Identification of selective inhibitors of cyclin dependent kinase 4

Article abstract of DOI:10.1016/S0960-894X(01)00416-4

A new structural type of kinase inhibitor, containing a benzocarbazole nucleus, has been identified. Members of the series are selective for inhibition of the cyclin dependent kinase family of enzymes. Although the cdks are highly homologous, representatives of the series showed intra-cdk selectivities, especially for cdk4. SAR studies elucidated the important features of the molecules for inhibition.

Full text of DOI:10.1016/S0960-894X(01)00416-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2209–2211
Identification of Selective Inhibitors of Cyclin Dependent
Kinase 4
David J. Carini,a,* Robert F. Kaltenbach, III,a Jie Liu,a Pamela A. Benfield,a
John Boylan,a Michael Boisclair,b Leonardo Brizuela,b Catherine R. Burton,a Sarah Cox,a
Robert Grafstrom,a Barbara A. Harrison,a Kimberly Harrison,a Emeka Akamike,a
Jay A. Markwalder,a Yuki Nakano,b Steven P. Seitz,a Diane M. Sharp,a
George L. Trainora and Thais M. Sieleckia,*
aDuPont Pharmaceuticals Company, Wilmington, DE 19880, USA
bMitotix Inc., Cambridge, MA, USA
Received 19 April 2001; accepted 7 June 2001
Abstract—A new structural type of kinase inhibitor, containing a benzocarbazole nucleus, has been identified. Members of the
series are selective for inhibition of the cyclin dependent kinase family of enzymes. Although the cdks are highly homologous,
representatives of the series showed intra-cdk selectivities, especially for cdk4. SAR studies elucidated the important features of the
molecules for inhibition. # 2001 DuPont Pharmaceuticals Company. Published by Elsevier Science Ltd. All rights reserved.
The eucaryotic cell cycle is regulated by the coordinated
activity of a family of cyclin dependent kinases (cdks).1
The enzymes, consisting of a kinase and a regulatory
cyclin subunit, become active at specific phases of the
cycle. For example, cdk1/cyclin B becomes active late in
the G2 phase of the cell cycle, while cdk2/cyclin E is
important late in the G1 phase. The complexes of cdk4
and cdk6 with D-type cyclins are especially interesting
as they control cell cycle entry. Activation of cdk4
results in a specific phosphorylation of the retino-
blastoma protein (pRb) which is critical to the release of
transcription factors of the E2F family.2 The Rb kinase
pathway is frequently deregulated in transformed cells.3
For example, pRb and p16 (a specific inhibitor of cdk4/
6) have been identified as the products of tumor sup-
pressor genes while cyclin D1 is a proto-oncogene pro-
duct. These factors make specific inhibitors of cdk4 highly
interesting targets for pharmacological intervention in
cancer.4
cdk4/D1 over other related cdks prompted us to initiate
a chemistry effort to explore the activity of this class of
molecules.
The literature route to 1 consists of the reaction of 2,7-
dihydroxynaphthalene with 2-hydrazinobiphenyl in the
presence of sodium bisulfite. We found that the slightly
longer route consisting of a Fischer indole synthesis
followed by aromatization of the B ring (Scheme 1)
resulted in higher yields and was more convenient for
analoging purposes.5
A high-throughput screening campaign using cdk4/
cyclin D1 complex and a GST-tagged, truncated pRb
succeeded in identifying the carbazole 1 as an inhibitor.
Kinetic analysis suggested that 1 is a competitive inhi-
bitor with respect to ATP. The apparent selectivity for
Several findings emerge from the data in Table 1. The
presence of a hydrogen bond donor in the A ring
appears to be important. Removal (example 9) or
methylation (example 4) results in analogues devoid of
activity. Movement of the hydroxy moiety to the 1- or
*Corresponding author. Tel.: +1-302-695-8501; fax: +1-302-695-
1502; e-mail: thais.m.sielecki@dupontpharma.com
0960-894X/01/$ - see front matter # 2001 DuPont Pharmaceuticals Company. Published by Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00416-4

