Current Inhibitors of Cyclin-Dependent Kinases as Anti-Cancer Therapeutics … · 2015-02-27 ·...

Current Medicinal Chemistry, 2000, 7, 1213-1245 1213

Inhibitors of Cyclin-Dependent Kinases as Anti-Cancer Therapeutics

Peter M. Fischer* and David P. Lane

Cyclacel Limited, Dundee Technopole, James Lindsay Place, Dundee DD1 5JJ, Scotland, UK

Abstract: Initiation, progression, and completion of the cell cycle are regulated byvarious cyclin-dependent kinases (CDKs), which are thus critical for cell growth.Tumour development is closely associated with genetic alteration and deregulation ofCDKs and their regulators, suggesting that inhibitors of CDKs may be useful anti-cancer therapeutics. Indeed, early results suggest that transformed and normal cellsdiffer in their requirement for e.g. cyclin/CDK2 and that it may be possible to develop novelantineoplastic agents devoid of the general host toxicity observed with conventional cystostatic drugs.Numerous active-site inhibitors of CDKs have been studied; the main limitation with these ATPantagonists is kinase specificity for CDKs. However, screening of compound collections, as well asrational design based on enzyme-ligand complex crystal structures, are now yielding pre-clinicalcandidates, particularly certain purine and flavonoid analogues, with impressive potency and selectivity.Natural CDK inhibitors (CKIs), e.g. the tumour suppressor gene products p16INK4, p21WAF1 , and p27KIP1,form the starting point for the design of mechanism-based CDK inhibitors. A number of these smallproteins have been dissected and inhibitory lead peptides amenable to peptidomimetic development havebeen identified. Conversion of these peptides into pharmaceutically useful molecules is greatly aided bythe recent elucidation of CKI/CDK crystal and solution structures. Additional interaction sites on CDKsbeing exploited for the purposes of inhibitor design include: phosphorylation/dephosphorylation sites,macromolecular substrate binding site, CKS regulatory subunit binding sites, cyclin-binding site, cellularlocalisation domain, and destruction box. Finally, progress has recently been made in the application ofantisense technology in order to target CDK activity.

INTRODUCTION CDK DIVERSITY

The topic of CDK inhibition and cancer therapyhas been reviewed several times, most recently byGarrett and Fattaey [1], Walker [2], and Meijer &Kim [3]. Significant advances have been made inthe field since these last overviews, including abetter understanding of the pharmacologicalconsequences of inhibiting cell-cycle regulatoryCDKs, the discovery of several new classes ofsmall-molecule CDK inhibitors, as well as thesolution of important crystal structures of CDKs incomplex with various regulatory proteins andchemical inhibitors. The latter topic was reviewedrecently [4]. The present review aims to cover insome depth the medicinal chemistry of bothinhibitors of the enzymatic function of CDKs, aswell as inhibitors of the various protein-proteininteractions leading to inactivation of CDKfunction.

To date nine different CDKs and at least 16different cyclins in mammalian cells have beendescribed. As can be seen from Table 1 , not allCDKs and cyclins function in cell cycle regulation.In this review only the major CDK/cyclin pairsinvolved in the regulation of the cell cycle will beconsidered. CDKs consist of a single polypeptidechain with ca. 200 amino acid residues and containa catalytic core that is completely inactive when theCDK is not phosphorylated and complexed with acyclin partner. CDKs are held in an inactive state byat least two major structural constraints: thesubstrate binding site is blocked and the ATPbinding site is distorted in such a way that the ATPphosphates are poorly positioned for efficientphospho-transfer [5]. The conformational changestaking place upon phosphorylation and cyclinbinding are now starting to become clear [6,7].

*Address correspondence to this author at the Cyclacel Limited, DundeeTechnopole, James Lindsay Place, Dundee DD1 5JJ, Scotland, UK;Telephone: +44 (0) 1382 43 21 42; Fax: +44 (0) 1382 43 21 55; E-mail:[email protected]

0929-8673/00 $19.00+.00 © 2000 Bentham Science Publishers Ltd.

1214 Current Medicinal Chemistry, 2000, Vol. 7, No. 12 Fischer and Lane

Table 1. Mammalian CDKs and Cyclins (Adapted from [8])

CDK Assoc ia ted cyc l ins Funct ion

CDK1 (CDC2), CDK2 A S phase entry and transition

CDK1 B1, B2 G2 exit, mitosis

CDK8 C Transcriptional regulation, G0-S phase transition

CDK4, CDK6 D1, D2, D3 G0-S phase transition

CDK2 E G1-S phase transition

? F G2-M phase transition

CDK5 G1, G2 DNA damage response

CDK7 H CDK activation, transcriptional regulation, DNA repair

? I ?

? K CDK activation, transcriptional regulation

CDK9 T1, T2 Transcriptional regulation

Until recently the majority of CDK inhibitorshave been designed and screened against CDK1and CDK2 complexes with cyclin A. This probablyreflects the comparatively ready availability of theseproteins and the fact that of all CDK/cyclin pairs theCDK2/cyclin A system is best understood at themolecular level. However, the CDK/cyclin pairsregulating cell cycle events preceding S phasewould seem equally interesting targets for inhibitordesign. Recent progress in purification andcharacterisation of e.g. CDK4/cyclin D1 andCDK7/cyclin H now make it possible to designmore readily inhibitors specifically against thesetargets [9,10].

CDKs to specific substrates or sub-cellularlocations; their function is primarily controlled bychanges in cyclin levels, which increasecharacteristically at specific cell cycle stages. Inaddition to cyclin binding, complete CDK activationrequires phosphorylation at conserved Thr residues(e.g. Thr160 in CDK2) located in the so-called T-loop. In vertebrate cell cycle control it appears that asingle CDK-activating kinase (CAK) canphosphorylate and thus activate all the majorCDK/cyclin substrates. CAK activity is due to thep40MO15 /cyclin H/MAT1 complex. CAK itself, likeits substrates, is a CDK, and p40MO15 has thereforebeen renamed CDK7 [10]. At first glance, CDK7,due to its role in activating all of the cell cycleCDK/cyclin complexes, would appear to offer aparticularly attractive target for inhibitor design.However, it has to be remembered that CAKactivity does not appear to change during the cellcycle and that the stimulation of CDKphosphorylation probably reflects cyclin bindingrather than CAK activity [5]. CAK activity is thusnot rate limiting during cell proliferation, althoughits regulation, which is as yet poorly understood,may be important under certain growth conditions[10].

THE CELL CYCLE

Activation of CDKs

The transitions between the phases of theeukaryotic cell cycle, summarised in Fig. (1), areregulated at checkpoints, including the restrictionpoint late in the G1 phase, at which a cell commitsitself to another round of DNA replication, as wellas the initiation and the completion of DNAreplication (S phase) and of cell division (mitosis).Entry into and progression through the cell cycleare triggered by CDKs. In order to assure theproper timing and co-ordination of the various cellcycle events, CDK activity is tightly regulated.Each CDK associates with and is partially activatedby a specific set of cyclin regulatory subunits.Cyclins are thought to contain regions that target the

Cell Cycle Progression

Once a cell has been stimulated to enter the cellcycle (sometimes called the G0-G1 transition), D-type cyclins are induced; unlike all other cyclins,their levels do not oscillate during the cell cycle but

Cyclin-Dependent Kinases Current Medicinal Chemistry, 2000, Vol. 7, No. 12 1215

Fig. (1) . The cell cycle (adapted from [11]).

are controlled by growth factors. They associatewith CDK4 and CDK6, whose primary substrate isthe retinoblastoma tumour suppressor protein (pRb)[12]. It plays a critical role in regulating G1progression through binding and regulating a largenumber of cellular proteins, including members ofthe E2F family of transcription factors. These inturn regulate the expression of many genes, whichencode proteins involved in cell cycle progressionand DNA synthesis. Binding of pRb to E2F factorsinhibits the latter’s transcriptional activationcapacity and converts them from transcriptionalactivators to repressors [8]. Phosphorylation ofpRb by D-type CDKs results in dissociation of pRbfrom E2F and expression of the E2F-regulatedgenes. In cells lacking pRb, D-type CDK activity isnot required for cell cycle progression [13].Through activation of E2F, cyclin E is the nextcyclin to be induced. It associates with CDK2 andthe resulting complex is important for the G1-Stransition by maintaining pRb in its hyper-

phosphorylated form. Unlike the D-type cyclins,cyclin E-type CDK activity is still required in cellslacking pRb, suggesting the existence of otherimportant substrates for CDK2/cyclin E [14]. Likemost CDK/cyclin complexes, CDK2/cyclin E alsophosphorylates histone H1, an activity probablyimportant for the chromatin rearrangementnecessary for replication. Cyclin A accumulates atthe G1-S transition and persists through S phase.Initially it associates with CDK2, later with CDK1.These CDKs are thus required for entry into Sphase, completion of S phase, and entry intomitosis. During the second gap phase G2, the cellprepares for the process of division. Mitosis then isregulated by CDK1 in association with cyclins A,B1, and B2. These complexes phosphorylateproteins such as lamins, histone H1, andcomponents of the mitotic spindle. Cyclins A and Bmust be degraded for a cell to exit mitosis andCDK1/cyclin B activity is involved in the regulationof this destruction process.


Table 2. In vitro Anti-Kinase Activity and Cytotoxicity of Some ATP Antagonists with ActivityAgainst CDKs

IC 50 (µM) l o gGI50 (M)

Compound

CD

K1/

cycl

in B

CD

K2/

cycl

in A

CD

K2/

cycl

in E

CD

K4/

cycl

in D

1

CD

K7/

cycl

in H

MA

P ki

nase

ER

K1

MA

P ki

nase

ER

K2

Prot

ein

kina

se C

α

cAM

P-de

pend

ent k

inas

e

cGM

P-d

epen

dent

kin

ase

EG

F-re

cept

or T

yr k

inas

e

C-S

RC

Tyr

kin

ase

MG

_MID

a

Staurosporinb 0.006 0.007 > 10 0.02 0.005 0.008 0.009 0.006 - 8

UCN-01c 0.031 0.03 0.032 0.91 0.007

9-Hydroxyellipticined ca. 1

Suramine 4 29 - 50 656 70

Butyrolactone-If 0.68 0.82 94 0.68 160 260 > 590 (- 3.9)

Isopentenyladenineg 45 50 200 90 50 43 50 50 1000

Olomoucineh 7 7 7 > 1000 30 7 > 1000 > 2000 > 2000 440 1000 - 4.1

Roscovitinei 0.65 0.7 0.7 > 100 34 14 > 100 > 1000 1000 250 - 4.8

Purvalanol Aj 0.004k

0.04l

0.5m

0.07 0.035 0.85 9 > 10 9 > 10 - 5.7

Purvalanol Bj 0.006k

0.05l

0.25m

0.006 0.009 > 10 3.3 > 100 3.8 > 100 Littleeffect

Paullonen 7 0.68 7.5 > 100 20 9 > 100 > 1000 > 1000 15

Kenpaullonen 0.4 - 4.4

Alsterpaulloneo 0.035 - 6.5

Indigop > 1000 70 > 1000 > 100 > 100 > 100 > 100 > 1000 900 28

Indirubinp 10 2.2 7.5 12 > 100 43 > 100 > 1000 > 1000 18

Indirubin-5-sulphonatep

0.055 0.035 0.15 0.3 38 > 100 > 100 > 1000 480 3.8 Littleeffect

Genisteinq > 100 2.6

Quercetinq > 100 16.6

L86 8276r 2.1

Flavopiridols 0.3 t 0.28 0.4 0.3 6 145 21 - 7.2

a) Mean graph mid-point in NCI human tumour cell line anti-cancer drug screen [20]. b) Ref. [21,22]. c) Ref. [21,23]. d) [ATP] = 50 µM; Ref. [24]. e) [ATP] = 100 µM; Ref.

[25,26]. f) [ATP] = 50 µM; Ref. [27,28]. g) [ATP] = 15 µM; Ref. [29]. h) [ATP] = 15 µM; Ref. [30,29]. i) [ATP] = 15 µM; Ref. [31]. j) [ATP] = 15 µM; Ref. [32,33]. k) [ATP]

= 15 µM. l) [ATP] = 150 µM. m) [ATP] = 1.5 mM. n) [ATP] = 15 µM; Ref. [34]. o) [ATP] = 15 µM; Ref. [35]. p) [ATP] = 15 µM; Ref. [36]. q) Ref. [37,38]. r) Ref. [38]. s)

Ref. [39,38]. t) [ATP] = 375 µM.


CDK Inhibition The different classes of small-moleculecompounds possessing CDK inhibitory activity byvirtue of ATP antagonism are discussed in detailbelow. A summary of the biological activity for themore important representatives, both in terms ofbiochemical kinase inhibition potency andselectivity, as well as effect on in vitro cellproliferation, is presented in Table 2 . Thecytotoxicity data is given, where available, in theformat used by the National Cancer Institute (NCI),using the human cancer cell line screen [18]. Itshould be noted that kinase assay IC50 values canonly be compared reliably if the ATP concentrationused is known, which is not the case for manyCDK inhibitor IC50s reported in the literature. Thispoint is illustrated by the entries for the purvalanols(CDK1/cyclin B). K i values, on the other hand, canbe compared directly, unfortunately, these arerarely reported. Only if the ATP concentration usedis well below the Km for the enzyme (ca. 40 µM forCDK2 [2]), are the K i app and IC50 valuesapproximately equal. Furthermore, ATPconcentration is also one of the factors affecting therelationship between enzyme IC50 values and invitro cytotoxicity effect. Intracellular ATPconcentrations are typically in the low millimolarrange, concentrations rarely used in kinase assays(typically 15 – 150 µM). An attempt has been madeto collect relevant comparable information in adatabase project but unfortunately this has not asyet progressed to a useful stage [19].

Biological CDK inhibition can be achieved inmany different ways, e.g. indirectly but veryimportantly through the changes in the levels of thevarious cyclins during the cell cycle. Directinhibition by phosphorylation at Thr14 and/or Tyr15

in human CDK1 and CDK2 is due to theWEE1/MYT1-related kinases [15]. This inhibitionmechanism is particularly important for CDK1regulation at mitosis. CDK1/cyclin B complexes arekept inactive until Thr14/Tyr15-dephosphorylationby the dual specificity phosphatase CDC25 at theend of G2 activates CDK1. Finally,dephosphorylation of Thr160/161 represents anadditional inactivation mechanism, the biologicalrelevance of which remains unclear.

