3-Amino-1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazoles: A new class of CDK2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.10.071

We have recently reported about a new class of Aurora-A inhibitors based on a bicyclic tetrahydropyrrolo[3,4-c]pyrazole scaffold. Here we describe the synthesis and early expansion of CDK2/cyclin A-E inhibitors belonging to the same chemical class. Synthe

Full text of DOI:10.1016/j.bmcl.2005.10.071

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1084–1090
3-Amino-1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazoles:
A new class of CDK2 inhibitors
Paolo Pevarello,a,* Daniele Fancelli,a Anna Vulpetti,a Raffaella Amici,a Manuela Villa,a
a,
Valeria Pittala, Paola Vianello,a Alexander Cameron,a Marina Ciomei,b Ciro Mercurio,b
`
James R. Bischoff,b,à Fulvia Roletto,b Mario Varasia and Maria Gabriella Brascaa
aDepartment of Chemistry, Nerviano Medical Science, BU-Oncology, Via Pasteur 10, 20014 Nerviano MI, Italy
bDepartment of Biology, Nerviano Medical Science, BU-Oncology, Via Pasteur 10, 20014 Nerviano MI, Italy
Received 1 September 2005; revised 12 October 2005; accepted 20 October 2005
Available online 14 November 2005
Abstract—We have recently reported about a new class of Aurora-A inhibitors based on a bicyclic tetrahydropyrrolo[3,4-c]pyrazole
scaffold. Here we describe the synthesis and early expansion of CDK2/cyclin A–E inhibitors belonging to the same chemical class.
Synthesis of the compounds was accomplished using a solution-phase protocol amenable to rapid parallel expansion. Compounds
with nanomolar activity in the biochemical assay and able to efficiently inhibit CDK2-mediated tumor cell proliferation have been
obtained.
Ó 2005 Elsevier Ltd. All rights reserved.
Uncontrolled cell growth and proliferation are hall-
marks of all cancers and are directly linked to cell cycle
dysregulation.1 Cyclin-dependent kinase 2 (CDK2) in
complex with cyclins E and/or A is a key cell cycle reg-
ulator and continues to be an attractive target for the
discovery of new antitumor agents.2 In particular, inhib-
itors of CDK2/cyclin A/E have already progressed into
clinical trials with encouraging early results.3 Recent
findings in the biology of the CDKs show that CDK2
may be dispensable for tumor formation and mainte-
nance4 but compounds able to inhibit both CDK2 and
CDK1 may still be regarded as potential anti-tumor
agents.5 We recently reported on a series of 3-aminopy-
razoles as potent, orally available inhibitors of CDK2/
cyclin A/E and CDK1/cyclin B, displaying activity in vi-
vo in a murine tumor xenograft model.6 A limitation
associated with the 3-aminopyrazole scaffold is that
high-throughput synthesis can be applied easily only
for placing substituents at position 3, that is, toward
the solvent-accessible region when the ligand is placed
in the ATP-pocket of the kinase.7 Positions 4- and 5-,
which could be exploited to gain accessibility to the bur-
ied and phosphate binding regions to direct substituents
are not synthetically amenable to rapid parallel medici-
nal chemistry.
We reasoned that by condensing a second heterocyclic
ring onto the 3-aminopyrazole moiety we could gain ac-
cess to the phosphate binding region of the ATP-kinase
pocket. Thus, we would obtain a new chemical class that
could be used to exploit an additional pocket in the ATP
binding site (Fig. 1). The generation of this new chemi-
cal class, 1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazoles, was
recently reported by us in a paper dealing with the char-
acterization of potent Aurora-A inhibitors.8 Here, we
describe the solution-phase synthesis and early expan-
sion of this class toward CDK2/cyclin A inhibitors. A
general solution-phase synthetic methodology was need-
ed to allow the generation of many derivatives using
parallel medicinal chemistry methodologies. The synthe-
sis of the scaffold 1,4,5,6-tetrahydropyrrolo[3,4-c]pyra-
zole was accomplished as previously described.8
Abbreviations: CDK, cyclin-dependent kinase; ATP, adenosine tri-
phosphate; FACS, fluorescence activated cell sorting; BrdU, bro-
modesoxyuridine; WB, Western blots; pRb, retinoblastoma protein;
ADME, absorption-disposition-metabolism-excretion.
