins undergo a series of protein–protein interactions which regulate the intrinsic cell‐death pathway. The pro‐survival members of the Bcl‐2 family, including Bcl‐2, Bcl‐xL, and Mcl‐1, are commonly overexpressed in a number of human cancers. Effective modulators of members of the Bcl‐2 family have been developed and are undergoing clinical trials, but the efficient modulation of Mcl‐1 is still not represented in the clinic. In addition, Mcl‐1 is a major cause of resistance to radio‐ and chemotherapies, including inhibitors that target other Bcl‐2 family members. Subsequently, the inhibition of Mcl‐1 has become of significant interest to the scientific community. This review covers the progress made to date in modulating the activity of Mcl‐1, by both stapled peptides and small molecules. The development of peptides as drug candidates, and the advancement of experimental and computational techniques used to discover small molecules are also highlighted.


Introduction
Protein-protein interactions (PPIs) control many important physiological processes within human cells. [1] The ability of specific proteins to interactw ith high specificity and affinity has been observed in processes both beneficial and detrimental to health. Additionally,c hanges in these interactions can lead to specific cellular processes malfunctioning, potentially resultingi nb iologically undesirablee ffects, for example the ability of cancerc ells to avoid apoptosis. Them odulation of PPIs has the potentialt or eturn physiological processes to healthys tates,a nd as such there is much interest in the development of agents which will allow such control. [2] This task, however, is not atrivial one, with PPIs historically being considered "undruggable" because of their large and shallow interfaces. [3] In recent years new approaches have been developed to effectively target PPIs using both computational and experimental approaches. [3b, 4] This has resulted in the preparation of an umber of small-molecule modulators, with more than 12 candidates currentlyi nc linical trials. [5] Proteins of the Bcl-2 family regulate the intrinsic mitochondrial cell death pathway with the familyc onsisting of both pro-and anti-apoptotic members. [6] Over the past decade there has been significant interesti nt he PPIs of the Bcl-2 family due to the role they playi na poptosis (we refer the reader to reference [7b] for ad etailed discussion of the Bcl-2 family proteins). [7] The ability of cancerc ells to avoid apoptosis plays al arge role in tumour progression and drug resistance. [7a] The Bcl-2 family consists of both pro-apoptotic proteins, antiapoptotic proteins and regulators. The pro-apoptosis proteins include BAXa nd BAK, and are triggered through their interaction with activating regulators BIM, BID, and PUMA. [8] The interaction between the pro-apoptotic proteins and regulators is ak ey elemento fc ell death.A nti-apoptotic proteins such as Bcl-2, Bcl-x L and Mcl-1 prevent activation of BAX or BAK by binding the activating regulators or the pro-apoptosis proteins themselves, keeping cells alive. [9] This prevents BAX and BAK from oligomerising and puncturing the outer mitochondrial membrane resulting in the release of cytochrome Ci nto the cytoplasm. This is ar apid and irreversible process which activates downstream caspases and is the point of no return for the cell. The anti-apoptotic proteins are modulated by sensitiser regulators including BAD and NOXA, which do not activate BAX or BAK, but preferentially bind to Bcl-2, Bcl-x L and Mcl-1, deactivating them. The overexpression of anti-apoptotic proteins is commonly observed in an umber of human cancers, resulting in cell survival. [10] These proteins share regions of homology named Bcl-2-homology (BH) domains. The regulating proteins are BH3-only proteins (meaning they only possess homology in the BH3 region) and are induced through cellular stress death signals. These proteins bind through hydrophobic and electrostatic interactions between the BH3 regiona nd the binding groove formed by the BH1, BH2 and BH3 regions of the pro-and anti-apoptotic proteins. [11] Candidates are already in clinical trials which inhibitB cl-2 andB cl-x L and have been shown to induce apoptosis as BH3 mimetics (see reference [12] for ar ecent review). However,t he inability to inhibita ll anti-The ability of protein-protein interactions to regulate cellular processesi nb oth beneficiala nd detrimental ways has made them obvious drug targets. The Bcl-2 family of proteins undergo as eries of protein-protein interactions which regulate the intrinsic cell-death pathway.T he pro-survival members of the Bcl-2 family,i ncludingB cl-2, Bcl-x L ,a nd Mcl-1, are commonly overexpressed in an umber of human cancers. Effective modulators of members of the Bcl-2 family have been developed and are undergoing clinicalt rials, but the efficient modulation of Mcl-1 is still not represented in the clinic. In addition, Mcl-1i samajor cause of resistance to radio-and chemotherapies, including inhibitors that target other Bcl-2 family members. Subsequently,t he inhibition of Mcl-1 has becomeo fs ignificant interestt ot he scientificc ommunity.T his review coverst he progress made to date in modulating the activity of Mcl-1, by both stapled peptides and small molecules. The development of peptides as drugc andidates, andt he advancement of experimental and computational techniquesu sed to discover small molecules are also highlighted. apoptotic proteins, in particular Mcl-1, has been shown to result in drug resistance. [7a] Despite being one of the most frequently amplified genes in cancera nd being am ajor factor in resistance to chemotherapy,M cl-1 hasp roven the most elusive, with no candidates currently in clinicalt rial. The pro-survival Bcl-2 family proteins sequester the a-helicalB H3 domain of the pro-death Bcl-2 family,i nt he binding groove createdb y the BH1-3 domains. However,M cl-1 differs from the other prosurvival members, possessing am ore electropositiveb inding groove, [8] with an umber of different residues. [13] Additionally, the Mcl-1 groove has been shown to be more rigid than the other pro-apoptosis members, [14] making specific modulators difficult to obtain.