Tumour suppressor functions are provided bythe so-called cyclin-dependent kinase inhibitors(CKIs). These are small proteins and fall into twoclasses: p21KIP1/CIP1 and p27KIP1 are specific forCDK2/CDK4 complexes, whereas the INK4proteins are specific for CDK4/6-cyclin complexes.The CKIs are capable of inhibiting directly theenzymatic activity of CDKs or through interactionwith the kinase substrate binding site. The mainregulatory mechanisms for the CKIs aretranscriptional or responses to mitogens and growthfactors. E.g. p21KIP1/CIP1 is transcriptionallyinduced by p53, which in turn is a transcriptionalregulator mediating cell-cycle arrest in response toDNA damage and in senescense. For p27KIP1,regulation by growth factors appears to beimportant. The CKIs will be discussed in moredetail below in connection with peptide andpeptidomimetic CDK inhibitors derived from them.

Staurosporins

Staurosporin, a microbial alkaloid, was firstisolated from Streptomyces sp. strain AM-2282(now classified as a new species with the nameStreptomyces staurosporeus) and was reported topossess anti-microbial properties against bothCDK INHIBITORS WITH ATP

ANTAGONIST ACTIVITY

N N

HNO

OH

NH

O

R1

R2

Staurosporin (R1 = R2 = H)UCN-01 (R1 = H, R2 = OH)UCN-02 (R1 = OH, R2 = H)

1

2

34

5 7

89

10

11

2'

3'4'5'

6'

Fig. (2) . Structure of staurosporins.

The ATP binding site is a very common motifpresent in a host of proteins, including the ca.2,000 kinases encoded in the mammalian genome.For this reason it has been thought that thedevelopment of kinase-specific ATP antagonistsdevoid of severe toxic effects would be verydifficult. However, recent advances in rational drugdesign and tumour pharmacology suggest thatnovel cancer therapeutics based on CDK inhibitionwill be available in the not-too-distant future. Atpresent the first two agents, flavopiridol andCGP60474 (see below), falling within the class ofCDK-specific ATP antagonists, are undergoingearly clinical evaluation [16,17].


yeasts and fungi [40]. The compound contains anoxadiazepine skeleton with a fused sugar moietyand possesses the absolute configuration shown inFig. (2).

instructive in terms of hints for the design of novelCDK inhibitors (see below).

Butyrolactone-IThe isolation from microbial fermentations of

further analogues, differing from staurosporin onlyin that they contain an additional hydroxyl functionat C7, was reported later [41]. These 7-hydroxystaurosporins are known as UCN-01 andUCN-02 and they are C7 diastereomers. Unlikestaurosporin, these compounds had no anti-microbial properties but they displayed cytotoxiceffects on certain mammalian cell lines, UCN-01being about twice as potent as UCN-02. Thesecytotoxic effects were hypothesised to emanatefrom cell growth disturbance through inhibition ofprotein kinases [41].

This compound was identified as a CDKinhibitor in the course of a screen ofmicroorganisms. It was found that Aspergillusstrain F-25799 produced strong inhibitory activityin an assay against murine cyclin B/CDK1. Theactive component was subsequently isolated fromthe mycelia and purified [27]. It turned out to beidentical to a compound previously namedbutyrolactone-I (3.3 ), first isolated as a newmetabolite from Aspergillus terreus var. africanusIFO8835 [44]. Its biosynthesis in this organism hasbeen elucidated [45] and involves conversion ofphenylalanine or tyrosine to methyl p-hydroxyphenylpyruvate 3.1 , two molecules ofwhich are then condensed between positions 2 and3’. The product 3.2 is lactonised, followed byisoprenylation in the last step, see Fig. (3).

It is now clear that both staurosporin and UCN-01 are indeed potent inhibitors of a wide variety ofkinases, including both Ser/Thr-, as well as Tyrkinases. The fact that at least part of thecompounds’ biological properties may be due toCDK inhibition was shown for both staurosporin[22] and UCN-01 [23]. However, one reportsuggested that inappropriate CDK activation eithercorrelated with or actually mediated cell growthinhibition with apoptosis in T lymphoblastsexposed to UCN-01 [42]. Interestingly, whereasstaurosporin inhibits CDK4/cyclin D1comparatively poorly [25], UCN-01 appears to be apotent CDK4 inhibitor [23].

As far as can be ascertained, no derivatives oranalogues of butyrolactone-I possessing CDK-inhibitory properties have been reported. Theelegant synthesis method used by nature suggests,however, that the substituted butyrolactonerepresents a good analogue template amenable tochemical synthesis.

Butyrolactone-I appears to be a kinase inhibitorwith high selectivity towards CDK1 and CDK2,with little activity on other kinases. In vitro growthof p53-negative pancreatic tumour cells wasinhibited by butyrolactone-I with IC50 values of 94µM (PANC-1 cell line) and 138 µM (AsPC-1 cellline), respectively [46]. Furthermore, consistent

The structural complexity of the staurosporinmolecules does not present a suitable template forthe synthesis of analogues with differingselectivity. Nevertheless, the staurosporin/CDK2complex crystal structure [43] has been very

O

OH

O O

HO

O

OH

HO

O O

O

OHO

O

OO

O O

O

OH

OHOH

OH

12 3

2

3'

Fig. (3) . Biosynthesis of butyrolactone-I.


with the proposed CDK-inhibitory mechanism forthe compounds, pRb phosphorylation and cyclin Aexpression were significantly inhibited in treatedcells. The fact that apoptotic cell death wasobserved suggests that exogenous inhibition ofCDK1 and/or CDK2 causes restraint of cellproliferation by induction of programmed celldeath. A study using various human prostatecarcinoma cell lines [47], on the other hand,revealed inhibition of cell proliferation andincomplete cell cycle arrest in G2/M, which allowedfor occasional skipping of mitosis and subsequentprogression through the cell cycle.

involved in the anti-cancer mechanism of 9-hydroxyellipticine [24].

The nucleoside antibiotic sangivamycin has beenknown for some time to inhibit protein kinases Aand C and to possess potent anti-proliferativeproperties [49]. In the course of a screen of culturemedia from soil microorganisms against CDK1, apotent inhibitor was isolated from Streptomyces sp.LPL931 culture medium. This compound wasshown to inhibit CDK1 with a K i value of 1.55µM, as well as CDK2 with similar potency and wasidentified as toyocamycin [50]. Toyocamycin isstructurally related to sangivamycin, differing onlyin a nitrile instead of carboxamide substituent on thepyrrolopyrimidine ring as shown in Fig. (4). Bothcompounds inhibit e.g. protein kinase C with IC50values of 12 µM (sangivamycin) and 90 µM(toyocamycin) [51]. These data suggest thattoyocamycin is a CDK-selective ATP antagonist.Furthermore, certain toyocamycin analogues haverecently been reported to inhibit CDK1 at lownanomolar concentrations, as well as to cause G1and G2/M cell cycle arrest, and to induce apoptosisin A549 human lung carcinoma cells [52].However, the impressive anti-proliferative effectsof sangivamycin, toyocamycin, as well asnumerous analogues [53,51] may only partlycorrelate with inhibition of kinases in general andCDKs in particular, since nucleoside antibioticspossessing furanose groups are phosphorylated inliving cells, leading to interaction with additionalcellular targets [54]. Nevertheless, toyocamyin-

Other non-Specific CDK Inhibitors

Suramin, a polysulphonylnaphthylurea, hasbeen in use not only as an anti-tumour agent, butalso for anti-helminthic and anti-protozoalindications [2]. Suramin appears to interact withnumerous intra- and extracellular targets, includingCDKs. Thus it was demonstrated that CDK1 isinhibited by suramin with an IC50 value of ca. 4 µM[26]. It remains unclear to what extent CDKinhibition contributes to the anti-proliferative effectsof suramin. 9-Hydroxyellipticine, an indolicalkaloid, represents a similar case. This molecule,and ellipticines in general, are known to exert theiranti-proliferative effects through interactions withDNA and/or topoisomerase II [48]. However, ithas been suggested that inhibition of pRbphosphorylation via CDK2 inhibition may be

O

HN

HN

NH

OHN

OHO3S

SO3HHO3S

NH

OHN

O

HO3S SO3H

SO3H

HO

NH

N

N

N N

RNH2

O

HO OH

HO

Suramin

9-HydroxyellipticineSangivamycin (R = CONH2)

Toyocamycin (R = CN)

Fig. (4) . Some non-specific CDK inhibitors.


derived pyrrolo[2,3-d]pyrimidines represent auseful starting point for the development of novelCDK inhibitors, particularly as the syntheticmethodology for these compounds is quite mature[55,56,53] and thus permits the preparation ofnumerous analogues in order to probe more fullythe SARs responsible for CDK inhibition and anti-proliferative effects.

(5), represent a large class of selective EGF-receptor tyrosine kinase inhibitors [59], some ofwhich have seen clinical evaluation as cancertherapeutics. It has been speculated that the effect offlavonoids on cancer cells may be partly due tomechanisms other than tyrosine kinase inhibition,e.g. genistein and quercetin have been reported toinhibit topoisomerase II [37]. Whereas both thesecompounds exhibit only very weak CDK inhibitoryactivity (IC50

> 100 µM), myrecetin displays anIC50

of 10 µM ([ATP] = 40 µM) against CDK2 [2].However, CDK inhibition is probably moreimportant for the biological effects of rohitukineanalogues L86-8276 and flavopiridol, which arealso potent tyrosine kinase inhibitors (compareTable 2).

Flavonoids

Flavonoids play an important role in plantphysiology and their interactions with humanmetabolism are numerous, since they form part ofour diet. They are believed to possess propertiescapable of potential protection againstcarcinogenesis [57] and have long been known tointeract with various kinases at the ATP-bindingsite [58]. Flavone acetic acid, genistein, andquercetin, whose structures are presented in Fig.

The isolation from the stem bark of Dysoxylumbinectariferum and the total synthesis of acompound [60] previously described as rohitukine[61,62], enabled the synthesis of flavopiridol and

OH

HO O

O

N

OH

OH

HO O

O

N

OH

Cl

OH

HO O

O

N

OH

OH

HO O

O

OH

OH

OH

OH

OH

HO O

O

OH

O

OH

HO O

OOH

OH

HO O

O

OH

OH

OH

Rohitukine L86-8276

Flavopiridol

Myrecetin

Flavone acetic acid Genistein Quercetin

Fig. (5) . Structures of some flavonoids with kinase inhibitory properties.


its analogues [38], summarised in Fig. (6). Themain problem in the synthesis of flavones 6.9concerns the installation of the cis-alcohol functionin the piperidine ring. Trimethoxyphenyltetrahydro-

substitutions of OH-modified derivatives of 6.2with various oxygen nucleophiles do not give 6.7 ;instead, ring contraction to 6.6 is observed.Similarly, attempted inversion via Mitsunobureaction failed and resulted only in elimination andreversion to 6.1 . For this reason, a rather clumsyroute had to be followed, involving conversion toketone 6.3 , followed by borohydride reduction to6.4 . Even here stereoselectivity is poor and thegeometric isomers had to be separated throughfractional crystallisation or chromatography. Thelevorotatory enantiomer with the absolutestereochemistry shown for 6.7 is accessiblethrough crystallisation of diastereomeric tartratesalts. Partial demethylation and simultaneousacetylation then provide 6.8 , which upondeprotonation and reaction with benzoate esters

pyridine 6.1 is readily obtained by Friedel-Craftsacylation of trimethoxybenzene with 1-methyl-4-piperidinone. In principle, epoxidation of the alkeneshould provide 6.5 , which could be reduceddirectly to the (racemic) cis-arylpiperidinol 6.7 .Unfortunately, epoxide 6.5 does not appear to beaccessible from 6.1 using peracids. The fact thatthe reaction also fails with various salt forms of6.1 indicates that not only interference through N-oxidation is involved. Hydroboration/oxidation ofalkene 6.1 is possible but affords the trans-arylpiperidinol 6.2 , as expected from themechanism of this reaction. Again, SN2

OH

HO O

O

N

OH

R

O

O O

N

O

O

O

N

O

O

O

N

OO

O

O

N

O

O

O

N

O

OOH

O

O

N

O

OH

O

O

N

O

OH

O

O

N

OH

OH

O

O

O O

NOH

2: predominantlytrans

4: ca. 7:3 cis/trans

1.) S

epar

atio

n of

isom

ers

2.) R

esol

utio

n of

ena

ntio

mer

s

Epoxidation

HydroborationOxidation Swern

oxidation

1 3

5 6

789

NaBH4

Fig. (6) . Synthesis of rohitukine analogues.


readily form the required chromone ring. The finalstage of the synthesis is achieved by demethylationof the remaining two O-methyl groups.

appearance of apoptotic cells in xenografted headand neck squamous cell carcinomas have beenreported [65]. The promising preclinical effectsseen with flavopiridol [38] have led to phase I/IIclinical evaluation of the drug. Initial results areencouraging in that the drug was comparativelywell tolerated, concentrations needed for CDKinhibition in preclinical models could be achievedsafely, and some antineoplastic effects wereobserved [16].

Around 100 rohitukine derivatives were testedfor EGF-receptor tyrosine kinase inhibitory activity[38]. It was found that the enzyme- and the in vitroanti-tumour SARs correlated well and thatflavopiridol was the most potent compound.Subsequent studies revealed that flavopiridolinhibited CDKs approximately 100-fold morepotently than tyrosine kinases. This observation,together with mechanistic studies, lead to theconclusion that CDK inhibition was more relevantto flavopiridol’s cell growth inhibition propertiesthan EGF-receptor tyrosine kinase inhibition. Infact the SARs for the inhibition of these twoenzymes by rohitukine analogues overlap at least inpart. Most relevant to CDK2 inhibition are thefollowing structural correlations: removal of thechlorine substituent on the phenyl group offlavopiridol reduced activity ca. 10-fold;conversely, introduction of chloro substituents atthe 4- or 5-position of the same ring also reducedactivity. Further loss of activity was seen uponreplacing the chlorophenyl group with smallaliphatic groups. Replacement of the hydroxygroups at the 5- and 7-positions of the flavonemoiety also dramatically reduced inhibitorypotency. As far as stereochemistry is concerned,the cis(-) isomer possessed higher potency thaneither the cis(+) or the trans(+/-) isomers. CertainSARs suggest that selectivity towards individualCDKs can be achieved. Thus it was found thatreplacement of the chloro substituent by fluoro,bromo, methyl, or hydroxyl groups reducedactivity relative to flavopiridol; however, the bromoand fluoro derivatives showed about 2-fold greaterselectivity towards CDK2. Replacement of thechlorophenyl group with 2- or 3-pyridyl groupsresulted in reduced activity. However, thecorresponding 4-pyridyl derivative was equipotentwith flavopiridol against CDK2 but significantlyless selective for CDK4 and CDK7. Substitution ofa pyridyl group for the N-methylpiperidine ring, onthe other hand, increased selectivity for CDK4.

