Structure-activity relationship studies of flavopiridol analogues

Article abstract of DOI:10.1016/S0960-894X(00)00156-6

Cyclin dependent kinases (CDKs) along with the complementary cyclins form key regulatory checkpoint controls on the cell cycle. Flavopiridol is a synthetic flavone that shows potent and selective cyclin-dependent kinase inhibitory activity. In this paper, we report modifications of the 3-hydroxy-1-methylpiperidinyl (D ring) of flavopiridol and their effect on CDK inhibitory activity. (C) 2000 Elsevier Science. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00156-6

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1037±1041
Structure±Activity Relationship Studies of Flavopiridol Analogues
Krishna K. Murthi, a,* Marja Dubay, b Christopher McClure, c Leonardo Brizuela, b
Michael D. Boisclair, c Peter J. Worland, d Muzammil M. Mansuri a and Kollol Pal a
aDepartment of Medicinal Chemistry, Mitotix Inc. One Kendall Square, Bldg. 600, Cambridge, MA 02139, USA
bDepartment of Biochemistry, Mitotix Inc. One Kendall Square, Bldg. 600, Cambridge, MA 02139, USA
cDepartment of Screening, Mitotix Inc. One Kendall Square, Bldg. 600, Cambridge, MA 02139, USA
dDepartment of Pharmacology, Mitotix Inc. One Kendall Square, Bldg. 600, Cambridge, MA 02139, USA
Received 7 July 1999; accepted 2 March 2000
AbstractÐCyclin dependent kinases (CDKs) along with the complementary cyclins form key regulatory checkpoint controls on the
cell cycle. Flavopiridol is a synthetic ¯avone that shows potent and selective cyclin-dependent kinase inhibitory activity. In this
paper, we report modi®cations of the 3-hydroxy-1-methylpiperidinyl (D ring) of ¯avopiridol and their e€ect on CDK inhibitory
activity. # 2000 Elsevier Science Ltd. All rights reserved.
The family of cyclin-dependent kinases (CDKs) are
important regulators that control the timing and coor-
dination of the progression of the cell cycle.1 CDKs
form reversible complexes with their obligate cyclin
partners to control transition through key junctures in
the cell cycle. For example the activated CDK4-cyclin
D1 complex controls progression through the G1 phase
while the CDK1-cyclin B1 complex controls entry into
the mitotic phase of the cell cycle.2 Endogenous cyclin
dependent kinase inhibitory proteins (CDKIs) are
known which bind either the CDK or cyclin component
and inhibit the kinase activity.3 In many tumors such as
melanomas, pancreatic and esophaegal cancers these
natural CDKIs are either absent or mutated.4 Thus
selective CDK inhibitors may prove to be e€ective che-
motherapeutic agents.
CDK inhibitory activity.8 Studies have shown that its
tumor cell growth inhibitory activity occurs in a cell
cycle speci®c manner.6a In addition, kinetic studies have
shown that ¯avopiridol binds at the ATP binding site of
the CDKs.6b The recently reported X-ray crystal struc-
ture of des-chloro ¯avopiridol, 1a, bound to CDK2 also
con®rms that the ATP binding pocket of the enzyme is
occupied by the ¯avone nucleus.9 The total synthesis of
¯avopiridol10 and some SAR around ¯avopiridol
have been reported.8 Two naturally occurring poly-
hydroxylated ¯avonoids quercetin and genistein (Fig. 1)
are potent tyrosine kinase inhibitors but have poor CDK
inhibitory activity.6b,11 Structurally both quercitin and
genistein lack the 3-hydroxy-1-methylpiperidinyl ring
(D-ring) present in ¯avopiridol. Therefore, we explored
the SAR around the D-ring of ¯avopiridol to determine
the key structural requirements for CDK inhibitory
activity.
