Kinesin spindle protein (KSP) inhibitors. Part 3: Synthesis and evaluation of phenolic 2,4-diaryl-2,5-dihydropyrroles with reduced hERG binding and employment of a phosphate prodrug strategy for aqueous solubility

Article abstract of DOI:10.1016/j.bmcl.2005.12.094

2,4-Diaryl-2,5-dihydropyrroles have been discovered to be novel, potent and water-soluble inhibitors of KSP, an emerging therapeutic target for the treatment of cancer. A potential concern for these basic KSP inhibitors (1 and 2) was hERG binding that can be minimized by incorporation of a potency-enhancing C2 phenol combined with neutral N1 side chains. Aqueous solubility was restored to these, and other, non-basic inhibitors, through a phosphate prodrug strategy.

Full text of DOI:10.1016/j.bmcl.2005.12.094

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1780–1783
Kinesin spindle protein (KSP) inhibitors. Part 3: Synthesis
and evaluation of phenolic 2,4-diaryl-2,5-dihydropyrroles
with reduced hERG binding and employment of a
phosphate prodrug strategy for aqueous solubility
Robert M. Garbaccio,a,* Mark E. Fraley,a Edward S. Tasber,a Christy M. Olson,a
William F. Hoffman,a Kenneth L. Arrington,a Maricel Torrent,a Carolyn A. Buser,b
Eileen S. Walsh,b Kelly Hamilton,b Michael D. Schaber,b Christine Fernandes,b
Robert B. Lobell,b Weikang Tao,b Vicki J. South,b Youwei Yan,c Lawrence C. Kuo,c
Thomayant Prueksaritanont,d Donald E. Slaughter,d Cathy Shu,d
David C. Heimbrook,b Nancy E. Kohl,b Hans E. Huberb and George D. Hartmana
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Cancer Research, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Structural Biology, Merck Research Laboratories, West Point, PA 19486, USA
dDepartment of Drug Metabolism, Merck Research Laboratories, West Point, PA 19486, USA
Received 4 November 2005; revised 20 December 2005; accepted 20 December 2005
Available online 24 January 2006
Abstract—2,4-Diaryl-2,5-dihydropyrroles have been discovered to be novel, potent and water-soluble inhibitors of KSP, an emerg-
ing therapeutic target for the treatment of cancer. A potential concern for these basic KSP inhibitors (1 and 2) was hERG binding
that can be minimized by incorporation of a potency-enhancing C2 phenol combined with neutral N1 side chains. Aqueous solu-
bility was restored to these, and other, non-basic inhibitors, through a phosphate prodrug strategy.
Ó 2006 Elsevier Ltd. All rights reserved.
In the preceding communication, the design, synthesis,
and characterization of dihydropyrrole-based inhibitors
F
F
F
F
of KSP (kinesin spindle protein) as novel antimitotics
were described.1 Representative inhibitors 1 and 2 dis-
played low nanomolar potency against KSP and were
shown to induce mitotic arrest in transformed cells
(Fig. 1). These inhibitors were promising candidates
for further development as intravenous anticancer
agents due to their pharmacokinetic profile and aqueous
solubility. However, a potential concern for these
amide- and urea-based dihydropyrroles 1 and 2 was
their ancillary activity against the IKr potassium channel
hERG (human Ether-a-go-go-Related Gene). Blockade
H
H
N
Me
N
N
O
O
NH2
1
2
N
H
KSP IC50 = 2.0 nM
Cell EC50 = 8.6 nM
hERG IC50 = 3500 nM
aq. sol. >10 mg/mL
KSP IC50 = 2.6 nM
Cell EC50 = 6.8 nM
hERG IC50 = 1300 nM
aq. sol. >10 mg/mL
- enhance potency
N1 acyl groups
- aqueous solubility
- potentiate hERG binding
Keywords: KSP; Kinesin spindle protein; Kinesin; hERG; Phosphate
prodrug; Cancer; Antimitotic; Dihydropyrrole; Phenol.
with basic amines
*
Corresponding author. Tel.: +1 215 652 5391; fax: +1 215 652 6345;
e-mail: robert_garbaccio@merck.com
Figure 1. Lead dihydropyrrole KSP inhibitors.
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.12.094