2210
D. J. Carini et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2209–2211
the 3-position on the A ring results in less potent inhi-
bitors, although example 7 in which the hydroxyl group
is at the 4-position is surprisingly active in this regard.
Similar investigations indicate that the phenyl group on
the D ring contributes less to the overall binding affinity
than the phenol and is optimally disposed when
attached to the 7-position.
donor is critical and the lack of activity of 4 is likely not
the result of an unfavorable steric interaction.
The SAR of the D ring phenyl may be conveniently
explored via Suzuki couplings on the bromide 20
(Scheme 2). As the data in Table 3 indicate, substitution
is well tolerated on the phenyl ring. Significantly,
example 22 is 10-fold more soluble than 1. Example 22
also retains the selectivity within the CDK family dis-
played by 1 (cdk1/B IC50=74.5 mM). In addition, IC50’s
for inhibition of other kinases including c-abl, PKA,
and PKC are at least two orders of magnitude higher
than that observed against cdk4.
To further define the nature of the A ring interaction,
we prepared the aza analogues and their corresponding
N-oxides. The route uses a stilbene photocyclization to
assemble the ring system.6 As shown in Table 2, these
analogues are all less potent than the reference com-
pound 13. The fact that the N-oxides cannot substitute
for the phenol strongly suggests that a hydrogen bond
The properties of four of the analogues in in vitro cell
assays are shown in Table 3. Exposure of colon carci-
noma HCT116 to these agents causes cell growth inhi-
bition accompanied by cell death with IC50’s in the
10 mMrange as judged by an SRB assay. The Skut1A/
1B pair are Rbꢀ and Rb+, respectively. The prediction
is that a fully selective cdk4 inhibitor would effect only
the Rb+ cells since Rb is the only known substrate of
Scheme 1. Synthesis of positional isomers.
Table 1. Inhibition of Cdk4/D1
Scheme 2. Synthesis of substituted phenyl analogues.
Table 3. Cellular activity of selected cdk4/D1 inhibitors
Compd
R1
R2
Cdk4/D1
IC50 (mM)
1
4
5
6
7
8
9
10
11
12
13
2-OH
2-MeO
1-OH
3-OH
4-OH
2-NH2
H
2-OH
2-OH
2-OH
2-OH
7-Ph
7-Ph
7-Ph
7-Ph
7-Ph
7-Ph
7-Ph
8-Ph
9-Ph
10-Ph
H
3.6
>309
19.5
29
Compd
R3
Cdk4/D1
HCT116
Skut 1A
Skut 1B
IC50 (mM) IC50 (mM) IC50 (mM) IC50 (mM)
1
H
3.6
2.3
1.9
9.9
10.2
9.6
2.2
51
10.7
11.6
5.6
2.8
4.1
2.3
10.2
5.8
22
23
24
30-NH2
40-OH
40-Cl
27.3
>341
22
36.0
37.5
83
17.4
Table 2. Inhibition of Cdk4/D1 by aza analogues
Compd
A
B
C
D
Cdk4/D1
IC50 (mM)
14
15
16
17
18
19
N
CH
N
CH
CH
NO
CH
CH
CH
N
CH
CH
NO
CH
CH
CH
N
CH
CH
442
159
30
197
605
59
Figure 1. FACS analysis of compound 1 on HS27 fibroblasts. Primary
human fibroblasts (HS27) were arrested using serum starvation and
released with 10% serum in the presence or absence of compound 1.
Profile 1 (blue) serum starved cells. Profile 2 (green): 24 h after the
addition of 10% serum to serum starved cells. Profile 3 (red): 24 h after
the addition of 10% serum and compound 1 to serum starved cells.
CH
CH
CH
CH
CH