Flavopiridol (1) (NSC 649890, L86-8275) [cis-5,7-dihy-
droxy-2-(2-chlorophenyl)-8-[4-(3-hydroxy-1-methyl)-
piperidinyl]-1-benzopyran-4-one] is a synthetic ¯avone
that has been shown to have antitumor activity against
various tumor cell lines such as human lung carcinoma,
breast carcinoma, and also inhibits tumor growth in
xenograft models.5 It has been shown to induce arrest in
both the G1 and G2 phases of the cell cycle.6 Flavopir-
idol, which is currently in clinical trials as an anticancer
therapeutic,7 is a potent and selective inhibitor of the
CDKs (Fig. 1), and its antitumor activity is related to its
Flavopiridol 1 and its trans isomer 2 were synthesized by
methods described previously.10 The D ring analogues
cis and trans 8-[3-(4-hydroxy-1-methyl)-piperidinyl], 4a
and 4b, and cis-8-[2-hydroxycyclohexyl], 5, were synthe-
sized by routes analogous to the synthesis of ¯avopiridol
(Scheme 1). Condensation of 1,3,5- trimethoxybenzene
with 1-methyl-3-piperidinone in acetic acid saturated
with hydrogen chloride provided ole®n 20. Hydrobora-
tion of ole®n 20 via the in situ generation of borane
(BF3-OEt2, NaBH4) followed by an oxidative work up
a€orded the trans-alcohol 21. Inversion of the alcohol
stereochemistry was accomplished in two steps: Swern
oxidation to the corresponding ketone followed by
*Corresponding author. Tel.: +1-617-225-001 ext. 278; fax: +1-617-
225-0005; e-mail: murthi@mitotix.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00156-6

1038
K. K. Murthi et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1037±1041
Figure 1.
Scheme 1. (a) (i) BF3-OEt, NaBH4, THF; (ii) conc HCl, then NaOH, H2O2; (b) (COCl)2, DMSO, Et3N, CH2Cl2; (c) NABH4, EtOH; (d) BF3-OEt,
Ac2O, CH2Cl2; (e) 20-Cl-benzoylchloride, pyridine; (f) (i) NaH, THF, (ii) HCl gas, (iii) Na2CO3; 9g) pyridinium HCl, 180 ꢀC.
reduction with NaBH4 in methanol to produced a mixture
of cis-alcohol 22 and the trans-alcohol 21 in a 1:7 ratio. The
alcohols 21 and 22 were separated by silica gel chromato-
graphy. Selective demethylation followed by a Fries rear-
rangement of cis-alcohol 22 gave hydroxyacetophenone
(23). Benzoylation of 23 with 2-chlorobenzoyl chloride
followed by cyclization (NaH/THF, HCl) provided the
5,7-dimethoxy¯avone (24). Finally, demethylation was
accomplished by heating 24 with pyridinium hydro-
chloride under melting conditions (180 ꢀC). The corre-
sponding trans-alcohol, 4b and the cyclohexyl analogue
5, were prepared in a similar manner. The only notable

K. K. Murthi et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1037±1041
1039
di€erence in the synthesis of the 6, is that the ole®n
derivative 27, was prepared by the reaction of 2-lithio-
1,3,5-trimethoxybenzene with cyclohexanone.
ole®n 32. Acylation of the aromatic nucleus was accom-
plished by reaction with acetic anhydride followed by
BF3-OEt promoted Fries rearrangement. Modi®cation
of the C-ring could easily be accomplished by acylation
of 33 with various acid chlorides followed by cyclization
(NaH/THF, HCl) to generate ¯avones 34±43. Finally,
demethylation with pyridinium hydrochloride provided
¯avones 6±15.