R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1780–1783
1781
of hERG has been implicated in the drug-induced
Pd(OAc)2
Ph3As
Chiracel
F
F
F
F
F
alteration of cardiac ventricular repolarization that is
typically manifested in prolongation of the heart rate-
corrected QT interval (QTc).2 Prolongation of the QTc
interval may potentiate cardiac arrhythmia in patients
and in rare instances lead to torsades de pointes or sud-
den death,3 and therefore part of our optimization ef-
forts were directed to minimize hERG binding.
OD
OTBS
OTBS
ArI (6)
µw
150 ˚C
56%
α = 1.18
N
N
Boc
Boc
5
7
F
F
F
acylation
TFA
H
H
OTBS
(see
below)
Structure–activity relationships (SAR) in this series re-
vealed that hERG binding was potentiated by the basic
amine in the N1 acyl group. However, deletion or acyl-
ation of this amine compromised potency and, to a
greater extent, aqueous solubility. An alternative strate-
gy was devised to capitalize on known SAR and X-ray
crystallographic information from dihydropyrazole 34
and monastrol5 wherein a 30-hydroxyl on the ‘eastern’
phenyl ring contributes to potency through formation
of a hydrogen bond to a backbone carbonyl (Glu118)
within the allosteric pocket of KSP (Fig. 2). We hypoth-
esized that for the dihydropyrrole series, similar incor-
poration of the potency-enhancing phenol would
enable replacement of the basic amine at N1 with
neutral acyl groups and result in minimization of hERG
binding affinity.
CH2Cl2
>95%
N
H
N
Boc
8
9
F
F
F
F
K2CO3
H
H
OTBS
OH
DMF
90%
N
N
R1
O
R1
O
10
11
Acylation Conditions:
For Amides:
PyBOP
Et3N
F
F
9
H
OTBS
DMF
O
N
R2
R2
O
OH
R3
12
An effort to adopt this structural feature into the dihy-
dropyrrole series was undertaken (Scheme 1). Utilizing
the three-step racemic route to 2,4-diaryl 2,5-dihydro-
pyrroles6 described in the preceding communication,
intermediate 2,3 dihydropyrrole 5 was examined as a
substrate for Heck coupling to a phenol precursor. It
was discovered that 2,3-dihydropyrrole 5 undergoes
microwave promoted Heck coupling with aryl iodides,
and this procedure was optimized for 3-(tert-butyldim-
ethylsiloxy)iodobenzene (6). The optimized conditions
utilized Pd(OAc)2 (0.1 equiv) with Ph3As (0.2 equiv) as
ligand and proceeded for 30 min at 150 °C under micro-
wave irradiation to produce the desired 2,4-diaryl-2,
5-dihydropyrrole (7) in 56% yield following purification.
R3
yields: 50-90%
1. CDI, 70 ˚C
2. MeI, 60 ˚C
90%
For Ureas:
F
F
9
H
OTBS
3. R4R5NH
N
R4
yields: 25-75%
N
O
R5
13
Scheme 1. Synthesis of dihydropyrrole C2-phenols.
Incorporation of the phenol into the amide- and urea-
based dihydropyrroles produced a 3- to 5-fold potency
enhancement for most inhibitors evaluated (e.g., 1 vs
16). A representative set of KSP inhibitors is presented
(Table 1) along with their potency against KSP,9 cell
potency in A2780 human ovarian carcinoma cells10,
and hERG activity.11 Combination of the C2-(30-phe-
nol) with the potency-optimized basic N1-acyl groups
afforded highly potent (IC50 < 3 nM) KSP inhibitors
(15 and 16) with excellent cell-based activity; however,
the hERG binding was only minimally improved by
the presence of the phenol alone. The strategy for signif-
icant reduction of hERG binding was realized in the
combination of the phenol with neutral N1-acyl groups
that, together, retained KSP inhibitory activity with
IC50 < 10 nM. KSP inhibitors 17–20, that possessed
neutral a-hydroxy amides and ureas, displayed excellent
potency and reduced hERG binding. Interestingly, the
basic urea 21 did not show significant potency enhance-
ment when combined with the phenol.
The racemic dihydropyrrole (7) was resolved into pure
enantiomers using chiral chromatography and only the
S-enantiomer7 (8) was carried forward. Removal of
the Boc-protecting group using TFA revealed dihydro-
pyrrole 9 and acylation under conditions described in
the preceding communication: N1 amides (12) were syn-
thesized by PyBOP coupling, whereas ureas (13) were
produced either via the N1-acyl imidazolium salt or di-
rect acylation with a carbamoyl chloride. Final depro-
tection of the phenol was accomplished using K2CO3
in DMF to give the phenol-bearing KSP inhibitors (11).8
S
F
F
F
F
NH
OH
HN
Me
EtO
H
H
OH
OH
N
H
O
2
N
N
Me
O
R
O
3
4
(S)-Monastrol
C2-Phenol
Dihydropyrrole
KSP IC50 = 10,000 nM KSP IC50 = 26 nM
Based on their overall profile including diminished
hERG activity, phenols 17–19 were considered for
Figure 2. Selected and proposed phenolic KSP inhibitors.