D. J. Carini et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2209–2211
2211
factors that are downstream of cdk4 are present in
diminished quantities. Thus, cdk2, cyclin B, cyclin E,
and E2F-1 are expressed at levels similar to serum
starved cells.
In conclusion, we have found a series of carbazole ana-
logues with interesting intra-cdk selectivity profiles.
These activities translate into a consistent profile in
whole cell assays. The dim relationship of these com-
pounds to known classes of kinase inhibitors suggests
that elucidation of the structural basis of cdk4 inhibi-
tion will provide useful clues for the design of more
potent and selective analogues.
References and Notes
1. Morgan, D. O. Annu. Rev. Cell Dev. Biol. 1997, 13, 261.
2. Dyson, S. Genes Dev. 1998, 12, 2245.
3. Pines, J. Sem. Cell. Biol. 1994, 5, 399.
4. Sielecki, T. M.; Boylan, J. F.; Benfield, P. A.; Trainor, G. L.
J. Med. Chem. 2000, 43, 1.
Figure 2. Western-blot analysis of fibroblast G1 release. Equal protein
from cellular extracts from the G1 release experiment described in
Figure 1 were tested for the expression of selected cell cycle related
proteins using Western-blot analysis. Lane 1: Serum starved cells.
Lane 2: 24 h after the addition of 10% serum to serum starved cells.
Lane 3: 24 h after the addition of 10% serum and compound 1 to
serum starved cells.
5. Katritzky, A. R.; Wang, Z. J. Heterocycl. Chem. 1988, 25, 671.
6. DeSilva, O.; Snieckus, V. Synthesis 1971, 5, 254.
7. Cell cycle analysis. Primary human fibroblast cells were
obtained from ATCC (HS27) and maintained in MEM media
supplemented with glutamine and 10% heat-inactivated FCS.
Cells were G1 arrested using serum starvation and released
with the addition of 10% FCS. Cells were harvested 24 h fol-
lowing serum release. Cell cycle analysis was performed using
the standard Propidium Iodine (PI) assay (in Current Proto-
cols in Cytometry; Robinson, J. P., Ed.; John Wiley & Sons:
New York, 1998; Vol. 1).
the enzyme. Compounds 1 and 22 distinguish between
the Skut 1A and Skut 1B cells, consistent with their
enzymatic selectivity.
The behavior of 1 in cells was examined in more detail.
Figure 1 shows a FACS analysis of serum starved cells
arrested in G0/G1 and treated with 1.7 The compound
prevents serum induced cell cycle re-entry, consistent
with a cdk4 blockade and failure to progress to S phase.
BrDU incorporation analysis (data not shown) also
demonstrated a loss of S phase cells.
8. Protein extracts were tested for the expression of selected
cell cycle related proteins using Western-blot analysis. For
Western blots, cellular extracts were diluted in sample buffer
(2% SDS, 10% glycerol, 10 mm b-mercaptoethanol, and
60 mMTris–HCl, pH 6.8) and heated for 5 min at 95 ꢁC.
Samples were stored at ꢀ70 ꢁC. Proteins were separated by
SDS-PAGE, transferred to nitrocellulose, and probed with the
indicated primary antibody using standard techniques. The
level of protein expression was visualized using the ABC Vec-
tastain kit (Vector Labs, Burlingame, CA) and Chemilumine-
sence (NEN, Boston, MA). Antibodies were obtained from the
following sources: p16 (Calbiochem, San Diego, CA), cyclin E,
cdk2, cdk4, E2F-1, cyclin B, cyclin D1 (Santa Cruz, CA), pRb
(Pharmingen, San Diego, CA), and actin (Amersham, Piscat-
away, NJ).
Western-blot analysis of fibroblasts treated with 1 is
consistent with a cdk4 inhibition mechanism.8 There is a
clear increase in dephosphorylated pRb without detect-
able alteration in cyclin D1 or cdk4 levels. The endo-
genous protein inhibitor p16 is also unaffected by
exposure to 1 (Figure 2). Enzymes and transcription