Purines

The story of substituted purines as CDKinhibitors begins with 6-dimethylaminopurine,shown in Fig. (7), first described as an inhibitor ofcell division in embryos [66] and later found to be aCDK1 inhibitor [67]. Subsequently, using a newscreening test based on starfish CDK1/cyclin B forantimitotic compounds, isopentenyladenine, a non-selective kinase inhibitor with activity againstCDK1, was identified [68]. In a systematic studyusing cytokinin-derived purine analogues,olomoucine was then discovered as the first trulyCDK-selective purine [29]. Further SAR studiesaround 2,6,9-trisubstituted purines lead to thediscovery of a closely related compound,roscovitine, with approximately 10-fold increasedpotency against CDKs and similar selectivity[31,69]. The purine analogues with the highestCDK inhibitory potencies reported to date are thepurvalanols [70]. They were obtained aftercombinatorial synthesis and screening of severalhundred 2,6,9-trisubstituted purines. Purvalanol Bis more potent in biochemical CDK assays thanpurvalanol A; the former compound possesses littleactivity on cells, however.

Of all the classes of compounds with CDKinhibitory activity, the purines are best understoodin terms of their SARs. The order of importance ofthe three purine ring substituents with respect tokinase inhibition is 2- > 6- > 9-. Furthermore,approximate additivity of the three substituents isobserved. Presumably this is due to the fact that therigid purine scaffold does not permit the 2-, 6-, and9-substituents to bind in overlapping regions of theactive site. The SARs are summarised in Fig. (8)and are discussed in detail below.

In accordance with its proposed biologicalmechanism as a general CDK inhibitor, flavopiridolwas shown to inhibit both CDK2 and CDK4 invivo and to inhibit cell cycle progression in eitherthe G1 or G2 cell cycle phase [63]. Furthermore,significant in vivo anti-tumour effects wereobserved with xenografted lung, colon, mammary,and ovarian tumours, as well as glioblastomasusing daily oral dosing [64]. Similarly, sustainedtumour reduction (60 – 70 %), as well as

The majority of purine analogues reportedpossess a 2-substituent containing anhydroxyethylamino or aminoethylamino [71] motif.However, in some analogues with conformationallyconstrained 2-substituents, the number of carbonatoms separating the two functional groups may be


N

N N

N

HN

NH

HO

N

N N

N

HN

NH

HO

N

N NH

N

N

H2NN

N NH

N

HN

H2N

N

N N

N

HN

NH

HO

Cl

R

IsopentenyladenineDimethylaminopurine Olomoucine

Roscovitine Purvalanol A (R = H)Purvalanol B (R = COOH)

Fig. (7) . Structures of some purine derivatives with CDK inhibitory properties.

as high as four. In non-cyclic hydroxyalkylamino2-substituents, a (preferably branched) alkyl groupvicinal to the amino function has yielded some ofthe most potent compounds to date. Certain cyclichydroxyalkyl and aminoalkyl groups, either vicinalto N2 or including it, also yield potent compounds.Clearly the amino group bonded to the purine ringcan be 2o or 3o, also shown by the fact that(dihydroxyethyl)amino 2-substituents providepotent inhibitors [72]. Furthermore, this aminogroup can be replaced by an alkynyl group, butapparently not by alkenyl or alkyl groups [73]. Thecrystal structure complexes of CDK2 witholomoucine [74], roscovitine [75], and purvalanolB [32] all show that the hydroxy group in the 2-substituent contributes an important H-bond to theactive site. Nevertheless, potent compoundscontaining a thiomorpholine function, the sulphuratom being unable to act as an H-bond donor, havebeen reported [76].

substitution patterns on the aromatic ring varybetween the anilino- and benzylamino series. In theformer, substituents of any kind are poorlytolerated in the ortho position. Activity gains can bemade with certain electron-withdrawingsubstituents in the meta position, particularly chloroand bromo groups. The same is true, although to amuch lesser extent, of the para position.Interestingly, amino- and carboxyl-substitution inaddition to the meta-chloro group, has yielded themost potent analogues. For purvalanol B, wherethis additional group is a carboxylic acid, a salt-bridge from this group to the Lys89 side chain in theATP binding pocket, as well as further additionalinteractions, can be observed [70]. As pointed outabove, addition of such protic substituents, whileincreasing CDK inhibitory activity, appears to bedetrimental to anti-tumour activity, perhaps due topoor cellular uptake of such compounds. In the 6-benzylamino series, addition of further substituentsto the aromatic ring does not generally have thesame dramatic effects as in the 6-anilino series.However, certain electron-releasing groups in thepara- (particularly methoxy), and electron-withdrawing groups in the meta position (notablyiodo), correlate with increased potency.

The 6-position of purine analogues with potencyand selectivity towards CDKs is invariablysubstituted by benzyalmino or anilino groups.Unlike in the 2-substituent, the amino group hereneeds to be 2o rather than 3o. CDK2/purine crystalstructures show an important hydrogen bondbetween the purine N6H group and the Leu83

backbone carbonyl of the ATP active site. The bestThe 9-position of the purines is very restrictive

towards substitution in terms of CDK inhibition.


N

N N

N

HNR2

R3R1

NH

HO

NH

HONH

HO

N

HO

OHHO N

S

NHO

OH

N

H2N

NH

H2N

XCl

OH

O

OI

Cl

NH2

O CF3

X

NH

HO

NH

NH2

29

61

3

7

8

Fig. (8). Structure-activity relationships of 2,6,9-trisubstituted purines. Some of the substituents optimal with respect toCDK inhibitions for each of the positions are shown.

Only small aliphatic and alicyclic groups, optionallycontaining hydroxyl groups [77], are tolerated. The9-isopropyl group appears to be optimal. Additionof methyl substituents at N1, N3, or N7 of purinesdramatically reduces CDK2 inhibitory activity [69].Little appears to be known about the influence ofsubstituents at C8, although this position wouldappear to be synthetically accessible, particularlyusing the synthetic route of Legraverend et al. [78].

solution are summarised in Fig (9). The mostconvenient starting material is 2,6-dichloropurine(9.2 , X = Cl). Due to the pronounced difference inreactivity towards substitution of the chlorosubstituents in the 2- and 6-positions of the purinering, selective amination under comparatively mildconditions readily affords aminochloropurines 9.3 .The reaction succeeds even with deactivated arylamine educts R1-NH2 under suitable conditions.Deprotonation of the N9H group in 9.3 with eitherNaH or K2CO3 in polar aprotic solvents, followedby alkylation with alkyl halides R2-X cleanlyaffords the disubstituted intermediates 9.8 . Finally,if the desired trisubstituted purine 9.9 has analkylamino R3 substituent, as is usually the case,the 2-chloro group is substituted at elevatedtemperature with excess alkyl amine. If the alkylamine is a liquid, the sealed tube method is usuallyfavoured, else heating a solution in N-methylpyrrolidinone or Me2SO is adopted.However, poor reactivity of the 2-chloro group in9.8 precludes this final amination reaction for

As far as selectivity within the CDKs isconcerned, only minor differences between thevarious 2,6,9-trisubstituted purines can bediscerned. There is one notable exception,however: purvalanol A (as opposed to purvalanolB) is the only purine with submicromolar IC50against CDK4. Apart from flavopiridol, this wouldin fact seem to be the only CDK-selectivecompound with appreciable activity against CDK4.

The routes commonly used [79,76,69,80,77]for the synthesis of 2,6,9-trisubstituted purines in


N

N NH

N

Cl

H2N

N

N NH

N

Cl

N

N NH

N

HN

X

R1

N

N N

N

HN

X

R1

R2

N

N N

N

HN

R3

R1

R2

X

N

N N

N

Cl

X

R2

N

N N

N

Cl

H2NR2

N

N

Cl

H2N

N

N

Cl

H2N

NH2

Cl

NH2

NH

R2

N

N N

N

Cl

H2NR2

N

N N

N

Cl

R3

R2

1 2

3

4

5 6 7

89

10 11

Fig. (9) . Solution synthesis routes to 2,6,9-trisubstituted purines.

many sterically hindered and poorly nucleophilicalkyl- and aryl amines. The 2-fluoro group (9.8 , X= F) renders the purine 2-position much moreelectrophilic and has been found to be moreversatile for the final amination step. It isintroduced (9.2 , X = F) by diazotisation/fluorino-

cyclised to 9.7 under acid-catalysed conditionsusing excess triethylorthoformate. Furthertransformation to 9.4 (X = I) involvesdiazotisation/substitution of the amino group in9.7 , the iodide source being provided by a mixtureof CH2I2, I2 and CuI. Similar to the fluoropurines9.4 (X = F), the iodopurines 9.4 (X = I) aresuperior to chloropurines 9.4 (X = Cl) for the finalamination reactions. Additionally, iodopurines 9.4(X = I) are suitable if the desired products 9.9contain alkyl rather than alkylamino R3

substituents, since they can be used directly in Pd-catalysed cross-coupling reactions with terminalalkynes [73]. In a further synthetic route [82],regioselective N9-alkylation of 9.1 is performed asthe first step, followed by Mitsunobu alylation ofthe product 9.10 with alcohols R3-OH. In order torender the 2-amino group in 9.10 sufficientlyacidic to undergo Mitsunobu alkylation, it was firstacylated with trifluoroacetic anhydride. Thetrifluoroacetyl group was then removed by basic

lysis of 2-amino-6-chloropurine 9.1 [81]. In thosecases where the R1 group contains functionssensitive to alkylation, the order of 6-amination and9-alkylation can be reversed. Thus 9.2 can be N9-alkylated directly using the same conditions asabove, although formation of 9.4 is accompaniedby N7-alkylation. Regioselectivity has beenreported [82] to be good if alkylation is carried outusing 1o or 2o alcohols R2-OH under Mitsunobuconditions [83]. A different route to the 9-alkylpurines 9.4 has recently been described [78].In this case 2,5-diamino-4,6-dichloropyrimidine9.5 [84] is the starting material, which isselectively C4-monosubstituted using oneequivalent of R2-NH2. The product 9.6 is then


NH

O

OO

O

O

H

O

OO

O

O

H

HO

NH2

NH

O

OO

O

HN

R1

N

N N

N

Cl

F

O

N

N NH

N

Cl

F

Linker

N

N N

N

N

F

O

R1Linker

N

N N

N

N

F

R2

R1Linker

N

N N

N

N

NH

R2

R1

N

N N

N

HN

NH

R2

R1

R3

R3

Si

Si

Linker

+

1

2

3

4

5 6

789

10

CF3COOH/H2O

Fig. (10) . A solid-phase synthesis route for 2,6,9-trisubstituted purines.

hydrolysis prior to purification of the alkylatedproducts 9.11 . Since the R3 group in most targetcompounds 9.9 itself contains an alcohol function,this requires protection, e.g. as a silyl ether. Thefinal step, i.e. 6-amination of 9.11 , was carriedout as usual.

followed by alkylation as usual to afford the 9-alkylated products 10.8 . These were aminated atthe 2-position (refer above) to give the immobilisedtarget compounds 10.9 . The remaining step wasacidolytic cleavage of the purine-linker bond inorder to obtain the free 2,6,9-trisubstituted purines10.10 . The limitations of the approach are that 2o

amines at the 6-position cannot be introduced andthat the reactivity of the solid-phase boundfluoropurines 10.8 towards bulky amines R3-NH2was much lower compared to the correspondingtransformations in solution.

Various solid-phase synthesis approaches for thepreparation of trisubstituted purines have beendeveloped [81,70,82]. The most versatile of theseis summarised in Fig. (10). Amino-functionalisedsolid-phase synthesis resins 10.1 were acylatedwith 5-(4-formyl-3,5-dimethoxyphenyloxy)valericacid 10.2 . Resin 10.3 was then transformed withamines R1-NH2 by reductive amination. In order tobe able to perform the amination reaction at the 6-position of the purine 10.5 with weaklynucleophilic amines 10.4 (depending on the natureof R1), the overall electrophilicity of the purinesystem had to be increased. The best solution tothis problem was derivatisation of the N9 positionwith the trimethylsilylethoxymethyl activatinggroup. Condensation of solid-phase bound 10.4with the derivative 10.6 did indeed occur undervery mild conditions. Product 10.7 was thendeprotected with tetrabutylammonium fluoride,

Paullones

Fused [1]benzazepine-2,5-diones were firstinvestigated as anti-tumour agents on the basis thatthey may represent anti-metabolites of riboflavinand vitamin B12, both of which play a role in certaintumours. A number of such compounds, e.g. thosewith structures 11.1 – 11.4 shown in Fig. (11),were indeed found to possess marked anti-tumouractivity. The lack of tumour cell line selectivityobserved with e.g. 11.4 suggested a non-specificmode of cytotoxicity, however [85].


HN

O

O

HN

O

ONH

N

HN

O

O

HN

O

O

S

1 2 43Fig. (11) . Benzazepinediones with anti-tumour activity.

Compounds with the basic scaffold 7,12-dihydro-indolo[3,2-d][1][benzazepin-6(5H)-onewere first reported in 1992 [86] and they seemed topossess only modest anti-tumour activity. Someyears later, however, analysis of the NCI humantumour cell line anti-cancer drug screen data usingthe COMPARE algorithm [87] to detect similaritiesin the pattern of compound action to flavopiridolrevealed that 7,12-dihydro-indolo[3,2-d][1][benzazepin-6(5H)-ones, now termedpaullones (in honour of the late inventor of theCOMPARE algorithm, Ken Paull), were in factpotent CDK inhibitors [34]. This finding spurredfurther investigation of this class of compounds,leading to the discovery of the as yet most potentanalogue in the series, known as alsterpaullone[35], see Fig. (12). This compound displays invitro anti-tumour activity similar to flavopiridol.

>> Br). Similar substituents in the 2-, 10-, or 11-position are detrimental to activity. Addition ofsubstituents in the 2- and/or 3-position has minoreffects on potency. The 4- and 7-positions, on theother hand, are more sensitive to the introduction ofsubstituents; e.g. the 4-methoxy-substitutedkenpaullone had only little activity. Derivatisationof either the indole or lactam N’s was poorlytolerated, as was replacement of the lactone with athioimidate function. Overall, CDK1 inhibitorypotency was not exactly paralleled by in vitro anti-tumour activity. Most strikingly, 9-cyano-paullone,the most potent CDK1 inhibitor in the series,showed only low in vitro anti-proliferative activity.Amongst other effects, a shift of the bromo-substituent in kenpaullone from the 9- to the 10-position resulted in reduced CDK1 inhibition butimproved anti-tumour activity. Most paullones donot appear to have been screened extensivelyagainst CDKs other then CDK1 or against non-CDK kinases. The discrepancies between CDK1inhibition and anti-tumour activity are likelyconnected with varying kinase-selectivity betweenthe different analogues. Compounds which are

So far, ca. 40 paullones have been studied and aSAR is starting to emerge: Electron-releasinggroups in the 9-position decrease CDK inhibitoryactivity, whereas electron-withdrawing substituentsin the same position enhance activity (CN > NO2

HN

O

HN1

2

3

4

7

8

9

1011

HN

O

HN

NO2

HN

O

HN

HN

O

HN

Br

Paullone Kenpaullone Alsterpaullone

Fig. (12) . Structures and numbering system of paullones.