Severale xamples of small molecules which modulate Mcl-1h ave been reported, with av ariety of reviewsa ddressing patented compounds, [15] compounds in clinicalt rials [8] and pan-Bcl-2 inhibitors. [16] Most recently, Belmara nd Fesik presented an excellent review of Mcl-1 bindersw hich are known to be BH3 mimetics. [17] The scopeoft his review will target ac omprehensiveo verview of known Mcl-1 binders which have been demonstrated to inhibit the PPIs of Mcl-1 with the pro-apoptotic and regulating proteins from the Bcl-2 family,c omplementing the work which has already been presented before us.

Stapled Peptides
In general terms there are two major classes of approved drugs:s mall molecules andp rotein therapeutics. [18] As mallmolecule inhibitor of aP PI is generally much smaller than the protein it is inhibiting, therefore limiting the number of interactions it can make. However,t he presence/requirement of ah ydrophobic groove or pocket enables the design of small molecules capable of inhibiting the protein of interest, [19] and small-molecule inhibitors of Mcl-1 are discussed in detail below.B yc ontrast, protein therapeuticsa re much larger and have ag reater surface area in which to make contact with the target. Therefore, they do not need such defined binding pockets and can effectively bind to flatter,s hallower surfaces which are not necessarily hydrophobic.However,t his class of drug is unable to cross the cell membrane, so is therefore limited to extracellular targets. An alternative and attractive option for stabilising or disrupting PPIs are peptides. [20] However, in vivo their efficacy is compromised due to al oss in secondary structure as well as poor cellular uptake and susceptibility to proteolysis. Apromisingsynthetic approacht oo vercome these limitations is to "staple"t he peptidet of ix its orientation. [21] In addition to being ap otential new class of therapeutics capable of inhibiting the Bcl-2 proteins,t hey may also prove useful chemical biology tools for probing these key proteins and furthering our understanding of the processes governinga poptosis.
Peptide staplingw as first introduced by Verdine and co-workers in 2000 [22] who then went on to employ the concept to identify peptide-based inhibi-tors of Bcl-2 proteins designed on the structure of the BH3 domain of the BID protein (Table 1, Figure1), so called "stabilised a-helix of BCL-2 domains" (SAHBs). [20] BID SAHB A was able to specifically activate the apoptosis pathway in leukaemia cells and inhibit the growth of leukaemiax enografts in vivo. Further characterisation of the peptided emonstrated it bound to and activated BAX directly. [23] Following this pioneering work av ariety of BH3-only peptided omains were used to synthesise stapled peptidescapable of modulating apoptosis. [24] In 2010L oren Walensky'sg roup identifiedt he first stapled peptidec apable of selectively inhibiting Mcl-1 (Table 1). [25] The development of such selectivei nhibitors is am ajor challenge due to the very subtle differences betweent he Bcl-2 family proteins in terms of their hydrophobic grooves. As eries of stapled peptides based on the BH3 domains of both pro-and anti-apoptotic Bcl-2 proteins were designed and assessed for their ability to inhibit Mcl-1 by fluorescencep olarisation (FP) assay.T hose that bound Mcl-1 with ah igh affinity were further evaluated against ap anel of anti-apoptotic targets and the Mcl-1 BH3 helix was found to be ap otent and selectivei nhibitor of Mcl-1 capable of sensitising cells to caspase-dependent apoptosis. Alanine scanning,s ite-directed mutagenesis and staple scanning were used to determine the key binding sites and specificity determinants as well as the optimal helicity.I nterestingly,X -ray crystallography and mutagenesis studies revealed the hydrocarbon staple itself is capableo fm aking additional hydrophobic interactions, which along with the en- Table 1. SAHB peptides reported by Verdine [20] and Walensky. [25] Compound Sequence [  hanced a-helicity may be responsible for the enhanced binding affinity of Mcl-1 SAHB D .T his was later supported by molecular dynamics simulations. [26] The Walensky group then went on to use theses tapled peptides, specifically Mcl-1 SAHB A as as creening tool to identify small molecules (described below). [27] In 