Keywords: CDK2; Cyclins; Kinase selectivity; Tumor cell proliferation
inhibition.
*
Corresponding author. Tel.: +390331581927;
+390331581347; e-mail: paolo.pevarello@nervianoms.com
`
fax:
Present address: Istituto di Chimica Farmaceutica, Universita di
Catania, Via Doria 9, 95125 Catania, Italy.
Present address: Centro Nacional de Investigaciones Oncologicas
à
Scheme 1 reports a general solution-phase synthesis for
this class of compounds that was applied to obtain
´
(CNIO), Melchor Ferna´ndez Almagro n°3, 28029 Madrid, Spain.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.071

P. Pevarello et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1084–1090
1085
Figure 1. Schematic representation of the 1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazole scaffold in the kinase binding protein.
EtOOC
EtOOC
EtOOC
N
N
N
N
NH
NH
a)
c)
b)
N
N boc
N
N boc
N
N boc
HN
H2N
HN
O
H2 N
O
R
R
EtOOC
O
N
O
NH
N
N
N
N
e)
d)
R1
R1
HN
HN
O
O
R
R
Scheme 1. Reagents and conditions: (a) EtOCOCl, DIEA, THF, 2 h, 0–5 °C; (b) RCOCl (1.2 equiv), DIEA (7.2 equiv), THF, 16 h, 22 °C; (c) TFA/
DCM (1:1), 3 h, 22 °C; (d) R1NCO or R1COOH, TBTU, NMM, DCM or DMF, 24 h, 22 °C; (e) aq NaOH, MeOH, 72 h, 40 °C, then 35% HCl.
derivatives 1–18. The 1,4,5,6-tetrahydropyrrolo[3,4-c]-
pyrazole scaffold was selectively protected on the pyra-
zole ring using ethyl chloroformate. Acylation of the
3-position was accomplished using the corresponding
acyl chlorides and di-isopropyl-ethylamine as a base.
Removal of the N-Boc protection and treatment with
a suitable acyl chloride, followed by a final deprotection
of the ethylcarbamate group gave compounds 1–18.9
Aurora-A, a mitotic serine-threonine kinase, since, at
least in this compound series, inhibitors of CDK2/cyclin
A often inhibited Aurora-A. Our primary goal at this
stage was to dissociate CDK2/cyclin A from Aurora-A
activity. While in principle a double specificity inhibitor
of both serine-threonine kinases may be a superior tu-
mor cell proliferation blocker, we wanted to get selective
CDK2/cyclin A and selective Aurora-A inhibitors to
possibly decoupling additive toxic effects and make an
assessment of the advantages (or disadvantages) of a
selective versus double specificity inhibitor.
Our strategy for class expansion was based on three
main steps:
(a) Synthesize a limited number of inhibitors with mod-
erate to good activity toward CDK2/cyclin A;
(b) Obtain X-ray structures of such inhibitors with
CDK2/cyclin A;
A first series of 5-acetamido-3-benzamido-1,4,5,6-tetra-
hydro[3,4-c]pyrazoles was synthesized according to
Scheme 1 and inhibitory activities are reported in Table 1.
(c) Use these structures to decide how to further expand
this class.