R1

R1

R1

OOEt

O

ClR1

HN

H2N

HN

O

HN

NH

OEt

O

R1

R2

R2

NH

OEt

O

OOEt

O

HN

O

O

HN

O

HOO

OEt

5

1 2 3

4Fig. (13) . Synthesis of paullones.

structurally related to the paullones, viz. quinolino-and pyrido[3,2-d][1]benzazepines, also possessanti-tumour properties [88], although these do notappear to be due to CDK-inhibition.

13.5 . The results of acylation and alkylationreactions of 13.5 (R2 = Br) under variousconditions suggest that both the lactam and indoleN’s, as well as C7, are amenable to furthermodification. Furthermore, the lactam function canbe modified successively to the thiolactam,alkylthioimidated by S-alkylation, andhydroxamidine by nucleophilic exchange.

The key intermediates in the synthesis ofpaullones, 3,4-dihydro-1H-1-benzazepin-2,5-diones 13.4 , can be prepared by a variety ofmethods [85], see Fig. (13). The most useful interms of availability of starting materials are thermalring-closure of o-succinoylated anilines anddealkoxycarbonylation ofethoxycarbonylbenzazepindiones 13.3 underneutral conditions with wet Me2SO [89]. The latterare readily elaborated starting from anthranilic acidesters 13.1 , followed by N-succinoylation todiesters 13.2 and KH-catalysed Dieckmanncyclisation. Paullones 13.5 are then obtained byFischer indole synthesis: the acetophenone functionin 13.4 is first transformed to the hydrazone bycodensation with suitable phenylhydrazines inAcOH at 70 oC, followed by addition of conc.H2SO4 and heating to effect ring closure to indoles

Indigoids

Chronic myelocytic leukaemia has traditionallybeen treated in Chinese medicine with a mixture ofherbs, one of the active ingredients of which wastraced to the dark blue powder prepared from theleaves of a variety of plants [36]. This powdercontains a high proportion of the blue dye indigo;however, the anti-leukaemic activity was attributedto the red-coloured isomer of indigo, known asindirubin [90]. The pharmacology of indirubin andsome of its analogues, refer Fig. (14), including 5-halogenoindirubins, N-ethylindirubin, and

NH

HN

O

ONH

NH

O

R1

R2

Isoindigo

1

2

39

4

5 6

7

82'

3'9'4'

5'

6'

7'8'

1'

Indirubin (R1 = O, R2 = H)

Fig. (14) . Structures of indigoids.


NH

NH

O

O

R1

R2

NH

OH

R1

NH

O

R2

O+

1 2 3

Fig. (15) . Synthesis of indirubins.

indirubin-3’-monoxide is fairly well understood butthe fact that indirubins are CDK inhibitors wasdiscovered only recently (Ref. [36] and referencescited therein). Indirubin and its analoguesselectively inhibit CDKs and block cell proliferationin the late G1 and G2/M phases of the cell cycle.

New Compounds

A series of purine-based CDK inhibitorsexemplified by a compound described as O6-cyclohexylmethylguanine, see Fig. (16).apparently with inhibitory activity comparable toolomoucine against CDK1, but > 50-fold moreactive against CDK4, were reported [95]. Crystalstructures of selected compounds from this seriesbound to CDK2 revealed a different binding modeto e.g. olomoucine, with the O6-alkyl substituentsoverlapping the ribose-binding domain of thesubstrate ATP. Interestingly, removal of the 2-amino group from NU2058 reduced potencyagainst CDK1 but not against CDK2, whilstmethylation at N9 reduced potency against both.Corresponding 4-substituted 5-nitroso-pyrimidineswere also shown to be potent CDK inhibitors [96].Thus NU6027 showed low micromolar activityagainst CDK1 and CDK2. The relationship of thepyrimidine analogues with the purine analogues isnot entirely clear, although alkylation of the 6-amino and replacement of the 5-nitroso groups inthe latter correlated with loss of potency.

Although a comparatively large number ofindirubin analogues have been reported [91-94],kinase inhibition SARs for only indigo and fourindirubin derivatives have been published [36]. Asis the case with most selective CDK inhibitors,activity against CDK4 was about one order ofmagnitude less than against the other cell-cycleCDKs for most indirubin analogues, with theexception of indirubin itself, which isapproximately equipotent against all CDKsexamined but has comparatively poor overallpotency. Both 5-chloroindirubin and indirubin-3’-monoxime show high nanomolar IC50s againstCDK1 and CDK2, whereas indirubin-5-sulphonicacid is even more potent. The latter compound,despite its high inhibitory activity against CDKs,had limited effects on cell proliferation, presumablydue its limited permeability.

A different pyrimidine-based compound, termedCGP60474 and representative of a group ofcompounds described as a phenylamino pyrimidinederivatives [97], was reported to be a potent andselective CDK inhibitor with specificity for CDK1and CDK2. The presumed structure of CGP60474is shown in Fig. (16) [4]. Potent (IC50s 10 – 100nM) in vitro anti-proliferative effects against avariety of human tumour cell lines, as well aspotent anti-tumour activity agaist various human

Indirubins 15.3 are synthetically accessiblethrough condensation between indoxyl derivatives15.1 and isatins 15.2 [93] as shown in Fig. (15).The generality of this reaction, together with thefact that both indoxyls and isatins can besynthesised by cyclisation of precursors derivedfrom anthranilic acid, renders the indirubin systeminteresting for further analogue synthesis andevaluation.

N

N NH

N

O

H2N

N

N

O

H2N NH2

NO

N

NNH

Cl

N

NH

HONU2058 NU6027 CGP60474

Fig. (16) . New purine- and pyrimidine-based CDK inhibitors.


O

NS

HN

H2N

O

NS

HN

H2N

X

X

SO

O

NH2

HN

O

R

R'

HN

O

NHN

R

S

H2N

OO

N NH

O O

N NH

O S

Cl

1a1b 2a

2b 3a 3bFig. (17) . Structures of some recently reported CDK inhibitors.

tumours transplanted onto nude mice, werereported [98].

analogues with the general structure 17.2b werereported as CDK inhibitors [101]. Indenopyrazole17.3a , found by high-throughput screeningagainst CDK4/cyclin D1, formed the starting pointfor a SAR program which afforded analogue17.3b , reputedly a potent and selective CDK4inhibitor, capable of inducing cell cycle arrest andapoptosis in tumour cell lines [102].

Interesting new compounds, variously found byselective screening against CDKs of compoundcollections and through structure-based design,have recently been reported. While little is known atpresent regarding their in vitro and in vivopotencies, selectivities, pharmacokinetics, etc.,some examples are shown in Fig. (17). A class of sulphonamides possessing anti-

tumour properties due to interference in theprogression through the M-phase of the cell cyclewas reported and compound E7010, the optimisedcompound from this class, was shown to be apotent tubulin polymerisation inhibitor. Furthersynthesis of a large number of sulphonamidesrevealed that certain N-(7-indolyl)benzenesulpho-namides, notably compound E7070, also had anti-proliferative properties but that these were not dueprimarily to antimitotic effects but due to a block inprogression through G1-phase [103]. Although themain cellular target for compound E7070 has not asyet been identified, it is possible that CDKinhibition may be involved. E7070 is currentlyundergoing phase I clinical trials in Europe [104].

The substituted diaminoacylthiazole 17.1a wasidentified as a CDK inhibitor through libraryscreening. Molecular modelling-assisted SAR-studies later identified more potent compounds17.1b (X = Cl or F), which apparently possesslow nM K is against CDK1/cyclin B, CDK2/cyclinA, and CDK4/cyclin D, as well as showingsignificant tumour growth delay in mouse tumourxenograft models [99]. Interestingly, the class ofcompounds exemplified by 17.1a and 17.1b wasreported to encompass examples with > 100-foldselectivity for CDK4, as well as examplesequipotent against CDK1, CDK2, and CDK4[100]. Indoline 17.2b was identified throughmolecular modelling. Starting from this lead, potent

HO NH

N

HN S

O

O

O HN S

O

O

SO2NH2HN

E7010 E7070

Fig. (18) . Novel sulphonamide anti-tumour drugs with M- or G1-phase cell cycle effects.


Insights from Crystal Structures BetweenCDK2 and ATP Antagonists

A range of structurally diverse molecules havebeen shown to inhibit CDKs by virtue of beingATP antagonists. Several complex structures ofsuch inhibitors with CDK2 have recently becomeavailable and some of these are showndiagramatically in Fig. (19). The CDK2 portions,with the exception of the positioning of certainactive-site residue side chains, in all the knowncomplexes with ATP antagonists are practicallyidentical to that found in the CDK2 apoenzyme[105], the ATP complex [105,106], and are verysimilar to that in the CDK2/cyclin A complex [107].The inhibitor portions invariably bind into the deepATP-binding cleft between the N- and C-terminallobes of CDK2.

Articulation about the hinge region connectingthese two lobes permits a range of possible relativedomain orientations in protein kinases. Togetherwith the observed flexibility of the residue sidechains within the active site, this goes some way inrationalising why selective inhibitors exist.Furthermore, exploitation of sequence differencesbetween various kinases as shown in Fig. (20) inthe residues surrounding the conserved ATP-binding points, as well as novel interactions withinthe binding cleft, form a useful basis for the designof novel selective inhibitors [108].

Purines

The purine ring of isopentenyladenine is rotatedca. 180o with respect to ATP around an axisthrough the N3 and N7 atoms, allowing theisopentenyl group to contact the binding pocket areaoccupied by the ribose in ATP. In the olomoucinecomplex, the purine ring is oriented differently:here the hydroxyethyl group of N2 binds into theATP ribose pocket, whereas the N6 benzyl groupbinds between β-strand 1 and the β5-α2 connectingloop, which bridges the two domains of CDK2[74]. The purine ring of roscovitine, as well as itssubstituents, are located in the CDK2 structure in asimilar orientation to those of olomoucine [75]. Theoverall geometry of purvalanol B bound to CDK2resembles that of both olomoucine and roscovitine[32]. A conserved pair of H-bond interactionsbetween the purine N7 and the backbone NH ofLeu83, and between the N6 amino group and thebackbone carbonyl of Leu83 can be observed in allthree (olomoucine, roscovitine, purvalanol B)CDK2/purine complexes. Another conserved H-bond in these complexes is that between the acidicC8 position of the purine ring and the carbonyl

Fig. (19) . Alignment of CDK inhibitors in the ATP-binding pocket of CDK2. ATP (a) and the inhibitorsstaurosporin (b), flavonoid L-868276 (c), purvalanol B (d),and kenpaullone (e) are shown in their receptor-boundorientation. The structures were generated bysuperimposition of the ATP-binding residues in CDK2shown in (a). For ATP [106], staurosporin [43], andpurvalanol B (only the main conformation shown) [32], thepublished crystal structure co-ordinates were used, flavonoidL868276 [109] and kenpaullone [110] were modelled.


L1:b1--:L2------:b2----:L3-:b3-----:L4------: a1----------:L5-------:b 10 20 30 40 50 60CDK2 MENFQKVEKIGEGTYGVVYKARNKLTG-EVVALKKIR-LDTETEG---VPSTAIREISLLKELN---HPNIV CDK1 MEDYTKIEKIGEGTYGVVYKGRHKTTG-QVVAMKKIR-LESEEEG---VPSTAIREISLLKELR---HPNIV CDK4 MATSRYEPVAEIGVGAYGTVYKARDPHSG-HFVALKSVR-VPNGGGGGGGLPISTVREVALLRRLEAFEHPNVV CDK6 MEKDGLCRADQQYECVAEIGEGAYGKVFKARDLKNGGRFVALKRVR-VQTGEEG---MPLSTIREVAVLRHLETFEHPNVV CDK7 MALDVKSRAKRYEKLDFLGEGQFATVYKARDKNTN-QIVAIKKIK-LGHRSEAKDGINRTALREIKLLQELS---HPNII ERK2 MVRGQVFDVGPRYTNLSYIGEGAYGMVCSAYDNVNK-VRVAIKKISPFEHQT-----YCQRTLREIKILLRFR---HENII KPCa PSNNLDRVKLTDFNFLMVLGKGSFGKVMLADRKGTE-ELYAIKILKKDVVIQDD---DVECTMVEKRVLALLD--KPPFLT KAPa ESPAQNTAHLDQFERIKTLGTGSFGRVMLVKHKETG-NHYAMKILDKQKVVKLK---QIEHTLNEKRILQAVN---FPFLV KPBb EAAFFANLKLSDFNIIDTLGVGGFGRVELVQLKSEESKTFAMKILKKRHIVDTR---QQEHIRSEKQIMQGAH---SDFIV EGFR NQALLRILKETEFKKIKVLGSGAFGTVYKGLWIPEGEKVKIPVAIKELREATSP--KANKEILDEAYVMASVD---NPHVC SRC LAKDAWEIPRESLRLEVKLGQGCFGEVWMGTWNGTT--RVAIKTLK-PGTMS------PEAFLQEAQVMKKLR---HEKLV :* * :. * : * :: : L9 4------: L6---:b5:L7-:a2------:L8-------:a3---------------: b6--:L10-:b7--:L11 70 80 90 100 110 120 130CDK2 KLLDVIHT-----ENKLYLVFEFLHQDLKKFMDASALTG--IPLPLIKSYLFQLLQGLAFCHSHRVLHRDLKPQNLLINTECDK1 SLQDVLMQ-----DSRLYLIFEFLSMDLKKYLDS-IPPGQYMDSSLVKSYLYQILQGIVFCHSRRVLHRDLKPQNLLIDDKCDK4 RLMDVCATSRTDREIKVTLVFEHVDQDLRTYLDKAPPPG--LPAETIKDLMRQFLRGLDFLHANCIVHRDLKPENILVTSGCDK6 RLFDVCTVSRTDRETKLTLVFEHVDQDLTTYLDKVPEPG--VPTETIKDMMFQLLRGLDFLHSHRVVHRDLKPQNILVTSSCDK7 GLLDAFGH-----KSNISLVFDFMETDLEVIIKD---NSLVLTPSHIKAYMLMTLQGLEYLHQHWILHRDLKPNNLLLDENERK2 GINDIIRAPTIEQMKDVYIVQDLMETDLYKLLKT---QH--LSNDHICYFLYQILRGLKYIHSANVLHRDLKPSNLLLNTTKPCa QLHSCFQT-----VDRLYFVMEYVNGGDLMYHIQ--QVG-KFKEPQAVFYAAEISIGLFFLHKRGIIYRDLKLDNVMLDSEKAPa KLEFSFKD-----NSNLYMVMEYVPGGEMFSHLR--RIG-RFSEPHARFYAAQIVLTFEYLHSLDLIYRDLKPENLLIDQQKPBb RLYRTFKD-----SKYLYMLMEACLGGELWTILR--DRG-SFEDSTTRFYTACVVEAFAYLHSKGIIYRDLKPENLILDHREGFR RLLGICLT-----S-TVQLITQLMPFGCLLDYVREHKDN--IGSQYLLNWCVQIAKGMNYLEDRRLVHRDLAARNVLVKTPSRC QLYAVVSE------EPIYIVTEYMSKGSLLDFLKG-ETGKYLRLPQLVDMAAQIASGMAYVERMNYVHRDLRAANILVGEN : : :: : . . : : . ::*** *:::