1782
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1780–1783
Table 1. Properties of selected C2-phenol
and KSP inhibitors 17–19, 28 and 32 were ideally suited
for this strategy. Interestingly, three distinct synthetic
strategies were required to access phosphates of a phe-
nol, primary alcohol and secondary alcohol, respective-
ly. For the phenols (17–19, Scheme 2), in situ generation
of dibenzylchlorophosphate13 gave in high yield diben-
zylphosphate esters (22–24). As reported, this procedure
is highly chemoselective for phenols over aliphatic alco-
hols. Transfer hydrogenation of the benzyl groups yield-
ed the desired phosphate prodrugs (25–27) in >95%
yield and purity.
F
F
H
OH
N
R1
11
Compound
R1
KSP IC50
(nM)
Cell EC50
(nM)
hERG IC50
(nM)
14a
13
37
7800
O
For the primary alcohol (28), the dibenzylphosphate
was synthesized in a convergent manner (Scheme 3).
Phosphorylated ethylene glycol was coupled to the
4-nitrophenylcarbamate (29) at 100 °C in neat Et3N.
Transfer hydrogenation produced the desired prodrug
31. In the final case, the hindered secondary alcohol
(32) was phosphorylated by chlorodimethylphosphate
in the presence of catalytic Ti(OtBu)4.14 Deprotection
was effected by methanesulfonic acid in dimethylsulfide
to give the prodrug 34. Importantly, the presence of a
phosphate rendered these compounds inactive (phenols
15
16
17
18
19
O
O
1.2
0.5
2.9
1.3
7.0
3.3
3.0
3.1
2.6
22
8500
NH2
5600
NH2
OH
18,000
7800
O
O
OH
(BnO)2POH
Me
Me
F
F
CCl4
N
Me
O
33,000
F
F
O
H
H
OH
O P OR
OR
i-PrNEt2
DMAP
CH3CN
N
N
R1
R1
N
O
O
20
4.0
5.5
9.1
8.5
15,000
6600
R1
0 to 25 ˚C
OH
Pd/C, MeOH
, >95%
22 R = Bn
25 R = H
19
17
79%
82%
Me2N
O
Me
N
Pd/C, MeOH
, >95%
23 R = Bn
26 R = H
21a
O
O
OH
N
H
Pd/C, MeOH
, >95%
24 R = Bn
27 R = H
a Tested as a racemic mixture.
18
84%
OH
Scheme 2. Synthesis of phenolic phosphate prodrugs.
further characterization along with two additional KSP
inhibitors, 28 and 32.1 A key limitation to these inhibi-
tors was that, with the absence of the basic amine, aque-
ous solubility was negligible and additional formulation
would be required for use as intravenous agents. It was
preferable to avoid the use of vehicles such as Crema-
phorÓ that bear their own side effects and toxicities.
To circumvent these potential issues, chemical derivati-
zation to a water-soluble prodrug was examined.
F
F
F
F
O
HO
O P OBn
OBn
H
H
N
N
Et3N
O
O
O
O
100 ˚C
50%
HN OR
HN
O P OR
OR
O
O
28 R = CH2CH2OH
29 R = 4-NO2Ph
Phosphate prodrugs have been widely used to enhance
the aqueous solubility of parent drugs, and this strategy
has been adopted for other clinical anticancer agents
including etoposide, pancratistatin and taxol.12 The
anionic charge imparted by the phosphate at nearly all
physiological pH values dramatically increases the
polarity and aqueous solubility over the parent mole-
cule. Upon dosing in vivo, these phosphates are expect-
ed to undergo rapid hydrolysis by non-specific
phosphatases and to produce the parent drug with bio-
equivalence and non-toxic inorganic phosphate.
Pd/C, MeOH
, >95%
30 R = Bn
31 R = H
F
F
(MeO)2POCl
F
F
H
N
H
Ti(OtBu)4
Et3N
N
O
O
OH
O
RO P OR
32
O
Me2S
MeSO3H
33 R = Me
34 R = H
34% for two steps
Typically the phosphate moiety is appended to an avail-
able hydroxyl functional group in a parent molecule,
Scheme 3. Synthesis of aliphatic phosphate prodrugs.