The syntheses of ketone 3, and analogues 16 and 17 are
shown in Scheme 2. Swern oxidation of the dimethoxy
¯avone 1b, followed by demethylation with pyridinium
hydrochloride provide ketone 3. The Stille cross-coupling
protocol was employed to introduce aromatic rings in
place of the piperidinyl moiety. Reaction of the appro-
priate aryl stannanes with 44 followed by demethylation
resulted in analogues 16 and 17. The synthesis of ole®n
analogues 6±15 is shown in Scheme 3. Condensation of
3,5-dimethoxy phenol with 1-methyl-4-piperidinone in
acetic acid saturated with hydrogen chloride provided
The synthesis of the quinolone and isocoumarin analo-
gues 18 and 19 is shown in Scheme 4. Condensation of
2-bromo-3,5-dimethoxy aniline with ethyl benzoylace-
tate gave ester 45.12 Thermal cyclization in diphenyl
ether, Stille cross coupling with vinyl stannane 48,13
followed by demethylation provided 18. Dimethoxy
Scheme 2. (a) (COCl)2, DMSO, Et3N, CH2Cl2; (b) pyridinium HCl, 180 ꢀC; (c) ArSnBu3, Pd(PPh3)2Cl2, DMF, 110 ꢀC.
Scheme 3. (a) N-methypiperidinone, CH3CO2H, HCl gas; (b) BF3-OEt, Ac2O, Ch2Cl2; (c) RCOCl, pyridine; (d) (i) NaH, THF, (ii) HCl gas, (iii)
Na2CO3; (e) pyridinium HCl, 180 ꢀC.

1040
K. K. Murthi et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1037±1041
Scheme 4. 9a) Ph2O, 230 ꢀC; (b) 48, Pd(PPh3)2Cl2, DMF 100 ꢀC; 9c) pyridinium HCl, 180 ꢀC; (d) Br2, CHCl3; (e) BBr3, DCE, re¯ux.
Table 1. CDK Inhibitory activity of ¯avopiridol D-ring analogues
isocoumarin 47 was synthesized as described pre-
viously.14 Bromination of 47, cross-coupling with 48
and demethylation gave analogue 19. The kinase assays
Compound
CDK4/
Cyc D1 (mM)
CDK1/
Cyc B (mM)
MFC-7
(mM)
were carried out as reported previously.15
1
2
3
4a
4b
5
6
7
0.2
56
8
120
250
31
0.8
2.4
0.55
1.8
0.65
2.5
1.0
1.2
0.8
7
0.2
49
10
NTa
NT
12
1.1
1.2
1.7
2.3
0.98
NTa
NTa
1.7
NTa
NTa
22
0.5
NTa
NTa
NTa
NTa
NTa
0.75
1
The biological results for compounds 1±19 are shown in
Table 1. In ¯avopiridol, the hydroxyl group and the
¯avone ring substituents on the piperidine ring (D ring)
have a cis-orientation. Inverting the hydroxyl sub-
stituent to give the trans-isomer 2 results in >1000-fold
loss in activity against both CDK4/cyclin D and CDK1/
cyclinB kinases.
8
9
3
1
1
NTa
NTa
1.5
NTa
NTa
NTa
NTa
NTa
NTa
10
11
12
13
14
15
16
17
18
19
Similarly, oxidation of the alcohol to the ketone analo-
gue 3 also results in a signi®cant loss of activity. The
importance of the piperidinyl nitrogen was explored
with the cyclohexyl D ring analogue 5. The cyclohexyl
D ring analogue 5 is more than an order of magnitude
less active than ¯avopiridol 1. Thus, the presence of the
nitrogen atom on the D ring is very important for CDK
inhibitory activity. Next we studied the position of the
nitrogen atom on the D ring with the 4-hydroxy-1-
methyl-piperidinyl analogues 4a and 4b. The cis- and
trans- hydroxyl isomers 4a and 4b which are derived
from 1-methyl-3-piperidinone are completely inactive
against CDK4/cyclin D. Thus the positioning of the
nitrogen atom on the D ring is also very important for
CDK inhibitory activity.
21
160
44
85
NTa
NTa
23
aNot tested.
with the ole®n 6 suggests that the hydroxyl group-Lys
33 interaction may not be very critical for CDK inhibi-
tory activity as long as the nitrogen-Asp 145 interaction
is maintained. The major loss in activity of trans ¯avo-
piridol 2 and ketone 3 could be due to a loss in one or
both hydrogen bonding interactions because of con-
formational changes of the hydroxy piperidine ring.