2012 Lin andc o-workersreported the designo fp roteolytically stable and cell permeable peptide-basedi nhibitors of Mcl-1 ( Table 2). [28] Rather than employing ah ydrocarbon crosslink as used by Verdine and Walensky,ab is-aryl staple was incorporated into aN oxa peptidew hich binds to Mcl-1 with high affinity and selectivity.T wo solvent-exposed i and i + 7 residues were replaced with d-o rl-cysteine and the peptide subjected to 4,4'-bis(bromomethyl)biphenyl (Bph)-mediated cross-linking. [29] The resulting cross-linked peptides possessed enhanced inhibitory activity when compared to the parent Noxa peptidew hilst still maintaining selectivity.H owever,t he compounds were inactive in cellular assays due to poor uptake. Replacement of three solvente xposed, positively chargedr esidues (Arg6, Arg14, andL ys16) improved affinity (roughlyt wofold increase) but more importantly led to increasedc ellular activity.N -Methylation of both N-terminal alanine residues led to improved activity in both the FP assay as well as in the cell viability assay.F luorescence-activated cell sorting (FACS) analysis of fluorescently tagged versions of the peptides confirmed increased cellular uptake and confocal microscopy showed the peptides were predominantly localised in the cytosol. Finally the stapled peptides weres hown to possess enhanced stability towardp roteases.
In 2014, Lin went on to report the effects of the flexibility/rigidity as well as the length of the cross-linker; [30] suggesting am ore flexible cross-linker may allow the peptidet oa dopt am ore favourable/active conformation when bound to its protein partner.Aseries of aryl and vinylaryl cross-linkers with varying linker length, rigidity and hydrophobicity were synthesised (again using the Noxa BH3 peptide) and the compounds evaluated for inhibitory Mcl-1 activity using ac ompetitive FP assay.A nalogues containing a6 ,6'-bis(bromomethyl)-3,3'-bipyridine( Bpy) or p-phenylene-3,3'-bis(allylbromide) cross-linker appeared to have the highest helicity which translated into the highest inhibitory activity in the FP assay,w hereas cellular uptake correlatedw ith hydrophobicity with the ana-logues containing the Bph staple or a3 ,3'-bis(bromomethyl)biphenylvariant showingt he most efficient cellular uptake.
Ta ken together these results suggest the stapling of peptides is ap romising new approach to targeting PPIs, specifically Mcl-1 and other Bcl-2 family proteins. The bioactive conformation of the peptide is maintained whilst overcoming the drawbacks of using peptides as drug molecules;c areful positioning and choice of staple can result in ah igh-affinity binder with improved cellular uptake and stability.

Antimycin A
In 2001 Tzung and co-workersreported that antimycin A, anatural product antibiotic isolated from a Streptomyces sp.,m arkedly increased the apoptosis induced cell death in cell lines possessing Bcl-x L associated multidrug resistance. [31] Through docking studies, they showedt hat the binding groove of the Bcl-2 family proteins was occupied by antimycin, confirmed by their fluorescencea ssays on Bcl-2 and Bcl-x L .R eed and coworkers subsequentlyd emonstratedt he ability of antimycin A to competitively bind to Mcl-1 at similar concentrations to Bcl-2a nd Bcl-x L (IC 50 = 2.51 mm,FITC-Bid BH3-only peptide). [32] 3.2 BH3I-1 Yuan and co-workersh ighlighted three compounds from ac ompetitive FP binding assay of the Bcl-x L -BH3 site and Bak. These three compounds, titled BH3I-1( BH3 Inhibitor-1), BH3I-1' and BH3I-2, induced apoptosis in JK cells, showing the characteristicf eatures of over-expressiono fp ro-apoptotic Bcl-2 family proteins. [33] Yuan showed that BH3I-1 acts by preventing the heterodimerisation of the pro-apoptotic and anti-apoptotic Bcl-2 proteins,a nd identified the bindings ite throughN MR studies. Reed demonstrated that BH3I-1 is ac ompetitive Mcl-1b inder with an IC 50 of 2.17 mm (FITC-Bid BH3-only peptide). [32] 3.3 BH3M6 In 2002 Hamilton and co-workers synthesised an umber of compounds which were designedt om imic the binding residues of the Bcl-2 family proteins. [34] Using the crystal structure of Bak/Bcl-2t hey identified an umber of hydrophobic residues Table 2. StapledNoxa peptides described by Lin and co-workers. [28] Compound Sequence [ [b] Data are the mean AE SD, n = 3.