All compounds 1–7 displayed moderate activity against
both CDK2/cyclin A and Aurora-A. CDK2/cyclin A
inhibition was less influenced by the nature of the substi-
tution on the benzamido aromatic ring than was Auro-
ra-A inhibition. The structure of CDK2/cyclin A was
In addition to testing compounds as CDK2/cyclin A
inhibitor, all compounds were tested in parallel against

1086
P. Pevarello et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1084–1090
Table 1. CDK2/cyclin A and Aurora-A inhibition of representative
5-acetyl-3-benzoylamino-1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazoles
Amides or ureas built on the tetrahydropyrrole nitrogen
(N-5) showed improved activity against the two refer-
ence kinases. In particular, ureas yielded activities in
the high nanomolar range. On the other hand, the com-
pounds did not show any selectivity for CDK2/cyclin A
versus Aurora-A. (Table 2). Within the kinase super-
family the largest sequence variation is seen in the
residues making up the solvent accessible region (see
Fig. 1). In this region, Aurora-A differs from CDK2
due to the presence of an inserted glycine in Aurora-A
(Gly216) that is absent in CDK2. This additional resi-
due in Aurora-A causes a different protein conformation
between the hinge region and the beginning of the C-ter-
minal domain that tends to disfavor inhibitors that
do not protrude out of the ATP pocket in a planar
fashion. This region can be exploited both to improve
selectivity as well as to modulate the ADME properties
of the ligands (Fig. 3).
H
N
N
N
HN
O
O
R
\
Compound
R
IC50 (lM)6,13
CDK2/cyclin A Aurora-A
1
2
3
4
5
6
7
H
1.15
1.08
2.54
1.37
1.59
2.11
3.11
2.14
4-Me
4-F
>10
4-OMe
4-CN
3-F
0.66
8.5
4.85
3,4-Methylenedioxy 2.76
>10
Based on this observation the hypothesis was made that
3-arylacetamido residues on the scaffold would generally
be much more potent against CDK2/cyclin A, whilst
3-benzamido compounds would preferentially inhibit
Aurora-A. 3-Benzamide containing compounds would
make positive lipophilic interactions with Leu139 and
Gly216 in Aurora-A, while 3-arylacetamido products
could fit nicely into the lipophilic space delimited by
Phe 82 and Ile10 in CDK2/cyclin A.
K33
F80
E81
E51
Indeed, when this substituent is varied the outcome is
striking (Table 3). Compounds 11–14 are inhibitors of
CDK2/cyclin A (down to nanomolar concentrations in
the case of 13 and 14) but are, at best, very weak inhib-
itors of Aurora-A.
L83
The same observation can be made for the substituents
pointing toward the solvent-accessible region of other
classes of CDK2/cyclin A inhibitors and leads us to be-
D86
Figure 2. The crystal structure of CDK2/cyclin A in complex with
compound 1 (PDB code 2C4G).
Table 2. CDK2/cyclin A and Aurora-A inhibition of representative
3-benzoylamino-5-substituted-1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazoles
H
N
solved in complex with compound 1 according to known
protocols.6 Figure. 2 shows its mode of binding in the
ATP-pocket of the kinase.
N
R1
N
HN
O
O
The 3-aminopyrazole moiety forms three hydrogen
bonds with Glu81 and Leu83 at the hinge region of
the CDK2 ATP binding site. The benzamido moiety
occupies the solvent accessible region, and the tetrahy-
dropyrrole moiety the hydrophobic pocket formed by
Ala31, Val64, Phe80, and Ala144 (also known as buried
region). In addition, the carbonyl oxygen of the amide at
position 5 forms a hydrogen bond with Lys33. The hit
compound was relatively simple with a low molecular
weight and seemed suitable for further decoration to
give improved activity, selectivity, and drug-likeness.
R
Compound
R
R1
IC50 (lM)6,13
CDK2/ Aurora-A
cyclin A
N
8
9
4-Me
1.05
0.43
0.12
0.44
0.16
0.39
3,4-Methylenedioxy
3,4-Methylenedioxy
v
Libraries of tetrahydropyrrole[3,4-c]pyrazoles were
made according to the method of Scheme 1 and three
representative compounds are reported in Table 2.
10
N
H

P. Pevarello et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1084–1090
1087
3-benzoic
3-arylacetic
(Aurora-A oriented)
(CDK2 oriented)
L139
I10
F82
G216
Flat shape
Chair-like shape
Figure 3. Compound 1 (flat-shape) and 11 (chair-like shape) docked in, respectively, Aurora-A and CDK2 crystal structures.