Replacement of the 3-hydroxy-1-methyl-piperidinyl ring
with aryl rings such as pyridyl (compound 16) and pyr-
imidyl (compound 17) also results in a major loss of
CDK inhibitory activity.
The ole®n analogue 6, which lacks the hydroxyl group
results only in a 4- to 5-fold loss in activity against both
the CDKs. The X-ray crystal structure9 of des-chloro
¯avopiridol (1b) with CDK2 shows that the hydroxy
piperidine D ring partially occupies the phosphate
binding region. The structure shows interactions
between the the piperidine nitrogen and Asp 145 and
the hydroxyl group and Lys 33 respectively. Our SAR
studies indicate that the piperidine nitrogen-Asp 145
interaction seems to be very critical for CDK inhibitory
activity. Presumably the lack of activity of compounds
4a, 4b, and 5 is due to the loss of the piperidine nitro-
gen-Asp 145 interaction. The modest loss in activity
Since the tetrahydropyridyl analogue (6), did not result
in a major loss of CDK inhibitory activity we pursued
additional SAR studies of this series involving changes
in the ¯avone nucleus and the C-ring. As shown in
Table 1, the SAR for the tetrahydropyridine series
diverges from the data that has been reported for ¯avo-
piridol. Modi®cation of the C-ring generally leads to a

K. K. Murthi et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1037±1041
1041
loss of CDK inhibitory activity in ¯avopiridol.8
Removal of the 220-chlorine atom from ¯avopiridol, 1a,
results in about a 10-fold loss in activity; however, a
similar modi®cation has little e€ect in the ole®n series
(compare 6 and 12). Other variations of the halogen
substituent (7±11,13) on the aromatic ring do not seem to
impair the CDK inhibitory activity. Heterocyclic mod-
i®cations such as pyridyl, 14, also seem to be well toler-
ated. It is noteworthy that the cyclohexyl analogue, 15,
results in a loss of activity. Modi®cation of the ¯avone
nucleus was more consistent with the ¯avopiridol SAR.
The C-5 and C-7 hydroxyl groups are critical for kinase
inhibitory activity. Thus, the corresponding dimethoxy
analogues, 34±43, are devoid of CDK inhibitory activ-
ity. Our results are consistent with the X-ray crystal
structure of 1a with CDK2 which indicates that the C-5
hydroxyl and the ¯avone carbonyl are involved in cri-
tical hydrogen bonds with E81 and L83 in the adenine
binding pocket.9 In addition, the C-ring is solvent
exposed and does not make appreciable protein inter-
actions. It is noteworthy, that in the present study
modication of the D-ring seems to attenuate these
binding interactions to some degree. While in the ¯avo-
piridol series, modi®cation of the C-ring results in at
least a 10-fold loss in activity, the tetrahydropyridine
series seems quite tolerant of changes in the C-ring (7±
15). On the other hand, the interactions with E81 and
L83 seem to be retained, since methylation abrogates
CDK inhibitory activity. The ole®n analogues 6±10 and
13 were tested against the MCF-7 tumor cell line to
measure growth inhibition and showed comparable
activities for both CDK inhibition and inhibition of cell
growth (Table 1). Finally alteration of the ¯avone
nucleus to either a quinol-4-one, 18, or an isocoumarin
nucleus, 19, resulted in reduced activity. This may
re¯ect ine€ective interactions in the adenine binding
pocket.
Draetta, G. F. Trends Biochem. Sci. 1990, 15, 378. (e) Sherr, C.
J. Cell 1993, 73, 1059.
2. Draetta, G. F. Curr. Opin. Cell Biol. 1995, 6, 842.
3. (a) Harper, J. W. In Checkpoint Controls and Cancer;
Kastan, M. B. Ed.; ICRF, 1997, pp 91±107. (b) Serrano M.
Hannon, G. J. Beach, D. Nature 1993, 366, 704.
4. Serrano, M. Exp Cell Res. 1997, 237, 7.
5. (a) Czech, J.; Ho€mann, D.; Naik, R.; Sedlacek, H. H. Int.