ChemMedChem 2016, 11,802 -813 www.chemmedchem.org which weres hown to participate in binding by alanine scanning. This structure guided design resulted in the preparation of BH3M6. Hamilton demonstrated that BH3M6 was able to inhibit the binding of Bak and Bcl-x L with a K d = 1.89 mm.I n2 011 Sebti and co-workers demonstrated that BH3M6 inhibits the binding of Mcl-1 to Bax and Bim in ad ose-dependentm anner in HEK293T and A549 cells. [35] Sebti also demonstrated that BH3M6 can induce apoptosis by acting as ap an-inhibitor and disrupting the formation of Bcl-x L /Bim, Bcl-2/Bim and Mcl-1/ Bim heterodimers.

YC137
In 2001 Wang and Yang performedavirtual screen of 206 876 compounds based on an NMR structure of Bcl-2, and then performed as ubsequent in vitro assay which highlighted 35 of these structures. [36] In 2004 Wang proceeded to report the exploration and elaboration of these structures,w hichh ighlighted YC137 as ap otent Bcl-2 inhibitor. [37] Reed again showed this compound to be as trong Mcl-1 binder, demonstrating an IC 50 of 2.47 mm in their FP assay (FITC-Bid BH3-only peptide). [32] Despite the competitive binding ability,a nd the recogniseds electivity of YC137 for Bcl-2 family members, the structure has received little attention in the literature regardingt he activity toward Mcl-1.

EGCG
There has been much attentioni nt he literature regarding the anticancer ability of compounds found commonly in green and black tea. [38] In 2003 Pellecchia and co-workersr eported the examination of an umber of natural compoundsf ound in significant quantitiesi ng reen and black tea, with ap articular interesti nt heir ability to bind the Bcl-2 family.T hey showed that (À)-epigallocatechin gallate (EGCG)h ad binding affinities in the nanomolar range for Bcl-2 (K i = 490 nm)a nd Bcl-x L (K i = 335 nm,F ITC-BAD-BH3-only peptide competitive binding assay). [39] Computational docking studies revealed that EGCG bound to the BH3 domain.R eed and co-workersd emonstrated that EGCG was ap an-inhibitor of the Bcl-2 family. [32] Recently, it has been shown that in several cancer cell lines (786-O renal cell carcinoma, HNSCC, Pc-3 and LNCaP), EGCG binds upstream of the Bcl-2 family,r esultingi nd ownregulation of the antiapoptosis proteins. [40] 3.6 (À)-Gossypol (AT-101) In 2002 Wang and co-workersp atentedt he use of (À)-gossypol (AT-101) as aB cl-2 family antagonist, after performing as tructure-based database screen. [41] Gossypol was shown to bind to Bcl-x L and Bcl-2 with high affinity (320 nm and 480 nm respectively), [42] as well as to Mcl-1 (180 nm,F ITC-Bad-BH3-only peptide). Subsequently,g ossypol was advanced to clinical trials as as mall-molecule inhibitor of Bcl-x L ,B cl-2 and Mcl-1. [43] In 2004 Bradford and co-workersr eportedt hat (À)-gossypol was acting as aB H3 domain mimetic, making it ap an inhibitoro f the Bcl-2 family of proteins. [44] Subsequently,( À)-gossypol was shown to delay the onset of androgen-independent prostate cancer in vivo, [45] chemosensitise prostate cancerc ells (PC-3) to docetaxel both in vitro and in vivo, and wasp rovent oi ncrease the availability of the pro-apoptotic proteinsP uma and Noxa. [46] 3.7 TW-37 In 2006 Wang et al. reportedt he structure-based design of an ew series of Bcl-2 family binders, based on the exploration of the structure activity relationship of gossypol. [47] The Bcl-2 protein was targeted, with computational docking calculations, to examine whichf unctionalg roups played the largest role in gossypol'sa ffinity to Bcl-2. It was shown that the polyphenol system was of high importance,a nd the hydrophobic section of the molecule could be extendedt ob etter fit into the binding pocket.A fter as eries of iterations Wang and co-workers prepared TW-37, which showedahigh affinity for Bcl-2 (K i = 290 nm)a nd Mcl-1 (K i = 260 nm), but lower affinity for Bcl-x L (K i = 1110nm,F AM-Bid-BH3-onlyp eptide competitive binding assay). Wang and co-workersw ent on to showt hat TW-37 in-ChemMedChem 2016, 11,802 -813 www.chemmedchem.org hibits cell growth in the prostate cancerc ell line PC-3 with an IC 50 value of 200 nm,a nd inducesa poptosis in 89.5 %o fc ells at 5 mm. [47] In 2007 Mohammad and co-workers demonstrated that TW-37 resulted in the inhibition of cell proliferation of lymphoma cells (WSU-DLCL 2 290 nm)a nd induced apoptosis in 71.4 %o f cells at 400 nm. [48] Furthermore, they showedt hat TW-37 enhances the efficacy of the four drug combination cyclophosphamide-doxorubicin-vincristine-prednisolone (CHOP) in mice models, displaying significant decrease in tumour weight relative to CHOP alone, TW-37a loneo rt he control. [48] These results were similarly reflected in the examination of B-cell tumourc ell lines by Katib and co-workers. [49] 3.8 Apogossypolone (ApoG2) Apogossypolone (ApoG2) is ag ossypol derivativep repared by Mohammad and co-workers in 2008. [50] It was shown to possess low nanomolar affinity for Bcl-2 (K i = 35 nm)a nd Mcl-1 (K i = 25 nm,c ompetitor not published). Examinationo ft he activity of apogossypolone in Follicular Lymphomac ells (WSU-FSCCL)s howed an IC 50 of 109 nm,a nd decreased cell numbers in af resh lymphoma sample. Apogossypolonew as demonstrated to prompt apoptosis, with an activation of the apoptosis inducing factor,s uggesting that apogossypolonew as acting as aBH3 mimetic.