b8---: aL12:L12------------------------------:a4-------------: 140 150 160 170 180 190CDK2 GAIKLADFGLARAFGVPVR---TYTHEVVTLWYRAPEILLGCKYYSTAVDIWSLGCIFAEMVTR CDK1 GTIKLADFGLARAFGIPIR---VYTHEVVTLWYRSPEVLLGSARYSTPVDIWSIGTIFAELATK CDK4 GTVKLADFGLARIYSYQM----ALTPVVVTLWYRAPEVLLQST-YATPVDMWSVGCIFAEMFRR CDK6 GQIKLADFGLARIYSFQM----ALTSVVVTLWYRAPEVLLQSS-YATPVDLWSVGCIFAEMFRR CDK7 GVLKLADFGLAKSFGSPNR---AYTHQVVTRWYRAPELLFGARMYGVGVDMWAVGCILAELLLR ERK2 CDLKICDFGLARVADPDHDHTGFLTEYVATRWYRAPEIMLNSKGYTKSIDIWSVGCILAEMLSN KPCa GHIKIADFGMCKEHMMDG---VTTRTFCGTPDYIAPEIIAYQP-YGKSVDWWAYGVLLYEMLAG KAPa GYIQVTDFGFAKRVKGRTW------TLCGTPEYLAPEIILSKG-YNKAVDWWALGVLIYEMAAG KPBb GYAKLVDFGFAKKIGFGKK----TWTFCGTPEYVAPEIILNKG-HDISADYWSLGILMYELLTG EGFR QHVKITDFGLAKLLGAEEKEYHAEGGKVPIKWMALESILHR--IYTHQSDVWSYGVTVWELMTF SRC LVCKVADFGLARLIEDNEYT-ARQGAKFPIKWTAPEAALYGR--FTIKSDVWSFGILLTELTTK :: ***:.: : . * *: * . *:

Fig. (20). Sequence alignment of ATP-binding domains of various kinases. ClustalW multiple sequence alignment wasperformed using the following sequences: CDK2, CDK2-human cell division protein kinase 2, sp|P24941|; CDK1, CC2-human cell division control protein 2 homolog, sp|P06493|; CDK4, CDK4-human cell division protein kinase 4 (EC 2.7.1.-)sp|P11802|; CDK6, CDK6-human cell division protein kinase, sp|Q00534|; CDK7, CDK7-human cell division protein kinase7, sp|P50613|; ERK2, ERK2-human extracellular signal-regulated kinase 2, (mitogen-activated protein kinase 2), sp|P28482|;KPCa, KPCA-human protein kinase C, alpha type, sp|P17252|; KAPa, KAPA-human cAMP-dependent protein kinase, alpha-catalytic subunit, sp|P17612|; KPBb, KGPB-human cGMP-dependent protein kinase, beta isozyme, sp|P14619|; EGFR,EGFR-human epidermal growth factor receptor precursor, sp|P00533|; SRC, SRC-human proto-oncogene tyrosine-proteinkinase SRC (P60-SRC) (C-SRC), sp|P12931|. Sequence positions known to be involved in ATP interactions in the case ofCDK2 are shaded [106]. The secondary structure designations refer to CDK2, as well.

oxygen of Glu81. Consistent with the SAR seen forN9 substituents in the purine analogues, this grouppacks into a small hydrophobic pocket involvingVal18, Ala31, Phe80, Leu134, and Ala144. As is thecase for olomoucine and roscovitine, purvalanolB’s C2 side chain is bound in the ATP ribosebinding pocket, although its precise position isdifferent: the R-isopropyl group makes numeroushydrophobic contacts with the Gly-rich L2 loop ofCDK2 and the hydroxyl forms an H-bond to thebackbone carbonyl of Gln131. This results in anopen pocket in the active site lined by polar sidechains of Lys33, Asn132, and Asp145. Since this

region in the flavonoid/CDK2 structure is occupiedby the N-methylpiperidine ring [109], it has beenspeculated that further increases in affinity of thepurvalanols may result from appending additionalsubstituents at the C2 position [70]. In all threepurine/CDK2 complexes, the aromatic groups onN6 bind in a region outside the conserved ATPbinding site. This region is also occupied by thephenyl group in the flavonoid/CDK2 complex. Theinteractions in question are hydrophobic in natureand involve mostly His84, Phe82, Lys89, Glu8, andIle10 in CDK2. It is thought that interactions in thisregion may be responsible for the increased CDK


selectivity of certain purines and flavonoids. In thestaurosporin/CDK2 complex structure the C1-containing phenyl ring binds in this region.However, it does not extend out of the cleft as faras e.g. the olomoucine N6 benzyl ring and does notmake any comparable significant contacts except avan der Waal’s interaction with Ile10 [43]. Anadditional interaction between the chloro group andthe side chain of Asp86 is observed in the case ofpurvalanol B. Two different orientations observed,involving ca. 160o flipping of the chlorophenylring, suggest a partially protonated state of Asp86.Numerous para-substituents in the N6-anilinopurine series are tolerated; this is consistent with thesolvent accessibility of that site and may provide anopportunity for the fine-tuning of the inhibitor’sphysico-chemical properties.

ATP-binding cleft more fully than either ATP orstaurosporin. Although not precisely mimicking theATP α-phosphate group, the sulphonate group inindirubin-5-sulphonic acid also interacts with theside chain of Lys33. Furthermore, the sulphonateoxygen contacts the backbones of Asn132 andAsp145. The monoxime group of indirubin-3’-monoxime occupies the ATP ribose-binding siteand makes no direct interactions with CDK2.

Staurosporin

This large hydrophobic kinase inhibitor becomesmostly solvent inaccessible when bound to CDK2[43]. Again the familiar H-bonds to Glu81 and fromLeu83 are observed, as well as numeroushydrophobic contacts. As pointed out above, lackof significant interactions of staurosporin with thehinge loops L2 and L7 may account for itspromiscuity with respect to kinase inhibition. Themain difference between the CDK2 complexes withstaurosporin on the one hand and ATP/otherinhibitors on the other hand appears to be theamount of apolar surface area buried. Thusformation of the CDK2/staurosporin complexburies 545 Å2 apolar surface (274 Å2 contributed byCDK2, 272 Å2 by staurosporin), compared with265 Å2 apolar surface buried in the ATP/CDK2complex (175 Å2 contributed by CDK2, 90 Å2 byATP). The staurosporin/CDK2 shows that onlycarbon atoms 1 – 3, 22, 24, & 26 of staurosporinretain any accessible surface area and that this isappreciable (16 Å2) only for C1. The possibilities todesign derivatives of staurosporin with alteredspecificity would therefore seem limited.

Flavonoids

The differences in the binding modes of closelyrelated purines is mirrored in the flavonoidsituation, where myrecetin apparently binds in theATP binding site of CDK2 in a reverse orientationcompared to flavopiridol [2]. The benzopyran ringof L86-8276 occupies about the same region as thepurine ring of ATP. The two rings overlap in thesame plane in such a way that the carbonyl O4 andhydroxyl O5 are close to the positions of the N6 andN1 atoms of the ATP adenine ring. These twooxygens are involved in H-bonds with Phe82/Leu83

and Glu81, respectively. Two of these residues arealso involved in H-bonding to the purine rings inolomoucine, roscovitine, and purvalanol B.Numerous hydrophobic interactions between thebenzopyran ring and CDK2 (mainly Ile10 andLeu134) are also observed. The positions of the N-methylpiperidine and phenyl rings were discussedabove in relation to the purines. The flavopiridolmolecule currently in clinical trials has achlorophenyl in place of the phenyl group in L86-8276. This modification increases kinase inhibitionby a factor of around 7. This is probably due tonew possible contacts between the chlorine andresidues Ile10, Phe82, and Leu83.

Apart from almost completely non-selectivekinase inhibitors, flavopiridol appears so far to bethe only CDK-specific ATP-antagonist with highinhibitory activity against CDK4 and CDK6. Todate no compounds have been reported with uniqueselectivity for CDK4/6, although Garrett andFattaey [1] claim to have found such an inhibitor byhigh-throughput screening of a large chemicallibrary; unfortunately neither chemical structure norany other details were given. Althoughflavopiridol’s CDK4/6-selectivity may well besignificant for its superior anti-tumour effects, thisselectivity is difficult to rationalise on the basis ofavailable structural information. Presumably thisquestion will not be answered until a high-resolution crystal structure of CDK4 or CDK6 incomplex with an appropriate inhibitor has beensolved.

Indigoids

Both indirubin 3’-monoxime and indirubin-5-sulphonic acid have been co-crystallised withCDK2 [36]. Like other small-molecule CDKinhibitors, they bind at the ATP binding site. Thefamiliar H-bonds involving the backbone carbonylof Glu81, as well as both the backbone NH andcarbonyl of Leu83, are seen in these structures aswell, here the bonding partners are the lactamnitrogen, lactam oxygen, and the indole nitrogen,respectively. Both indigoids fill the back of the

The fact that the receptor affinity, and hence thepotency, of ATP antagonists can potentially beimproved with the aid of structure-based design of


novel antagonists is indicated by the following:ATP itself has comparatively poor affinity for thecatalytic site of CDKs; the ATP binding pocket isnot fully occupied by ATP or any of the knowninhibitors, apart from the non-selective kinaseinhibitor staurosporin. From the relevant complexcrystal structures it has been estimated that ATP,olomoucine, flavopiridol, and purvalanol B occupyca. 78, 76, 74, and 86 %, respectively [82].

While both cyclin D - CDK4/6 and cyclin A/E -CDK2 kinases phosphorylate pRb, theirmacromolecular substrate specificity and preferredphosphorylation sites appear to be different.CDK4/6 preferentially phosphorylate Ser780 of pRbin a cell-cycle-dependent manner, leading to theinability of pRb to associate with E2F transcriptionfactors [113]. The INK4a gene is inactivated in alarge percentage of tumour cell lines and loss ofp16INK4a function is second only to inactivation ofp53 as the most frequent event in human tumours.Clearly p16INK4a thus acts as a tumour suppressor.It is now known that in mammalian organisms theINK4 family is composed of four closely relatedmembers: p16INK4a, p15INK4b, p18INK4c, andp19INK4d [114]. Sequence homology betweenp15INK4b and p16INK4a is ca. 80 %, whereasp18INK4c and p19INK4d are ca. 40 – 50 % identicalwith p16INK4a and with each other. The fourproteins have similar biochemical properties in thatall of them bind to and inhibit the kinase activitiesof CDK4 and CDK6, as well as their complexeswith cyclins D1 and D2. The four proteins do not

PROTEIN AND PEPTIDE INHIBITORSOF CDKS

INK4 CKIs

The 156 amino-acid residue protein p16INK4a

was first identified and demonstrated to be aspecific inhibitor of the CDK4/cyclin D kinases in1993. [111]. Inhibition of cell cycle progressionthrough the G1 phase by this protein was found tobe associated with suppression of phosphorylationof the CDK4 and CDK6 substrate pRb [112].

Fig. (21) . The high-resolution (1.9 Å) crystal stucture complex between p19 INK4d and CDK6 [118] is depicted. The secondarystructure of p19INK4d is shown in green, that of CDK6 in blue. Residues of p19INK4d which make contacts with CDK6 arecoloured white. In the CDK6 structure, the PLSTIRE helix is coloured light-blue, key residues of the catalytic cleft are shownwith side chains (red). The catalytic triad (Lys43 , Glu61 , and Asp163) is shown in yellow, the T loop in violet.


appear to be functionally redundant, however, astheir transcription seems to be regulateddifferentially, e.g. during senescence and inresponse to oncogenic stimuli (p16INK4a), inresponse to the growth-inhibitory factor TGFβ(p15INK4b), and particularly during S-phase. TheINK4 proteins contain four to five consecutiveankyrin repeats (motifs particularly associated withprotein-protein interactions [115]) stacked linearlyas helix bundles and consisting of a helix-turn-helixmotif, successive repeats being linked by β-structures. The structure of p16INK4a has beenfound to be both thermodynamically and kineticallyunstable, rendering the protein vulnerable to single-point mutations. This vulnerability provides oneexplanation for the striking frequency of tumour-associated p16INK4a mutations [116].

cyclin binding site, refer to Fig. (21). The bindingof cyclins to CDK4 or CDK6 is thus preventedindirectly through allosteric changes induced byINK4 CKI binding. These changes cause the N-and C-lobes of the kinase to twist away from eachanother, leading to a misalignment of these lobes,as well the PLSTIRE helix (PSTAIRE in CDK2)and the T loop. While both INK4 CKIs and cyclinsneed to interact with both lobes of the CDK fortheir biological function, they require them indifferent relative orientations [7]. Whenever INK4CKIs are present in vivo, they appear to dominateover the cyclins with respect to the interaction withCDK4/6. Additionally, the binding of INK4 CKIsto CDK4 or CDK6 appears to lead to distortion ofthe ATP-binding site. Portions of the catalytic cleftwhich normally make contact with ATP aredislodged upon CKI binding in such a way as tosuggest a significant decrease in the affinity of thekinases for ATP and to disorient severely boundATP with respect to catalysis. This fact, togetherwith the finding that INK4 CKIs do not directly

The structures of CDK6 bound to p16INK4a [117]and the closely related p19INK4d [118] CKIs showthat the INK4 inhibitors bind next to the ATP-binding site of the catalytic cleft, opposite to the

Fig. (22) . Interface between p16INK4a and CDK6.The illustration was constructed from the complex crystal structure [117].Residues 81 – 104 only of p16INK4a (red) are shown with the key contacts to CDK6 surface residues (blue/green) coloured pink.The three key interactions involve Asp92 , Phe90 , and Asp84 /Arg87 in p16INK4a, respectively forming a hydrogen bond withLys111, non-polar side-chain contacts with the two active site back wall loops Asp102-Gln103 & Thr154-Ser155, and beinginvolved in a hydrogen-bond network centred on Arg31 in CDK6.


occupy the cyclin binding sites on CDK4 andCDK6, explains why INK4 CKIs are neverthelesscapable of binding and inhibiting both uncomplexedCDKs, as well as their complexes with cyclins.