J. Oncol. 1995, 6, 31. (b) Sedlacek, H. H., Ho€mann, D.;
Czech, J.; Kolar, C.; Seeman, G.; Gussow, D.; Bosslet, K.
Chimia 1991, 45, 311.
6. (a) Kaur, G.; Stetler-Stevenson, M.; Sebers, S.; Worland, P.;
Sedlacek, H.; Myers, C.; Czech, J.; Naik, R.; Sausville, E. J.
Natl. Cancer Inst. 1992, 84, 1736. (b) Losiewiecz, M. D.;
Carlson, B. A.; Kaur, G.; Worland, P. J. Biochem. Biophy.
Res. Commun. 1994, 201, 589. (c) Carlson, B. A.; Dubay, M.
M.; Sausville, E. A.; Brizuela, L.; Worland, P. J. Cancer Res.
1996, 56, 2973.
7. (a) 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. (b) Wright, J.; Blatner, G. L.; Cheson,
B.D., Oncology 1998, 12, 1018. (c) Wright, J.; Blatner, G. L.;
Cheson, B. D. Oncology, 1998, 12, 1023.
8. Sedlacek, H.; Czech, J.; Naik, R.; Kaur, G.; Worland, P. J.;
Losiewiecz, M. D.; Parker, B.; Carlson, B. A.; Smith, A.;
Senderowicz, A.; Sausville, E. A. Intl. J. Oncol. 1996, 9, 1143.
9. Azevedo, W. F.; Mueller-Dieckmann, H. J.; Schulze-Gah-
men, U.; Worland, P. J.; Sausville, E. A.; Kim, S. H. Proc.
Natl. Acad. Sci. USA 1996, 93, 2735.
10. (a) Kattige, S. L.; Naik, R. G.; Lakdawalla, A. D.;
Dohadwalla, A. N.; Rupp, R. H.; Desouza, N. J. U.S. Patent
4 900 727, 1990. (b) Naik, R.; Lal, B.; Rupp, R. H.; Sedlacek,
H. H.; Dickneite, G.; Czech, J. U.S. Patent 5 284 856, 1988. (c)
Naik, R.; Lal, B.; Rupp, R. H.; Sedlacek, H. H.; Dickneite,
G.; Czech, J. Euro. Patent 0 366 061 1988. (d) Naik, R.; Kat-
tige, S. I.; Bhat, S. V.; Alrejy, B.; DeSouza, N. J.; Rupp, R. H.
Tetrahedron 1988, 44, 2081.
11. Chang, C. J.; Geahlen, R. L. J. Natl. Prod. 1992, 55, 1529.
12. Kuo, S. C.; Lee, H. Z.; Juang, J. P.; Lin, Y. T.; Wu, T. S.;
Chang, J. J.; Lednicer, D.; Paul, K. D.; Lin, C. M.; Hamel, E.;
Lee, K. H. J. Med. Chem. 1993, 36, 1146.
Our SAR studies on ¯avopiridol have shown that both
the presence and the position of the nitrogen moiety on
the D ring are critical requirements for CDK inhibitory
activity. We have also identi®ed a simpler analogue,
ole®n 6, which shows good CDK inhibitory activity.
Our SAR studies on ole®n 6 have shown that the C ring
tolerates substitutions without a major loss in activity.
We also found that the quinolone and the isocoumarin
rings are not e€ective ¯avone replacements.
13. Vinyl stannane 48 was prepared as shown below.
14. Ohta, S.; Kamata, Y.; Inagaki, T.; Masuda, Y.; Yama-
moto, S.; Yamashita, M.; Kawasaki, I. Chem. Pharm. Bull
1993, 41, 1188.
References and Notes
15. Wick, S. T.; Dubay, M. M.; Imanil, I.; Brizuela, L. Onco-
gene 1995, 11, 2013.
1. Sherr, C. J. Cell 1994, 79, 551. (b) Hunter, T.; Pines, J. Cell
1994, 79, 573. (c) Morgan, D. O. Nature 1995, 374, 131. (d)