Obatoclax
Obatoclax was developed by Gemin X, [15] ad evelopmental program focussed on modulating the anti-apoptototic proteins of the Bcl-2 family. [51] It was demonstrated to be ap an-Bcl-2 inhibitor by Reed and co-workers, [32] but was shown by Shore to be more potent than the original FP assay suggested, due to the low solubility in aqueous medium. [51] Shore went on to display that obatoclax overcomes resistance to apoptosis conferred by Mcl-1, restoring sensitivity to the known Bcl-2 antagonist ABT-737, by increasing the presence of Bim in the cell.A d-ditionally, obatoclax displayed single-agent antitumour activity in multiple standardm ouset umour models. [51] Obatoclax was shown to bind tightlyt oM cl-1 by inducing ah istidine (His252) side chain hydrogen bond,i nanarrow groove of the binding site, which is conserved in several Bcl-2 familym embers. [52] Obatoclaxw as also found to increaset he activity of cisplatin in both resistant and sensitive cell lines, overcoming platinum resistance in the former and restoring mitochondrial apoptosis. [53] Obatoclax has been investigated in severalp hase I/II clinical trials, both as as ingle agent and in combination. However to date, the data available suggests al ow therapeutica ctivityo f obatoclax. [12] 3.10 S1 Zhang andQ ian reported in 2007 the designo fc ompounds termedS 1a nd S2, and demonstrated them to be cytotoxic compounds. [54] Flow cytometry showed that S1 was inducing apoptosis in vitro, and subsequent analysis showed that this was also the case in vivo. Examining the process upstream, Zhang showedt hat S1 bound to Bcl-2, via NMR-basedb inding assays, and demonstrated through computational docking that S1 and S2 were bindingi nt he BH3 domain. S2 was shown to have an IC 50 of 1.3 mm to HL-60 cellsa nd S1 to have an IC 50 of 2.8 mm,r eproducing their computational predictions. [54a] Zhang reported in 2011t hat S1 was also an Mcl-1 inhibitor,p reventing the Mcl-1/Bak dimerisation, claiming it to be the first reported "authentic BH3 mimetic". [55] In 2013 Zhang demonstrated that S1 shows activity towardavariety of leukaemia cell lines, including acutel ymphoblasticl eukaemia, acute myeloid leukaemia,c hronic lymphoblastic leukaemia and chronic myeloid leukaemia. [56] While S1 inhibited Mcl-1 and Bcl-2 as expected, the protein Bcl-2 did provide resistance to S1, and the ratio of these proteins allowedf or the prediction of the cytotoxicity of S1. Elaboration of S1 and S2 led to product 1,w hich showeda9-to 35-fold higher affinity for Mcl-1, Bcl-2 and Bclx L than S1, displaying IC 50 values of 10, 20, and 18 nm,r espectively (ELISA). [57] ChemMedChem 2016, 11,802 -813 www.chemmedchem.org

3.11M eiogynin Ad erivatives
In 2009 Gueritte and co-workersisolated meiogynin Af rom the bark of Meiogyne cylindrocarpa,aMalaysian plant, and demonstrated it to be aB cl-x L inhibitor. [58] Subsequently,i n2 014 Roussi and co-workersr eported the preparation of an umber of meiogynin Ad erivatives, designed by as tructure activityr elationship study,w hich was reported at the time to be "one of the most potent dual inhibitors" of Mcl-1 and Bcl-x L . [59] The most effective compound,t itled Roussi compound 2a, possessed K i values of 106 nm for Mcl-1 (Bid-BH3 only FP assay) and 153 nm for Bcl-x L (Bak-BH3o nly FP assay). In addition, the selectiveM cl-1 inhibitor Roussi compound 2c showed K i values of 460 nm for Mcl-1 (Bid-BH3o nly FP assay) and > 23 mm for Bcl-x L (Bak-BH3o nly FP assay).