KIP/CIP CKIs

Members of the KIP/CIP family of tumoursuppressor proteins, which includes p27KIP1,p21CIP1/WAF1 and p53KIP2 [125,126], contain a 65-residue region with 38 – 44 % homology (refer toFig. (23)) in their N-terminus which is necessaryand sufficient to bind to and to inhibit cyclin/CDKcomplexes with broad specificity for the G1 and Sphase kinases over the mitotic ones. The functionof these cyclin-binding domains in CKIs and otherproteins will be discussed in more detail below.Their C-terminal domains vary in length, sequence,and function [127]. Furthermore, they can bindcyclins and CDKs independently, but they havehigher affinity for the preformed cyclin/CDKcomplexes. The only crystal structure available of aKIP/CIP CKI in complex with a CDK/cyclin is thatof an N-terminal fragment of p27KIP1 andCDK2/cyclin A [128]. The fact that p27KIP1 bindsto both the CDK and cyclin partner is depicted inFig (24).

Prior to elucidation of the INK4 CKI/CDK6complex crystal structures a systematic study [119]using overlapping 20mer synthetic peptidesspanning the entire sequence of p16INK4a suggestedthat residues 84 – 103 were directly involved inCDK4/6 binding. Furthermore, the peptidesubtending these residues was found to be capableof inhibiting pRb phsophorylation in vitro.Moreover, a linear construct of this peptide with thecellular delivery vector derived form the Drosophilaantennapedia homeodomain (residues 43 – 58)[120] mediated the same effect in vivo, as well asblocking cell-cycle progression into S phasethrough a pronounced G1 arrest. It is worth notingthat an independent study observed similar effectswith a recombinant form of the full-length p16INK4a

protein fused to the same antennapedia peptide[121]. Further optimisation of the p16INK4a(84 –103) peptide [122] identified residues 90 - 97 asbeing particularly important to the peptide’sbiological function. These findings correlate wellwith the mechanism proposed for CDK4/6inhibition by INK4 CKIs summarised in Fig. (22),which suggests a prominent role for residues in theα5 and α6 helices, as well as loop 6 of the CKIs(in particular Asp84, Arg87, Phe90, and Asp92).Structural and functional analysis of a number ofsite-specifically mutated p16INK4a analogues alsostressed the importance of the third helix-turn-helixmotif (mutants F90A, V95A, V96A, and R99A)[123]. In terms of peptide mimetics of p16INK4a itcan be imagined that the hydrogen-bondinteractions from Asp84/Arg87 and Asp92 to CDK4/6may serve as initial binding points leading topositioning of the Phe90 side chain, which plays acentral role in the dislocation of the ATP-interactingloop in the catalytic cleft.

The obvious way in which p27KIP1 inhibitsCDK2 is by insertion of a small 310-helix into thecatalytic cleft. The details of this interaction areshown in Fig. (25). A tyrosine side chain ofp27KIP1, conserved amongst all CIP proteins(boxed residue in Fig. (23)), mimics the apolarcontacts made by the ATP purine ring, as well asthe H-bonds made by N1 and N6 of ATP. At the C-terminus of the helix, backbone carbonyl groupsform H-bonds to an active site Lys side chain alsoinvolved in H-bonds to the phosphate groups inATP [7]. These interactions clearly show thatinsertion of the 310-helix into the catalytic cleftdirectly blocks ATP binding.

Even in the absence of the 310-helix, however,p27KIP1 is still capable of inhibiting CDK2. This isa result of the distortion of the catalytic cleftinduced by binding of the CKI to the CDK [7].

Studies with p21WAF1-derived synthetic peptides[129,130] clearly showed that efficient CDKinhibition, as well as inhibition of cell proliferationthrough cell cycle arrest, could be achieved usingthe protein-protein interaction between the CKI andcyclin alone. Here the mechanism involvesprevention of the macromolecular CDK substratepRb from binding to the CDK/cyclin complex insuch a way as to permit its phosphorylation. Thecyclin-binding motif is strongly conserved amongstthe CIP/KIP CKIs (refer to Fig. (23)) proteins.Only p21WAF1 contains two copies of this motif butthe biological significance of this finding remainsunclear. Furthermore, it is now clear that other

It can thus be concluded that the interactionssummarised in Fig. (22) represent a promisingstarting point for the rational design of selectivepeptidomimetic small-molecule CDK4/6 inhibitorswith potential as novel cancer therapeutics. Furtherinterest in INK4 CKIs as a potential source for thedevelopment of cancer therapeutics stems from thefact that e.g. both full-length p16INK4a, as well asthe CDK6-inhibitory peptides derived from itdiscussed above, have been shown [124] to be ableto inhibit αvβ3 integrin-mediated cell spreading, aphenomenon closely associated with tumour-relatedactivities such as neovascularisation and tumourmetastasis.


10 20 30 40 50p27KIP1 MSNVRVSNGSPSLERMDAR--QAEHPKPSACRNLFGPVDHEELTRDLEKHCRDMEEASQRp57KIP2 MSDASLRSTS-TMERLVARGTFPVLVRTSACRSLFGPVDHEELSRELQARLAELNAEDQNp21WAF1 ----------MSEPAGDVR---QNPCGSKACRRLFGPVDSEQLSRDCDALMAGCIQEARE : .* ..*** ****** *:*:*: : :.

60 70 80 90 100p27KIP1 KWNFDFQNHKPLEG--KYEWQEVEKGSLPEFY-----------YRPPRPPKGACKV----p57KIP2 RWDYDFQQDMPLRGPGRLQWTEVDSDSVPAFYRETVQVGRCRLLLAPRPVAVAVAVSPPLp21WAF1 RWNFDFVTETPLEG--DFAWERVRGLGLPKLY------------LPTGPRRGRDELG--- :*::** . **.* * .* .:* :* .. * :

110 120 130 140 150p27KIP1 -PAQESQD----VSGSRPAAPLIGAPANSEDTHLVDPKTDPSDS--QTGLAEQCAGIRKRp57KIP2 EPAAESLDGLEEAPEQLPSVPVPAPASTPPPVPVLAPAPAPAPAPVAAPVAAPVAVAVLAp21WAF1 -------------GGRRPGTSPALLQGTAEEDHVDLSLSCTLVP---------R-SGEQA *... ... : . . . .

160 170 180 190p27KIP1 PATDDSSTQNKRANRTEENVSDGSPNAGSVEQTPKKPGLRRRQTp57KIP2 PAPAPAPAPAPAPAPVAAPAPAPAPAPAPAPAPAPAPDAAPQESp21WAF1 EGSPGGPGDSQGRKRRQTSMTDFYHSKRRLIFSKRKP------- .. .. . . *

Fig. (23) . Multiple sequence alignment (ClustalW) of CIP/KIP CKIs. The numbering is based on the p27KIP1 sequence. Thep57KIP2 sequence is truncated with the actual C-terminus of p27KIP1. The conserved N-terminal cyclin-binding motif is shaded,as well as the additional C-terminal copy of the motif present in p21WAF1 . The conserved Tyr residue implicated in interactionwith the CDK active site is boxed.

proteins known to interact with cyclins do so byusing the same binding motif. These proteinsinclude the E2F transcription factors, as well as thepRb-related CDK substrates [131,132]. Analignment of the appropriate sequence portions isshown in Fig. (26). Cell-permeable forms ofpeptides containing this motif have been shown tobe able to induce preferentially transformed cells toundergo apoptosis relative to non-transformed cells[133]. The cyclin-binding motif is relevant to cyclinA/CDK2, cyclin E/CDK2, cyclin D/CDK4 but doesnot appear to be involved in cyclin B/CDK1complexes.

residues in this helix involved are highly conservedin the cyclin box [134].

Peptide Library Approaches

Since they contain numerous protein-proteininteraction sites whose inhibition with e.g. peptideligands can potentially lead to inactivation of kinaseactivity, CDKs represent particularly suitabletargets for genetic [135] or chemical [136] peptidelibrary approaches. One such approach, involvingpeptide aptamers, i.e. random-sequence peptidesdisplayed within the scaffold of a suitable protein,was reported recently [137]. Peptide aptamerstrategies are particularly attractive since they permitintra-cellular library deconvolution by two-hybridgenetic engineering methods. A number ofaptamers containing 20-residue variable regionsinserted in the active site of Escherichia colithioredoxin were thus shown to bind tightly toCDK2 and to inhibit protein function in vitro. Oneaptamer inhibited CDK function with distinctsubstrate specificity for histone H1 over pRb andits expression in mammalian cells retardedprogression through G1 phase. The aptamer wasshown to bind at or near the active site of CDK2.More important from a drug design point of view, itwas also shown that the aptamer’s variable regionwas sufficient for biological activity [138]. Thus a20-residue synthetic peptide containing this regioninhibited the phosphorylation by CDK2/cyclin E ofhistone H1 but not of pRb. Inhibition of the histoneH1 kinase activity had an IC50 value of 7 µM. Thiswas ca. 1000-fold higher than inhibition by the

The finding that blocking the interaction ofsubstrates with cyclin/CDK complexes is sufficientfor CDK inhibition offers a different approach toATP antagonism for the development of CDK-selective cancer therapeutics. The crystal structureof the p27KIP1/CDK2/cyclin A complex shows thatthe interaction of the above cyclin-binding motifinvolves a groove on the surface of cyclin, formedmainly by two α-helices, see Fig. (27). Thep27KIP1 binding coil has a rigid conformation. At itsN-terminus, it starts with a reverse turn (residues26 – 29), followed by an extended region (residues29 – 31), and ends with another turn, stabilised byan intramolecular H-bond. The first turn and theextended region interact with cyclin A mainlythrough H-bonds and apolar contacts involvingAla28. The second turn binds more deeply into ahydrophobic pocket in the cyclin groove. Multiplevan der Waal’s contacts are observed mainly to theLeu-Phe pair of p27KIP1. The contacts to the α1helix are particularly significant, because the


Fig. (24). Interactions between p27KIP1 and CDK2/cyclin A. The complex structure determined at 2.3 Å by X-raycrystallography [128] is shown: the protein backbone of p27KIP1, CDK2, and cyclin A are coloured, green, blue, and lightgreen, respectively. The portions of p27KIP1 making contact with cyclin A and CDK2 are shown in red and magenta,respectively.

corresponding thioredoxin aptamer. Clearly in thiscase the conformational constraint imposed by thescaffold upon the variable region was a primedeterminant of potency.

remains largely obscure [140]. The crystal structureof a CKS protein in complex with CDK2 wassolved recently [141] and the stage is thus set forthe rational design of inhibitors of CKS binding. Itwill be interesting to find out what biologicalfunctions such inhibitors possess.

Inhibitors of Other Regulatory Protein-Protein Interaction Sites of CDKs: FutureDirections Gene Transfer and Antisense Technologies

Other regulators of CDK activity which areamenable in principle to the development oftherapeutic agents include the proteins whichregulate the phosphorylation state of CDKs such asCAK, CDC25 phosphatases, as well as the WEE1and MYT1 kinases [2,8]. Furthermore, althoughmuch is now known about the molecularrecognition events leading to partial activation ofCDKs by their cyclin partners [139,7,107],inhibition of these events as the basis for drugdesign has apparently not yet been exploited. CKSproteins are small subunits which are essentialcomponents of the CDKs which regulate mitosis.However, the precise function of the CKS proteins

Loss of checkpoint control due to defects in keyproteins including pRb, p53, p16INK4a, p21WAF1,and p27KIP1, is a hallmark of cancer and thus anattractive target for therapeutic intervention [142].Gene therapy approaches, particularly thoseutilising replication-defective recombinantadenovirus vectors encoding CKIs, have recentlyshown some promise [143]. For example, achimeric p27/p16 tumour suppressor vector geneinduced cell growth arrest and apoptosis whenoverexpressed in certain tumour cell lines, as wellas showing therapeutic effect in several in vivoanimal tumour models [144]. Various approacheshave been applied using antisense


Fig. (25) . Details of the interaction between the p27KIP1 310 -helix and the catalytic cleft of CDK2. The backbone of CDK2 isshown in blue. CDK2 residues involved in binding residues Phe87 and Tyr88 of p27KIP1 are shown with their side chains:yellow for those also involved in binding the purine ring of ATP, green for Lys33 normally involved in binding the phosphatemoiety of ATP, and blue for Ala144 not directly involved in ATP binding. The backbone of residues in p27KIP1 forming part ofthe 310 -helix but not directly involved in contacts to CDK2 are shown in red.

oligodeoxynucleotides in order to target CDKactivity [1,2,143]. The main challenge whichcontinues to face antisense technologies in generalconcerns the problems of delivering suitableoligodeoxynucleotides to living cells.

CDK INHIBITION AND APOPTOSIS

Most of the inhibitors of CDKs discussed abovecan arrest proliferating cells at either the G1-Sand/or the G2-M boundaries and are capable undercertain circumstances of inducing apoptosis in a

p21WAF1 (C-terminal) F Y H S K R R L I F S K

p21WAF1 (N-terminal) G S K A C R R L F G P V

p21KIP1 K P S A C R N L F G P V

p57KIP2 R T S A C R S L F G P V

E2F1 R P P V K R R L D L E T

E2F2 R L P A K R R L D L E G

E2F3 G P P A K R R L E L G E

p107 pRBL1 A G S A K R R L F G E D

p130 pRBL2 A S T T R R R L F V E N

pRb P P K P L K K L R F D I

Fig. (26) . Cyclin-binding motif in CIP/KIP CKIs, E2Fs, and pRBs.


Fig. (27) . Details of the interaction between p27KIP1 and cyclin A. Residues 205 – 301 of cyclin A and residues 25 – 37 ofp27KIP1 are shown only. The p27KIP1 residues making hydrophobic contacts are shown in white, those forming hydrogen bondsin red. The unusual structural element guiding the p27KIP1 out of the cyclin binding groove and permitting Phe33 to penetratedeeply into a hydrophobic pocket is shown in yellow. The backbone of the cyclin A residues making van der Waal’s orhydrogen bonding contacts with p27KIP1 are coloured cyan (α1 helix), violet (α3 helix) and orange.

p53- and pRb-independent manner (Ref. [1], andreferences cited therein). Due to these properties,CDK inhibitors are particularly attractive therapeuticagents for the treatment of cancers, however, anumber of other proliferative diseases involvingCDK function are also starting to be addressed withthese agents [17]. The apparent ability of certainCDK2 inhibitors selectively to induce apoptosis intumour cells has been surprising; however, apossible explanation involving inappropriateexpression or activity of the transcription factorsbelonging to the E2F family was recently proposed[139]. The E2F transcription factors are designed tobe active during only a narrow window of timeduring the normal cell cycle: While pRb remains inits normal hypophosphorylated state, it binds E2Fs,preventing them from promoting transcription. Latein the G1 phase, however, pRb becomesphosphorylated by G1 cyclin/CDK complexes andthe transcription-activating domains of E2Fs areunmasked. This leads to a burst of E2F-activatedtranscription. Subsequently, progression through Sphase is enabled by activation of cyclin A/CDK2

complexes. But these complexes alsophosphorylate E2Fs, which causes them todissociate from DNA. Cessation of E2F-promotedtranscription in the cell cycle is then achieved.Experiments with transgenic mice lacking the E2F-1 factor showed [145,146] that the E2F-1 genepossesses properties of a tumour suppressor and itis now clear that while high levels of E2F-1 favourpassage through the G1 phase [147], continued andinappropriate E2F-1 transcription-promotingactivity triggers apoptosis. Recent data suggeststhat even minor reduction of the cyclin A/CDK2-mediated shutdown of E2Fs will selectively directtumour cells, which commonly express higherlevels of active E2Fs than normal cells, towardsapoptosis [148].