Table 3. CDK2/cyclin A and Aurora-A inhibition of representative
3-phenylacetylamino-5-substituted-1,4,5,6-tetrahydropyrrolo[4,3-c]pyr-
azoles
tumor cell lines, the selectivity profile on a panel of
CDKs and two other cross-reacting protein kinases,
GSK-3b and p42MAPK, and the preliminary in vitro
parameters for predicting in vivo ADME.
H
N
N
R1
Both compounds efficiently blocked at submicromolar
concentrations the proliferation of an ovarian (A2780)
and two colon (HCT116 and HT-29) tumor cell lines.
The two compounds also displayed moderate buffer sol-
ubility, were sufficiently stable to the important human
cytochrome CYP4503A4, showed moderate permeabili-
ty in a Caco-2 permeability assay, and were seen to bind
to plasma protein at high levels. When 15 and 18 were
tested on a panel of CDKs, they were seen to inhibit
CDK2/cyclin A and cyclin E at comparable levels, and
CDK1/cyclin B and CDK5/p25 in the low micromolar
range. On the other hand, CDK4/cyclin D1 was not sig-
nificantly inhibited. Both compounds 15 and 18 did not
significantly inhibit (IC50 > 10 lM) other protein ki-
nases in a standard panel of 22 enzymes. Compound
18 was equipotent on CDK2/cyclin A and GSK-3b. It
is frequently observed that CDK2 inhibition by small
molecules is accompanied by inhibition of the structur-
ally similar GSK-3b.10 Whether this double inhibition
may be advantageous (due to the recently reported acti-
vation of p53-dependent apoptosis by acute ablation of
GSK-3b in colorectal cancer cells)11 or detrimental due
to the known link between proliferation and GSK-3b
inhibition still remains to be established. However, that
most of the cellular activity of these compounds is med-
iated through CDK2/cyclin A/E inhibition is further
supported by an investigation into the mechanism of ac-
tion as reported in Figure 4.
N
HN
O
O
R
Compound
R
R1
IC50 (lM)6,13
CDK2/cyclin A
Aurora-A
11
12
13
14
H
CH3
0.44
0.73
2.31
>10
>10
>10
H
NHPh
NH2
NH-n-Bu
H
2-F
0.030
0.051
lieve that this simple rule is applicable across different
chemical classes and across different kinases once the
an inserted glycine is present among the aminoacids lin-
ing the solvent accessible region. We continued the
expansion of this chemical class and we obtained several
ureas (three of them, 15, 16, and 18 are reported in Ta-
ble 4) with an interesting antiproliferative activity in
cells (together with nanomolar inhibition of the target
CDK2/cyclin A).
Interestingly, within this subclass of compounds there
seems to be a correlation between lack of cellular activ-
ity and low ClogP and Caco-2 cell permeability (entry
17; Table 4).
FACS analysis was performed on colon HT-29 tumor
cells released from a nocodazole block. Untreated cells
are not blocked in the G1/S phase of their cell cycle
and the level of incorporation of BrdU is 62%. Con-
versely cells treated with the CDK inhibitor flavopiridol,
Based on the biochemical assay and cellular potency,
compounds 15 and 18 were selected for further assess-
ment. Table 5 reports their activity against additional

1088
P. Pevarello et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1084–1090
Table 4. CDK2/cyclin A, Aurora-A, A2780 tumor cell inhibition, C log P and Caco-2 permeability of representative 3-phenylacetylamino-5-ureidyl-
1,4,5,6-tetrahydropyrrolo[3,4-c]pyrazoles
H
N
N
HN
H
N
N
O
O
R1
R
Compound
R
R1
IC50 (lM)6,13
ClogP (ACD Labs)
Caco-2 perm6
CDK2/cyclin A
0.033
Aurora-A
A2780 cells
0.94
N
15
16
17
H
H
H
4.2
1.4
2.4
0.4
Moderate
Moderate
Low
0.045
0.051
>10
>10
0.54
8.27
N
O
18
Me (S-confg)
0.036
>10
0.13
2.8
Moderate
Table 5. Kinase selectivity, tumor cell proliferation inhibition, and
preliminary in vitro ADME parameters of compounds 15 and 18
used as a control, or compounds 15 or 18 showed a clear
cell cycle block in G1/S and did not incorporate BrdU.