Sabutoclax
In 2009 Pellecchia and co-workers examined the derivatisation of gossypol guided by NMR binding assays and computational docking. [60] From the library of compounds Pellecchia prepared, racemic BI97C1 (originally compound 8r) was the most effective, demonstrating IC 50 values of 0.28 mm for Mcl-1a nd 0.32 mm for Bcl-2 (FITC-Bak-BH3-only peptide), as well as binding to Bcl-x L and Bfl-1.A dditionally,B l97C1 was shown to induce apoptosisi nt he RS11846 human lymphoma cell line. In af ollow up publication, Pellecchia et al. examined the stereochemistry of Bl97C1, demonstrating that the (R,-,R) enantiomer (displayed) inhibits the binding of BH3 peptides to Bcl-x L ,B cl-2, Mcl-1, and Bfl-1 with IC 50 values of 0.31, 0.32, 0.20, and 0.62 mm (FITC-Bak-BH3-onlyp eptide). [61] Additionally,i nhibition of the cell growth of human prostate cancer, lung cancer,a nd BP3 Bcell lymphoma cell lines with EC 50 values of 0.13, 0.56, and 0.049 mm,respectively wasdemonstrated.

Marinopyrrole Aderivatives
Marinopyrrole Ai sanatural product antibiotic isolated from a Streptomyces sp. reported in 2009. [65] It was described by Wang and co-workerst op ossess the ability to overcome ABT-737 resistance by inhibiting the action of Mcl-1. [66] However,i t was subsequently reported to have no effect on Mcl-1i nc ells, and exhibits the same action on Bcl-2-dependent cells (2 mm) as Mcl-1-dependent cells (2.5 mm). [67] In 2015 Qin and co-workers reported the preparation of an umber of analogueso fm arinopyrrole A, followingastructurea ctivity relationship study. [68] The talismanic compoundso ft his study were titled Qin com-

Chai compounds 6&7
In 2010 Chai and co-workers addressed the problem of identifying as elective Mcl-1 inhibitor. [69] All Mcl-1 inhibitors previously reported had been pan-Bcl-2 inhibitors, with moderate to weak activity,m aking overcoming Mcl-1 induced resistance ac hallenge. Chai screened al ibrary of compounds which incorporated knownB cl-2 family binders BH3I-1 and sanguinarine with aFPa ssay using Mcl-1and Bak. [70] This screening highlighted two compounds, referred to in the originall iterature as compounds 6a nd 7. These constitutionali somers showeds ignificantd ifferences in selectivity,w ith Chai compound 6d isplaying bindingi nt he micromolar range to both Bcl-x L (K i = 3.7 mm)a nd Mcl-1 (K i = 6.9 mm), and Chai compound 7s howing selectivity toward Mcl-1 (K i = 8 mm)w ith no binding to Bcl-x L (K i > 100 mm,F lu-Bak-BH3 peptidec ompetitive binding). [69] Both compounds showed greater bindingi nt heir assays than BH3I-1, with NMR studies demonstrating that the compounds were binding in the BH3 domain. Docking studies demonstrated that the Mcl-1 binding groove is wider than that of Bcl-x L , which maye xplain the selectivity being displayed by the two constitutional isomers. [69] 3.15 MIM1 In 2010 Walensky and co-workers showedt hat the BH3 domain of Mcl-1 was ap otent and selective natural inhibitor of Mcl-1. [25] This prompted the use of the fluorescently tagged BH3 domain of Mcl-1a st he competitive binding agent for their FP assays, allowing for as elective and potent Mcl-1 binder to be identified. 71 296 compounds were screened for the ability to displace aF ITC tagged Mcl-1 BH3 domain peptide, and stringent selection processes highlighted MIM1 as ap otent and selectiveM cl-1 binder. [27] MIM1 displaced the FITC-Mcl-1-BH3 peptide at an IC 50 of 4.7 mm,b ut had no significant ability to displace Bid from Bcl-2, complementing the activity of ABT-737. The ability of MIM1 to induce apoptosis in Mcl-1-dependent leukaemia cells was also demonstrated.