REFERENCES

[1] Garrett, M. D.; Fattaey, A. Curr. Op. Gen. Dev.1999 , 9, 104.


[2] Walker, D. H. Curr. Top. Microbiol. Immunol.1998 , 149.

[23] Kawakami, K.; Futami, H.; Takahara, J.; Yamaguchi,K. Biochem. Biophys. Res. Commun. 1996 , 219,778.

[3] Meijer, L.; Kim, S.-H. Methods Enzymol. 1997 ,283, 113. [24] Ohashi, M.; Sugikawa, E.; Nakanishi, N. Jpn. J.

Cancer Res. 1995 , 86, 819.[4] Gray, N.; Détivaud, L.; Doerig, C.; Meijer, L. Curr.

Med. Chem. 1999 , 6, 859. [25] Meijer, L. Prog. Cell Cycle Res. 1995, 1, 351.

[5] Morgan, D. O. Nature 1995 , 374, 131. [26] Bojanowski, K.; Nishio, K.; Fukuda, M.; KraghLarsen, A.; Saijo, N. Biochem. Biophys. Res.Commun. 1994 , 203, 1574.[6] Russo, A. A.; Jeffrey, P. D.; Pavletich, N. P. Nature

Struct. Biol. 1996 , 3, 696.[27] Kitagawa, M.; Okabe, T.; Ogino, H.; Matsumoto,

H.; Suzuki-Takahashi, I.; Kokubo, T.; Higashi, H.;Saitoh, S.; Taya, Y.; Yasuda, H.; Ohba, Y.;Nishimura, S.; Tanaka, N.; Okuyama, A. Oncogene1993 , 8, 2425.

[7] Pavletich, N. P. J. Mol. Biol. 1999 , 287, 821.

[8] Johnson, D. G.; Walker, C. L. Annu. Rev.Pharmacol. Toxicol. 1999 , 39, 295.

[9] Konstantinidis, A. K.; Radhakrishnan, R.; Gu, F.;Rao, R. N.; Yeh, W.-K. J. Biol. Chem. 1998 , 273,26506.

[28] Nishio, K.; Ishida, T.; Arioka, H.; Kurokawa, H.;Fukuoka, K.; Nomoto, T.; Fukumoto, H.; Yokote,H.; Saijo, N. Anticancer Res. 1996 , 16, 3387.

[10] Kaldis, P. Cell. Mol. Life Sci. 1999 , 55, 284. [29] Vesely, J.; Havlicek, L.; Strnad, M.; Blow, J. J.;Donella-Deana, A.; Pinna, L.; Letham, D. S.; Kato,J.-y.; Detivaud, L.; Leclerc, S.; Meijer, L. Eur. J.Biochem. 1994 , 224, 771.

[11] Hunter, T.; Pines, J. Cell 1994 , 79, 573.

[12] Weinberg, R. A. Cell 1995 , 81, 323.

[30] Abraham, R. T.; Acquarone, M.; Andersen, A.;Asensi, A.; Belle, R.; Berger, F.; Bergounioux, C.;Brunn, G.; Buquet-Fagot, C.; Fagot, D.; Glab, N.;Goudeau, H.; Goudeau, M.; Guerrier, P.; Houghton,P. J.; Hendriks, H.; Kloareg, B.; Lippai, M.; Marie,D.; Maro, M.; Meijer, L.; Mester, J.; Mulner-Lorillon, O.; Poulet, S. A.; Schierenberg, E.;Schutte, B.; Vaulot, D.; Verlhac, M. C. Biol. Cell1995 , 83, 105.

[13] Lukas, J.; Bartkova, J.; Rohde, M.; Strauss, M.;Bartek, J. Mol. Cell. Biol. 1995 , 15, 2600.

[14] Ohtsubo, M.; Theodoras, A. M.; Schumacher, J.;Roberts, J. M.; Pagano, M. Mol. Cell. Biol. 1995 ,15, 2612.

[15] Fattaey, A.; Booher, R. N. Prog. Cell Cycle Res.1997 , 3, 233.

[16] Senderowicz, A. M.; Headlee, D.; Stinson, S. F.;Lush, R. M.; Kalil, N.; Villalba, L.; Hill, K.;Steinberg, S. M.; Figg, W. D.; Tompkins, A.;Arbuck, S. G.; Sausville, E. A. J. Clin. Oncol.1998 , 16, 2986.

[31] Meijer, L.; Borgne, A.; Mulner, O.; Chong, J. P.;Blow, J. J.; Inagaki, N.; Inagaki, M.; Delcros, J. G.;Moulinoux, J. P. Eur. J. Biochem. 1997 , 243, 527.

[32] Gray, N. S.; Wodicka, L.; Thunissen, A.-M. W. H.;Norman, T. C.; Kwon, S.; Espinoza, F. H.; Morgan,D. O.; Barnes, G.; LeClerc, S.; Meijer, L.; Kim, S.-H.; Lockhart, D. J.; Schultz, P. G. Science 1998 ,281, 533.

[17] Meijer, L.; Leclerc, S.; Leost, M. Pharmacol. Ther.1999 , 82, 279.

[18] Cragg, G. M.; Boyd, M. R.; Christini, M. A.;Kneller, R.; Mays, T. D.; Mazan, K. D.; Newman,D. J.; Sausville, E. A. Spec. Publ. - R. Soc. Chem.1997 , 200, 1.

[33] Rosania, G. R.; Merlie, J.; Gray, N.; Chang, Y.-T.;Schultz, P. G.; Heald, R. Proc. Natl. Acad. Sci. USA1999 , 96, 4797.

[19] Collin, O.; Meijer, L. Pharmacol. Ther. 1999 , 82,165.

[34] Zaharevitz, D. W.; Gussio, R.; Leost, M.;Senderowicz, A. M.; Lahusen, T.; Kunick, K.;Meijer, L.; Sausville, E. A. Cancer Res. 1999 , 59,2566.[20] Boyd, M. R. In Cancer Drug Discovery and

Development. Vol. 2; Drug Development; PreclinicalScreening, Clinical Trial and Approval; Teicher, B.A., Ed.; Humana Press:, 1997; pp 23.

[35] Schultz, C.; Link, A.; Leost, M.; Zaharevitz, D. W.;Gussio, R.; Sausville, E. A.; Meijer, L.; Kunick, C.J. Med. Chem. 1999 , 42, 2909.

[21] Meijer, L. Trends Cell Biol. 1996 , 6, 393.[36] Hoessel, R.; Leclerc, S.; Endicott, J. A.; Nobel, M.

E. M.; Lawrie, A.; Tunnah, P.; Leost, M.; Damiens,E.; Marie, D.; Marko, D.; Niederberger, E.; Tang,

[22] Gadbois, D. M.; Hamaguchi, J. R.; Swank, R. A.;Bradbury, E. M. Biochem. Biophys. Res. Commun.1992 , 184, 80.


W.; Eisenbrand, G.; Meijer, L. Nature Cell Biol.1999 , 1, 60.

[53] Huang, B.-G.; Bobek, M. Carbohydrate Res. 1998 ,308, 319.

[37] Markovits, J.; Linassier, C.; Fossé, P.; Couprie, J.;Pierre, J.; Jacquemin-Sablon, A.; Saucier, J.-M.; LePecq, J.-B.; Larsen, A. K. Cancer Res. 1989 , 49,5111.

[54] Suhadolnik, R. J. Prog. Nucleic Acid Res. Mol. Biol.1979 , 22, 193.

[55] Sharma, M.; Bloch, A.; Bobek, M. Nucleosides &Nucleotides 1993 , 12, 643.

[38] Sedlacek, H. H.; Czech, J.; Naik, R.; Kaur, G.;Worland, P.; Losiewicz, M.; Parker, B.; Carlson, B.;Smith, A. Int. J. Oncol. 1996 , 9, 1143.

[56] Sharma, M.; Li, Y. X.; Ledvina, M.; Bobek, M.Nucleosides & Nucleotides 1995 , 14, 1831.

[57] Boutin, J. A. Int. J. Biochem. 1994 , 26, 1203.[39] Losiewicz, M. D.; Carlson, B. A.; Kaur, G.;Sausville, E. A.; Worland, P. J. Biochem. Biophys.Res. Commun. 1994 , 201, 589. [58] Graziani, Y.; Erikson, E.; Erikson, R. L. Eur. J.

Biochem. 1983 , 135, 583.[40] Omura, S.; Iwai, Y.; Hirano, A.; Nakagawa, A.;

Awaya, J.; Tsuchiya, H.; Takahashi, Y.; Masuma, R.J. Antibiotics 1977 , 30, 275.

[59] Traxler, P.; Green, J.; Mett, H.; Séquin, U.; Furet, P.J. Med. Chem. 1999 , 42, 1018.

[60] Naik, R. G.; Kattige, S. L.; Bhat, S. V.; Alreja, B.;de Souza, N. J.; Rupp, R. H. Tetrahedron 1988 , 44,2081.

[41] Takahashi, I.; Saitoh, Y.; Yoshida, M.; Sano, H.;Nakano, H.; Morimoto, M.; Tamaoki, T. J.Antibiotics 1989 , 42, 571.

[61] Harmon, A. D.; Weiss, U.; Silverton, J. V.Tetrahedron Lett. 1979 , 721.

[42] Wang, Q.; Worland, P. J.; Clark, J. L.; Carlson, B.A.; Sausville, E. A. Cell Growth Differentiation1995 , 6, 927. [62] Houghton, P. J.; Hairong, Y. Planta medica 1987 ,

53, 262.[43] Lawrie, A. M.; Noble, M. E.; Tunnah, P.; Brown, N.R.; Johnson, L. N.; Endicott, J. A. Nature Struct.Biol. 1997 , 4, 796.

[63] Carlson, B. A.; Dubay, M. M.; Sausville, E. A.;Brizuela, L.; Worland, P. J. Cancer Res. 1996 , 56,2973.[44] Kiriyama, N.; Nitta, K.; Sakaguchi, Y.; Taguchi, Y.;

Yamamoto, Y. Chem. Pharm. Bull. 1977 , 25, 2593. [64] Czech, J.; Hoffmann, D.; Naik, R.; Sedlacek, H.-H.Int. J. Oncol. 1995 , 6, 31.[45] Nitta, K.; Fujita, N.; Yoshimura, T.; Arai, K.;

Yamamoto, Y. Chem. Pharm. Bull. 1983 , 31, 1528. [65] Patel, V.; Senderowicz, A. M.; Pinto, D.; Igishi, T.;Raffeld, M.; Quintanilla-Martinez, L.; Ensley, J. F.;Sausville, E. A.; Gutkind, J. S. J. Clin. Invest.1998 , 102, 1674.

[46] Wada, M.; Hosotani, R.; Lee, J.-U.; Doi, R.;Koshiba, T.; Fujimoto, K.; Miyamoto, Y.; Tsuji, S.;Nakajima, S.; Okuyama, A.; Imamura, M. AnticancerRes. 1998 , 18, 2559. [66] Nath, J.; Rebhun, L. I. Exp. Cell Res. 1973 , 77,

319.[47] Suzuki, M.; Hosaka, Y.; Matsushima, H.; Goto, T.;Kitamura, T.; Kawabe, K. Cancer Lett. 1999 , 138,121.

[67] Meijer, L.; Pondaven, P. Exp. Cell Res. 1988 , 174,116.

[48] Bourgois, J.; Bourgois, M.; Formisyn, P. TrendsHeterocycl. Chem. 1991 , 2, 205.

[68] Rialet, V.; Meijer, L. Anticancer Res. 1991 , 11,1581.

[49] Loomis, C. R.; Bell, R. M. J. Biol. Chem. 1988 ,263, 1682.

[69] Havlicek, L.; Hanus, J.; Vesely, J.; Leclerc, S.;Meijer, L.; Shaw, G.; Strnad, M. J. Med. Chem.1997 , 40, 408.[50] Park, S. G.; Cheon, J. Y.; Lee, Y. H.; Park, J.-S.;

Lee, K. Y.; Lee, C. H.; Lee, S. K. Mol. Cells 1996 ,6, 679.

[70] Gray, N. S.; Wodicka, L.; Thunnissen, A.-M., W.H.;Norman, T. C.; Kwon, S.; Espinoza, F. H.; Morgan,D. O.; Barnes, G.; LeClerc, S.; Meijer, L.; Kim, S.-H.; Lockhart, D.; Schultz, P. G. Science 1998 , 281,533.

[51] Bobek, M.; Bloch, A. Nucleosides & Nucleotides1995 , 13, 429.

[52] Bobek, M.; Kramer, D.; Leclerc, S.; Meijer, L.;Porter, C. Proceedings 17th International CancerCongress, 1998, Rio de Janeiro, Brazil, Abstract No.427

[71] Imbach, P.; Capraro, H.-G.; Furet, P.; Mett, H.;Meyer, T.; Zimmermann, J. Bioorg. Med. Chem.Lett. 1999 , 9, 91.


[72] Brooks, E. E.; Gray, N. S.; Joly, A.; Kerwar, S. S.;Lum, R.; Mackman, R. L.; Norman, T. C.; Rosete,J.; Rowe, M.; Schow, S. R.; Schultz, P. G.; Wang,X.; Wick, M. M.; Shiffman, D. J. Biol. Chem.1997 , 272, 29207.

[88] Link, A.; Kunick, C. J. Med. Chem. 1998 , 41,1299.

[89] Kunick, C. Arch. Pharm. (Weinheim) 1991 , 324,579.

[90] Chen, D.; Xie, J. Zhongcaoyao 1984 , 15, 534.[73] Legraverend, M.; Ludwig, O.; Bisagni, E.; Leclerc,S.; Meijer, L. Bioorg. Med. Chem. Lett. 1998 , 8,793. [91] Wu, K.; Zhang, M.; Fang, Z.; Huang, L. Yaoxue

Xuebao 1985 , 20, 821.[74] Schulze-Gahmen, U.; Brandsen, J.; Jones, H. D.;

Morgan, D. O.; Meijer, L.; Vesely, J.; Kim, S.-H.Proteins: Structure, Function, and Genetics 1995 ,22, 378.