This means that tumor cells are unable to synthesize
the DNA in the S-phase.12 Western blot (WB) analysis
from the same cell cultures using specific antibodies
labeling hypo- and hyper-phosphorylated forms of a
CDK2/cyclin A target, pRb, showed that the com-
pounds were able to block hyperphosphorylation of
pRb, at concentrations of 1 and 3 lM, respectively.
Compound
15
18
CDK2/cyclin A (IC50; lM)6,13
CDK2/cyclin E (IC50; lM)6,13
CDK1/cyclin B (IC50; lM)6,13
CDK4/cyclin D1 (IC50; lM)6,13
CDK5/p25 (IC50; lM)6,13
GSK3-b (IC50; lM)6,13
0.033
0.091
0.40
0.036
0.090
0.75
>10
0.28
>10
0.15
0.30
0.28
0.90
0.20
0.050
0.38
0.12
0.07
p42MAPK (IC50; lM)6,13
HCT116 cells (IC50; lM)6
HT-29 cells (IC50; lM)6
We have described the early expansion of a new class of
CDK2 inhibitors. 1,4,5,6-Tetrahydropyrrolo[3,4-c]pyra-
zole compounds were easily obtained from a versatile
scaffold using a simple solution-phase chemistry amena-
ble to rapid expansion.8,9 Compounds with nanomolar
Buffer solubility (lM)6
39
27
CYP4503A4 stability (% of remaining)6 Stable (100) Stable (20)
Plasma protein binding6
95%
99%
Figure 4. Molecular characterization of compounds 15 and 18 in HT-29 tumor cell line.

P. Pevarello et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1084–1090
1089
of (4-pyrrolidin-1-yl)-phenylacetic acid (2.67 g, 13.0
mmol) in DCM (60 ml) and DMF cat., (COCl)2
(1.32 ml, 15.4 mmol) in DCM (15 ml) was added drop-
wise. The mixture was stirred at rt for 90 min and then
concentrated under vacuum, taken up with toluene, and
activity against CDK2/cyclin A/E were obtained which
also show activity on cells at the submicromolar level.
Analysis of the cell cycle profile and of the CDK2 sub-
strate phosphorylation status points toward an antipro-
liferative effect that is mediated by CDK2 inhibition.
From this study compounds 15 and 18 emerged as inter-
esting leads. The unfavorable overall early ADME pro-
file called for an optimization of these compounds in
order to get inhibitors with drug-like in vivo pharmaco-
kinetic profiles suitable for in vivo administration. This
was the focus of further work around this class that will
be reported in due course.
concentrated again.
A solution of the so-obtained
chloride in THF (60 ml) and DCM (40 ml) was added
slowly to a mixture of 3-amino-4,6-dihydro-pyrrolo[3,4-
c]pyrazole-1,5-dicarboxylic acid 5-tert-butyl ester 1-ethyl
ester (3.5 g, 11.8 mmol) and DIEA (10.3 ml, 59.1 mmol)
in THF (70 ml) at 0–5 °C. The reaction was allowed to
reach rt and stirred at the same temperature overnight.
The mixture was evaporated to dryness under vacuum.
The resulting residue was taken up with DCM and the
solution was washed with brine, dried over sodium
sulfate, and evaporated to dryness. The crude product
was purified by flash chromatography (eluent: ethyl
acetate/cyclohexane 35:65 then 40:60) to give 2.4 g of
3-[2-(4-pyrrolidin-1-yl-phenyl)-acetylamino]-4,6-dihydro-
pyrrolo[3,4-c]pyrazole-1,5-dicarboxylic acid 5-tert-bu-
References and notes
1. Stein, G. S.; Baserga, R.; Giordano, A.; Denhard D.T.
Eds.; The Molecular Basis of Cell Cycle and Growth
Control; John Wiley & Sons: New York, NY, USA; 1999.