Fesikfragment-based screening
In 2013 Fesik and co-workersd escribed the use of an ow commonly employed technique to targetp rotein-protein interactions. They screeneds mall fragments for micromolar binding, and then using NMR studies identified the binding sites of each fragment. With this information in hand each fragment can be linked together to create as mallm olecule whichp ossesses greater binding affinity.I nt his way the relativelyl arge protein-protein binding site can be explored rapidly to identify the important bindingm otifs. [74] Employing this method, Fesik was able to identify two important regions, which were bound to by distinct small molecules, exemplified by 4 (K i = 131 mm) and 5 (K i = 60 mm,F ITC-Mcl-1-BH3-only peptide). Merging these two fragments generated 6 (K i = 0.32 mm,F ITC-Mcl-1-BH3-only peptide), which was confirmed by crystallography to occupy both of the pockets identified by the initial fragments. Subsequently this molecule underwent structure-guided synthetic design,w hich yieldeds everal high-affinity leads, exemplified by 7 which showed a K i = 0.055 mm,w ith 10-fold selectivity for Mcl-1 over Bcl-2 and Bcl-x L (FITC-Mcl-1-BH3-only peptide). Fesik also described several other potential leads with the same skeleton. [74] 3.19 Benzylpiperazine derivatives In 2013 Liu, Wang, and colleagues reported the use of an innovative computational modelling approacht odevelop four molecular scaffolds for synthetic evaluation. [75] Liu and Wang apply the assumption that critical binding residues of the PPI exist in small clusters, which can be exploited by small fragments.U sing computational docking, one can examinet hese small clusters to identify the most appropriate fragments, and then combine these virtually,w ith the appropriate chemical spacingt oc reate molecular scaffolds. In this manner Liu and Wang developed four backbones, termed series A-D, exemplified by the best performers in each series, A1, B3, C10 and D14. TheAa nd Ds eries weref ound to outperform series B and Ci nt heir FP binding assay, with A1 demonstrating a K i = 0.18 mm and D14 showing a K i = 0.32 mm (Mcl-1/5-FAM-Bid BH3-only peptide competitive binding assay). These com- ChemMedChem 2016, 11,802 -813 www.chemmedchem.org pounds also showed some selectivity forM cl-1, with A1 showing no appreciable binding to Bcl-2 or Bcl-x L (5-FAM-Bid BH3only peptide). [75] 3.20 UMI-77 In 2014 Nicolovska-Coleska performed ah igh-throughput screen to identify small-molecule inhibitors of Mcl-1. [76] They identifiedc ompound 8 which led to the developmento fUMI-77 throughs tructure-based chemical modifications.W ith FPbased screensN icolovska-Coleska showed that UMI-77 displaces Bid-BH3 from Mcl-1 with a K i = 0.49 mm,w ith a1 0-fold selectivity, with Bcl-w,B cl-2, and Bcl-x L displaying K i > 5 mm (5-FAM-BidB H3-only peptide).I tw as also demonstrated that UMI-77 selectively binds to Mcl-1 in preference to Noxa in ad ose-dependent manner,b yb inding to the BH3-binding pocket of Mcl-1. In cell lines, UMI-77 wass hown to inhibit the growth of pancreatic cancerc ells (MiaPaCa-2 and AsPC-1) and showeds ingle-agent antitumour activity towardB xPC-3 cells. Morgana nd co-workerst hen demonstratedt hat UMI-77 radiosensitised cancer cell lines BxPC-3 and Panc-1, but did not radio-sensitise normals mall intestinal cells. [77] Morgan demonstrated that ABT-737, whichd oes not bind to Mcl-1, did not perform in the same manner,s uggesting Mcl-1 plays ak ey role in radio-sensitising cancercells.

A-1210477
Souers andc o-workers recently identified 9 through ah ighthroughput screen aimed at identifying small-molecule inhibitors of Mcl-1 using the selective protein Noxa. [78] 9 showedm icromolar affinity for Mcl-1 (K i = 4.4 mm), as well as selectivity over Bcl-x L (K i > 10 mm), in an FP assay.S tructure guided design,i ncorporating hits which they had generated in previous reports, [79] allowed for the preparation of A-1210477, which showeda10 000-fold improvement in affinity for Mcl-1 (K i = 0.43 nm,N oxa BH3-only peptide) compared to 9.A -1210477 also showed similars electivity for Mcl-1 with K i > 0.66 mm for Bcl-2, Bcl-x L ,B cl-w and A1. Activity was provent o be through the disruption of the Mcl-1 Bim complex in live cells. A-1210477 was found to induce apoptosis in the cancer cell line H929 and restore navitoclax sensitivity in the resistant pancreatic cell line BxPC-3. [80] Similarly in the breast cancer cell line SKBR3 sensitivity to navitoclax was restored with the addition of A-1210477 resulting in the release of BAK, demonstrating the role of Mcl-1 in navitoclax resistance. [81] The molecule is believed to bind to Mcl-1 in as imilarm anner to analogue 10 reported in the same manuscript (where the sulfonamideo n the piperazine was replaced with an acetylg roup);s pecifically bindingi nt he BH3 binding domain and mimicking the BIM BH3 peptide. The central indole binds in the p2 pocket, with the carboxylate group forming ahydrogenbondinginteraction with Arg263. The naphthyl group wasp rojected towardt he p1 pocket and the extended piperazine spanned the p3/p4 pockets. A-1210477 represents the most potent small-molecule inhibitor described in the literature to date, and induces clear on-target cellulara ctivity.