[92] Ji, X.; Zhang, F. Yaoxue Xuebao 1985 , 20, 137.

[93] Gu, Y. C.; Li, G. L.; Yang, Y. P.; Fu, J. P.; Li, C.Z. Yaoxue Xuebao 1989 , 24, 629.

[75] De Azevedo, W. F.; Leclerc, S.; Meijer, L.; Havlicek,L.; Strnad, M.; Kim, S. H. Eur. J. Biochem. 1997 ,243, 518.

[94] Li, C.; Go, Y.; Mao, Z.; Koyano, K.; Kai, Y.;Kanehisa, N.; Zhu, Q.; Zhou, Z.; Wu, S. Bull.Chem. Soc. Jpn. 1996 , 69, 1621.

[76] Oh, C. H.; Lee, S. C.; Lee, K. S.; Woo, E. R.;Hong, C. Y.; Yang, B. S.; Baek, D. J.; Cho, J. H.Arch. Pharm. (Weinheim) 1999 , 332, 187.

[95] Grant, S.; Boyle, F. T.; Calvert, A. H.; Curtin, N.J.; Endicott, J.; Golding, B. T.; Griffin, R. J.;Johnson, L. N.; Newell, D. R.; Noble, M. E. M.;Robson, C. Proc. Amer. Assoc. Cancer Res. 1998 ,39, Abstract 1207.

[77] Schow, S. R.; Mackman, R. L.; Blum, C. L.;Brooks, E.; Horsma, A. G.; Joly, A.; Kerwar, S. S.;Lee, G.; Shiffman, D.; Nelson, M. G.; Wang, X.;Wick, M. M.; Zhang, X.; Lum, R. T. Bioorg. Med.Chem. Lett. 1997 , 7, 2697.

[96] Arris, C. E.; Boyle, F. T.; Calvert, A. H.; Curtin, N.J.; Jewsbury, P. J.; Endicott, J. A.; Gibson, A. E.;Golding, B. T.; Grant, S.; Griffin, R. J.; Johnson, L.N.; Noble, M. E. M.; Newell, D. R. Proc. Amer.Assoc. Cancer Res. 1999 , 40, Abstract 2021.

[78] Legraverend, M.; Ludwig, O.; Bisagni, E.; Leclerc,S.; Meijer, L.; Giocanti, N.; Sadri, R.; Favaudon, V.Bioorg. Med. Chem. 1999 , 7, 1281. [97] Ruetz, S.; Woods-Cook, K.; Solf, R.; Meyer, T.;

Zimmermann, J.; Fabbro, D. Proc. Amer. Assoc.Cancer Res. 1998 , 39, Abstract 3796.

[79] Imbach, P.; Capraro, H. G.; Furet, P.; Mett, H.;Meyer, T.; Zimmermann, J. Bioorg. Med. Chem.Lett. 1999 , 9, 91. [98] Meyer, T.; Zimmermann, J.; Geiger, T.; Mett, H.;

Buchdunger, E.; Müller, M.; O'Reilly, T.; Cozens,R.; Ruetz, S.; Fabbro, D. Proc. Amer. Assoc. CancerRes. 1998 , 39, Abstract 3794.

[80] Hocart, C. H.; Letham, D. S.; Parker, C. W.Phytochemistry 1991 , 30, 2477.

[81] Gray, N. S.; Kwon, S.; Schultz, P. G. TetrahedronLett. 1997 , 38, 1161. [99] Duvadie, R. K.; Chong, W. K. M.; Li, L.; Chu, S.

S.; Yang, Y. M.; Nonomiya, J.; Tucker, K. D.;Lewis, C. T.; Knighton, D. R.; Ferre, R. A.;Lundgren, K.; Koudriakova, T.; Escobar, J.; Price, S.M.; Huber, A.; Sisson, W.; Aust, R. M.;Verkhivker, G. M.; Schaffer, L.; Rose, P. W. “NovelATP-site cyclin-dependent kinase (CDK) inhibitors:selective CDK inhibitors”; 218th ACS NationalMeeting, 1999, New Orleans, LA, USA. Abstract214.

[82] Chang, Y. T.; Gray, N. S.; Rosania, G. R.;Sutherlin, D. P.; Kwon, S.; Norman, T. C.; Sarohia,R.; Leost, M.; Meijer, L.; Schultz, P. G. Chemistryand Biology 1999 , 6, 361.

[83] Toyota, A.; Katagiri, N.; Kaneko, C. Heterocycles1993 , 36, 1625.

[84] Legraverend, M.; Boumchita, H.; Bisagni, E.Synthesis 1990 , 587. [100] Lundgren, K.; Price, S. M.; Escobar, J.; Huber, A.;

O'Connor, P.; Chong, W.; Duvadie, R.; Li, L.; Chu,S. S.; Nonmiya, J.; Tucker, K.; Mroczkowski, B.;Lewis, C. Proc. Amer. Assoc. Cancer Res. 1999 ,40, Abstract 2022.

[85] Kunick, C. Curr. Pharmaceut. Design 1999 , 5, 181.

[86] Kunick, C. Arch. Pharm. (Weinheim) 1992 , 325,297.

[101] Veal, J. M.; Bramston, N.; Dickerson, S.; Frye, S.;Harris, P.; Hassell, A.; Holmes, B.; Hunter, R.;Kuyper, L.; Lackey, K.; Lovejoy, B.; Luzzio, M.;Montana, V.; Rocque, W.; Shewchuk, L.; Walker, D.218th ACS National Meeting, 1999, New Orleans,LA, USA. Abstract 314.

[87] Paull, K. D.; Hamel, E.; Malspeis, L. In CancerChemotherapeutic Agents; Foye, W. O., Ed.;American Chemical Society: Washington, D.C.,1995; pp 9.


[102] Seitz, P.; Benfield, P. A.; Boylan, J.; Boisclair, M.;Brizuela, L.; Burton, C. R.; Carini, D. J.; Chang, C.H.; Cox, S.; Czerniak, P. M.; Grafstrom, R. H.;Hoess, R. H.; Muckelbauer, J. K.; Nugiel, D. A.;Rossi, K. A.; Trainor, G. L.; Worland, P.; Yue, E.W. 218th ACS National Meeting, 1999, NewOrleans, LA, USA. Abstract 316

[118] Brotherton, D. H.; Dhanaraj, V.; Wick, S.; Brizuela,L.; Domaille, P. J.; Volyanik, E.; Xu, X.; Parisini,E.; Smith, B. O.; Archer, S. J.; Serrano, M.;Brenner, S. L.; Blundell, T. L.; Laue, E. D. Nature1998 , 395, 244.

[119] Fåhraeus, R.; Paramio, J. M.; Ball, K. L.; Lain, S.;Lane, D. P. Curr. Biology 1996 , 6, 84.

[103] Owa, T.; Yoshino, H.; Okauchi, T.; Yoshimatsu, K.;Ozawa, Y.; Sugi, N. H.; Nagasu, T.; Koyanagi, N.;Kitoh, K. J. Med. Chem. 1999 , 42, 3789.

[120] Derossi, D.; Joliot, A. H.; Chassaing, G.;Prochiantz, A. J. Biol. Chem. 1994 , 269, 10444.

[121] Kato, D.; Miyazawa, K.; Ruas, M.; Starborg, M.;Wada, I.; Oka, T.; Sakai, T.; Peters, G.; Hara, E.FEBS Lett. 1998 , 427, 203.

[104] Raymond, E.; Fumoleau, P.; Roche, H.; Schellens,H. H. M.; Ravic, M.; Dittrich, C.; Punt, C. J. A.;Droz, J. P.; Armand, J.-P.; Calvert, A. H.; Wanders,J.; Hanauske, A.-R. Proc. Amer. Assoc. Cancer Res.1999 , 40, Abstract 2545. [122] Fåhraeus, R.; Lain, S.; Ball, K. L.; Lane, D. P.

Oncogene 1998 , 16, 587.[105] De Bondt, H. L.; Rosenblatt, J.; Jancarik, J.; Jones,

H. D.; Morgan, D. O.; Kim, S.-H. Nature 1993 ,363, 595.

[123] Li, J.; Byeon, I.-J. L.; Ericson, K.; Poi, M.-J.;O'Maille, P.; Selby, T.; Tsai, M.-D. Biochemistry1999 , 38, 2930.

[106] Schulze-Gahmen, U.; De Bondt, H. L.; Kim, S. H. J.Med. Chem. 1996 , 39, 4540. [124] Fåhraeus, R.; Lane, D. P. EMBO J. 1999 , 18, 2106.

[125] Lee, M. H.; Reynisdottir, I.; Massague, J. Genes Dev1995 , 9, 639.

[107] Jeffrey, P. D.; Russo, A. A.; Polyak, K.; Gibbs,E.; Hurwitz, J.; Massague, J.; Pavletich, N. P.Nature 1995 , 376, 313. [126] Matsuoka, S.; Edwards, M. C.; Bai, C.; Parker, S.;

Zhang, P.; Baldini, A.; Harper, J. W.; Elledge, S. J.Genes Dev. 1995 , 9, 650.

[108] Toledo, L. M.; Lydon, N. B. Structure 1997 , 5,1551.

[127] Sherr, C. J.; Roberts, J. M. Genes Dev. 1999 , 13,1501.

[109] De Azevedo, W. F., Jr.; Mueller-Dieckmann, H. J.;Schulze-Gahmen, U.; Worland, P. J.; Sausville, E.;Kim, S. H. Proc. Natl. Acad. Sci.USA 1996 , 93,2735.

[128] Russo, A. A.; Jeffrey, P. D.; Patten, A. K.;Massagué, J.; Pavletich, N. P. Nature 1996 , 382,325.[110] Zaharevitz, D. W.; Gussio, R.; Leost, M.;

Senderowicz, A. M.; Lahusen, T.; Kunick, C.;Meijer, L.; Sausville, E. A. Cancer Res. 1999 , 59,2566.

[129] Chen, J.; Saha, P.; Kornbluth, S.; Dynlacht, B. D.;Dutta, A. Mol. Cell. Biol. 1996 , 16, 4673.

[130] Ball, K. L.; Lain, S.; Fåhraeus, R.; Smythe, C.;Lane, D. P. Curr. Biol. 1996 , 7, 71.[111] Serrano, M.; Hannon, G. J.; Beach, D. Nature 1993 ,

366, 704.

[131] Adams, P. D.; Sellers, W. R.; Sharma, S. K.; Wu,A. D.; Nalin, C. M.; Kaelin, W. G., Jr. Mol. Cell.Biol. 1996 , 16, 6623.

[112] Lukas, J.; Parry, D.; Aagaard, L.; Mann, D. J.;Bartkova, J.; Strauss, M.; Peters, G.; Bartek, J.Nature 1995 , 375, 503.

[132] Adams, P. D.; Li, X.; Sellers, W. R.; Baker, K. B.;Leng, X.; Harper, J. W.; Taya, Y.; Kaelin, W. G. J.Mol. Cell. Biol. 1999 , 19, 1068.

[113] Kitagawa, M.; Higashi, H.; Jung, H.-K.; Suzuki-Takahashi, I.; Ikeda, M.; Tamai, K.; Kato, J.-y.;Segawa, K.; Yoshida, E.; Nishimura, S.; Taya, Y.EMBO J. 1996 , 15, 7060. [133] Chen, Y.-N. P.; Sharma, S. K.; Ramsey, T. M.;

Jiang, L.; Martin, M. S.; Baker, K.; Adams, P. D.;Bair, K. W.; Kaelin, W. G. Proc. Natl. Acad. Sci.USA 1999 , 96, 4325.

[114] Serrano, M. Exp. Cell Res. 1997 , 237, 7.

[115] Bork, P. Proteins: Structure, Function, and Genetics1993 , 17, 363. [134] Noble, M. E. M.; Endicott, J. A.; Brown, N. R.;

Johnson, L. N. Trends Biochem. Sci. 1997 , 22, 482.[116] Tang, K. S.; Guralnick, B. J.; Wang, W. K.; Fersht,A. R.; Itzhaki, L. S. J. Molec. Biol. 1999 , 285,1869.

[135] Burritt, J. B.; Bond, C. W.; Doss, K. W.; Jesaitis, A.J. Anal. Biochem. 1996 , 238, 1.

[117] Russo, A. A.; Tong, L.; Lee, J.-O. D.; Jeffrey, P.D.; Pavletich, N. P. Nature 1998 , 395, 237.


[136] Terrett, N. K.; Gardner, M.; Gordon, D. W.;Kobylecki, R. J.; Steele, J. Tetrahedron 1995 , 51,8135.

[143] Brooks, G.; La Thangue, N. Drug Design Today1999 , 4, 455.

[144] McArthur, J. G.; Patel, S. D.; Lampere, L.; Tran, A.;Finer, M.; Gyuris, J. Proc. Amer. Assoc. CancerRes. 1999 , 40, Abstract 4157.

[137] Colas, P.; Cohen, B.; Jessen, T.; Grishina, I.;McCoy, J.; Brent, R. Nature 1996 , 380, 548.

[138] Cohen, B. A.; Colas, P.; Brents, R. Proc. Natl. Acad.Sci. USA 1998 , 95, 14272.

[145] Yamasaki, L.; Jacks, T.; Bronson, R.; Goillot, E.Cell 1996 , 85, 537.

[139] Lees, J. A.; Weinberg, R. A. Proc. Natl. Acad. Sci.USA 1999 , 96, 4421.

[146] Field, S. J.; Tsai, F.-Y.; Kuo, F.; Zubiaga, A. M.;Kaelin, W. G.; Livingston, D. M.; Orkin, S. H.;Greenberg, M. E. Cell 1996 , 85, 549.

[140] Pines, J. Curr. Biol. 1996 , 6, 1399.[147] Johnson, D. G.; Schwarz, J. K.; Cress, W. D.;

Nevins, J. R. Nature 1993 , 365, 349.[141] Bourne, Y.; Watson, M. H.; Hickey, M. J.; Holmes,W.; Rocque, W.; Reed, S. I.; Tainer, J. A. Cell1996 , 84, 863. [148] Chen, Y. N.; Sharma, S. K.; Ramsey, T. M.; Jiang,

L.; Martin, M. S.; Baker, K.; Adams, P. D.; Bair, K.W.; Kaelin, W. G. J. Proc. Natl. Acad. Sci. USA1999 , 96, 4325.

[142] Prabhu, N. S.; Somasundaram, K.; Tian, H.; Enders,G. H.; Satyamoorthy, K.; Herlyn, M.; El-Deiry, W.S. Int. J. Oncol. 1999 , 15, 209.

Current Inhibitors of Cyclin-Dependent Kinases as Anti-Cancer Therapeutics … · 2015-02-27 ·...

Documents

Transcript of Current Inhibitors of Cyclin-Dependent Kinases as Anti-Cancer Therapeutics … · 2015-02-27 ·...