2. For recent review see: Huwe, A.; Mazitschek, R.; Giannis,
A. Angew. Chem., Int. Ed. 2003, 42, 2122; Kong, N.;
Fotouhi, N.; Wovkulich, P. M.; Roberts, J. Drugs Future
2003, 28, 881.
3. Sausville, E. A. Curr. Med. Chem.: Anti-Cancer Agents
2003, 3, 47; Fischer, P. M.; Gianella-Borradori, A. Exp.
Opin. Invest. Drugs 2003, 12, 955.
4. Tetsu, O.; McCormick, F. Cancer Cell 2003, 3, 233;
Mendez, J. Cell 2003, 114, 398.
5. Martin, A.; Odajima, J.; Hunt, S. L.; Dubus, P.; Ortega,
S.; Malumbres, M.; Barbacid, M. Cancer Cell 2005, 7, 591;
For a commentary on this subject see: Bashir, T.; Pagano,
M. Nat. Cell Biol. 2005, 7, 779.
6. Pevarello, P.; Brasca, M. G.; Amici, R.; Orsini, P.;
Traquandi, G.; Corti, L.; Piutti, C.; Sansonna, P.; Villa,
M.; Pierce, B. S.; Pulici, M.; Giordano, P.; Martina, K.;
Fritzen, E. L.; Nugent, R. A.; Casale, E.; Cameron, A.;
Ciomei, M.; Roletto, F.; Isacchi, A.; Fogliatto, G.;
Pesenti, E.; Pastori, W.; Marsiglio, A.; Leach, K. L.;
Clare, P. M.; Fiorentini, F.; Varasi, M.; Vulpetti, A.;
Warpehoski, M. A. J. Med. Chem. 2004, 47, 3367.
7. Vulpetti, A.; Bosotti, R. Farmaco 2004, 59, 759; Vulpetti,
A.; Pevarello, P. Curr. Med. Chem.—Anti-Cancer Agents
2005, 5, 561.
1
tyl ester (42% yield). ESI MS: m/z 484 (MH+); H NMR
(400 MHz, DMSO-d6) (mixture of rotamers) d ppm 1.30
(t, J = 7.0 Hz, 3H), 1.41,1.42 (s, 9H), 1.92 (m, 4H), 3.16
(m, 4H), 3.43 (s, 2H), 4.34 (m, 4H), 4.50 (m, 2H), 6.46
(d, J = 6.7 Hz, 2H), 7.07 (d, J = 6.7 Hz, 2H), 11.10,
11.12 (s, 1H). 3-[2-(4-pyrrolidin-1-yl-phenyl)-acetylami-
no]-4,6-dihydro-pyrrolo[3,4-c]pyrazole-1,5-dicarboxylic
acid 5-tert-butyl ester 2.4 g (5.0 mmol) in 25 ml DCM
was treated with 16 ml of (1:1) TFA/DCM. The reaction
mixture was stirred for 1 h and evaporated to dryness.
The resulting residue was taken up with DCM, and
treated with aq NaHCO3. The precipitate was filtered,
washed with water and DCM and then dried under
vacuum to obtain 1.7 g (89%) of 3-[2-(4-pyrrolidin-1-yl-
phenyl)-acetylamino]-5,6-dihydro-4H-pyrrolo[3,4-c]pyra-
zole-1-carboxylic acid ethyl ester. ESI MS: m/z 384
(MH+); 1H NMR (400 MHz, DMSO-d6) d ppm 1.28 (t,
J = 7.3 Hz, 3H), 1.92 (m, 4H), 3.16 (m, 4H), 3.42 (s,
2H), 4.02 (s, 2H), 4.17 (s, 2H), 4.33 (q, J = 7.3 Hz, 2H),
6.45 (d, J = 8.5 Hz, 2H), 7.06 (d, J = 8.5 Hz, 2H), 11.0 (s,
1H). n-Butylisocyanate (530 ll, 4.8 mmol) was added to a
suspension of 3-[2-(4-pyrrolidin-1-yl-phenyl)-acetylamino]-
5,6-dihydro-4H-pyrrolo[3,4-c]pyrazole-1-carboxylic acid
ethyl ester (1.4 g, 3.6 mmol) in THF (30 ml). The reaction
mixture was stirred at room temperature for 18 h and
then evaporated. The resulting residue was taken up
with DCM and the solution was washed with brine,
dried over sodium sulfate, and evaporated to dryness.