Pyridoclax
In 2015 Voisin-Chiret and Poulain employed as trategy similar to that whichl ead to the discovery of BH3M6. [82] Examining the bindingp ocket of Mcl-1, andi ts differencesf rom the other Bcl-2 family members, led them to employ an oligopyridine backbonet oa ct as aB H3-mimetic.C omputational modelling and SAR studies resulted in the preparation of pyridoclax, which was designed based on the binding of the selective Mcl-1 binding protein Noxa. Using bioluminescence resonance energy transfer (BRET)i tw as shown that pyridoclax binds to

Summary and Outlook
The discovery of chemical entitiesc apable of modulating PPIs is ac hallenging prospect and PPIs have historically been described as "undruggable". However, the efforts of many research groups over the past two decades has defied this view point with an umber of inhibitors progressing through clinical trials which target the Bcl-2 family( ABT-737, navitoclax and ABT-199). [73,83] However, despite the recent success in targeting Bcl-2 and Bcl-x L, the modulation of Mcl-1 with inhibitors has provene lusive and there are currently no Mcl-1 inhibitors in clinicalt rial despite the factt hat Mcl-1 is one of the most commonly amplified genes in cancer. [84] Perhaps more promisingly,r ecent years have seen an increase in interesti nt argeting Mcl-1 and identifyingn ew inhibitors-both small-molecule and peptide based. Specifically,s tapled peptides are ap romisingn ew approach to developing moieties capable of inhibiting PPIs and the examples described above by Walensky [20] and Lin [28] demonstrate the potentialo f this class of therapeutic.T hrough fixingt he conformation of the a-helix, the active conformation of the peptidei sm aintained leading to an increase in affinity for the protein.I na ddition, stapling the peptidec an shield the amide bonds of the peptides protecting them from proteases and also decrease their ionic character,r esulting in an improvement in cellular uptake (through passive diffusion) andr educed clearance from the body.W ith the number of publications reporting stapled peptides accelerating( just one in 2000 vs 16 in 2013) and the first clinicalt rial successfully completed in 2013, the clinical potential of stapled peptides is currently being explored but it remains to be seen if this will translate into the clinic and if stapled peptides are able to fulfilthis potential.
The number of publications describing small-molecule inhibitors of Mcl-1 has increased rapidly in recent years. However, al arge proportion of the small molecules described have been identified through high throughput screening of larger libraries, resulting in large attrition rates. In addition achieving selectivity has provedp roblematic due to the very subtle differences in binding pockets between the Bcl-2 family;t he p2 pocket of Mcl-1 is dynamic compared to Bcl-2, whereas the p4 pocket is less well defined, shallower and less hydrophobic than Bcl-x L. [85] In fact most of the early compounds reported were pan Bcl-2 inhibitors with the first selectiveM cl-1 inhibitor reportedi n2 010. [69] There is also al ack of in vivo data for the majority of compounds reported in the literature, although this has been addressed very recently by A-1210477, where the authors argue this is the first description of as mallm olecule with sufficient potencyt oh ave ac lear on-target cellular effect. [80] Despite these challenges, advances have been made and the outlook is promisingf or the discoveryo fM cl-1 inhibitors. New methodologies are being employed to target PPIs which are provinge ffective and increasing the hit rate of targetbased screening. The computationala pproache mployed by Liu and Wang decreases screening time significantly. [75] The fragment-based approach employed by Fesik allows for smaller fragment libraries to be evaluated which quicklyg enerates information about effective ways to target the binding grooves of the PPI. [74] The recent work of Souers provides ap ositive outlook,w ith potent and selective Mcl-1 inhibitors being developed through the use of selective assays. [78,79] The exploitation of ac ombination of these techniques maya llow for the highly efficient development of drug candidates. The examples of Mcl-1 inhibitors described here demonstratet hat the "undruggable" challenge that PPIs present can be tackled effectively and is likely to lead to an ovel inhibitors for the treatment of cancer.