The crude product was treated with 10% TEA in MeOH
(36 ml) at 40 °C for 30 min and then evaporated. The
solid was triturated with Et2O, filtered, washed, and
dried under vacuum to give 0.8 g of 3-[2-(4-pyrrolidin-1-
yl-phenyl)-acetylamino]-4,6-dihydro-1H-pyrrolo[3,4-c]pyra-
zole-5-carboxylic acid butylamide (15) (53% yield). ESI
MS: m/z 411 (MH+); 1H NMR (400 MHz, DMSO-d6) d
ppm 0.85 (t, J = 7.3 Hz, 3H), 1.24 (m, 2H), 1.38 (m, 2
H), 1.92 (m, 4H), 3.01 (m, 2H), 3.17 (m, 4H), 3.42 (m,
2H), 4.29 (s, 4H), 6.18 (br s, 1H), 6.47 (d, J = 7.7 Hz,
2H), 7.08 (d, J = 7.7 Hz, 2H), 10.40 (s, 1H). HRMS:
exact mass calcd: 411.2503; exact mass found: 411.2523.
8. Fancelli, D.; Berta, D.; Bindi, S.; Cameron, A.; Catana,
C.; Forte, B.; Giordano, P.; Mantegani, S.; Meroni, M.;
`
Moll, J.; Pittala, V.; Severino, D.; Storici, P.; Tonani, R.;
Varasi, M.; Vulpetti, A.; Vianello, P. J. Med. Chem. 2005,
48, 3080.
9. Synthesis of 15: A solution of ethyl chlorocarbonate
(8.9 ml, 93 mmol) in THF (250 ml) was added slowly to
a mixture of 3-amino-4,6-dihydro-1H-pyrrolo[3,4-c]pyra-
zole-5-carboxylic acid tert-butyl ester (20 g, 89 mmol)
and DIEA (92 ml, 528 mmol) in THF (500 ml) at 0–
5 °C. The reaction was kept at the same temperature for
2 h, then allowed to reach rt and stirred overnight. The
reaction mixture was evaporated to dryness under
vacuum. The resulting residue was taken up with AcOEt
and water. The organic phase was separated, dried over
sodium sulfate, and evaporated to dryness. The mixture
was purified by flash chromatography (eluent: ethyl
acetate/cyclohexane 4:6 to 7:3) to give 19 g (72% yield)
of 3-amino-4,6-dihydro-pyrrolo[3,4-c]pyrazole-1,5-dicar-
boxylic acid 5-tert-butyl ester 1-ethyl ester. ESI MS:
m/z 297 (MH+); 1H NMR (400 MHz, DMSO-d6)
(mixture of rotamers) d ppm 1.25 (t, J = 7.1 Hz, 3H),
1.42 (s, 9H), 4.12, 4.17 (m, 2H), 4.25 (q, J = 7.1 Hz,
2H), 4.42, 4.45 (m, 2H), 5.65 (br s, 2H). To a solution
Combustion analysis: calcd: C (64.37); H (7.37); N
(20.47); found: C (63.99); H (7.43); N (20.35).
10. Knight, Z. A.; Shokat, K. M. Chem. Biol. 2005, 12,
621.
11. Ghosh, J. C.; Altieri, D. C. Clin. Cancer Res. 2005, 11,
4580.
12. The apparent percentage of cells in the G2 phase in the
FACS of 18 (Fig. 4) may be ascribed to a fraction of
non-cycling cells unable to escape the nocodazole block,

1090
P. Pevarello et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1084–1090
rather than to an additional G2 block caused by the
compound.
13. At least two independent experiments were performed
for each compound in order to determine IC50 in
replicates, and potency is expressed by the mean of
IC50 values obtained by nonlinear regression analysis.
Coefficient of variance /SD/mean) ranges from 10
to 24%.