R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1780–1783
1783
Table 2. Properties of selected phenols and prodrugs
readily converted to parent than did the corresponding
aliphatic phosphates. These potent, water-soluble KSP
inhibitor prodrugs are suitable for iv administration
and may exhibit safer ancillary profiles due to their
reduced affinity for hERG.
Druga
Clearance t1/2
(mL/min/kg) (h)
Prodruga t1/2
(h)
AUC of
parent (%)b
19 Dog
Rat
13
68
1.5
0.3
25
25
0.25 100
0.15 100
28 Dog
Rat
11
68
2.8
0.3
31
31
<0.1
<0.1
38
57
Acknowledgments
a Dosed in DMSO at 0.4 mpk iv to dog and at 1 mpk iv to rat.
b Ratio of AUC’s for parent from prodrug dosing and direct parent
dosing.
We thank Dr. Chuck Ross for high resolution mass
spectral analysis and Carl Homnick for chiral
separations.
>1 lM) or less active (alcohols >100 nM) towards KSP
in the enzymatic assay, and thus the phosphate prodrug
itself should not contribute to the antiproliferative effect.
In all cases, the aqueous solubility of these inhibitors
was dramatically improved by the phosphate group
from <1 to >20 mg/mL at pH 7.
References and notes
1. Preceding communication.
2. De Ponti, F.; Poluzzi, E.; Cavalli, A.; Recanatini, M.;
Montanaro, N. Drug Safety 2002, 25, 263.
3. Roden, D. M. N. Eng. J. Med. 2004, 350, 1013.
4. Cox, C. D.; Breslin, M. J.; Mariano, B. J.; Coleman, P. J.;
Buser, C. A.; Walsh, E. S.; Hamilton, K.; Huber, H. E.;
Kohl, N. E.; Torrent, M.; Yan, Y.; Kuo, L. C.; Hartman,
G. D. Bioorg. Med. Chem. Lett. 2005, 15, 2041.
5. Yan, Y.; Sardana, V.; Xu, B.; Homnick, C.; Halczenko,
W.; Buser, C. A.; Schaber, M.; Hartman, G. D.; Huber, H.
E.; Kuo, L. C. J. Mol. Biol. 2004, 335, 547.
An important consideration for the use of phosphate
prodrugs is their rate of conversion to parent in vivo,
and these varied significantly for the prodrugs studied
(Table 2). The dog and rat pharmacokinetic profiles
for the parent KSP inhibitors 19 and 28 and their corre-
sponding prodrugs 25 and 31, respectively, are shown in
detail. The key parameter is a comparison of AUC de-
rived from direct dosing of the parent and the AUC of
parent achieved by prodrug dosing. When the phenolic
phosphate 25 was dosed to rats and dogs, it provided
approximately 100% of the AUC seen with direct dosing
of the parent (iv), indicative of complete prodrug-to-
parent conversion. The prodrugs for both aliphatic alco-
hols (31 and 34), however, appeared to also undergo
clearance mechanism(s) other than conversion to parent;
for 31 the percent AUC of parent achieved was 57% in
rats and 38% in dogs when the prodrug was dosed.
Not surprisingly, the hindered secondary phosphate 34
produced little parent and showed poor conversion in
other in vivo studies. In general, the phenolic phos-
phates appeared to be well-behaved prodrugs showing
bioequivalence to their corresponding parent drug.
6. Carpes, M. J. S.; Correia, C. R. D. Synlett 2000, 1037.
7. Chiracel OD column, 85% hexanes/10% MeOH/5% EtOH,
a = 1.18, S was second to elute.
1
8. All compounds were characterized by H NMR and high
resolution mass spectrometry. For detailed experimental
procedures, see WO03105855 and WO04111193
(prodrugs).
9. KSP inhibitory activity was measured using a standard
ATPase assay. IC50 values are reported as averages of
at least two determinations; standard deviations are
25–50%.
10. Mitotic arrest was measured by antibody-based detection
of phosphonucleolin, an M-phase marker protein. Values
are reported as averages of at least two determinations;
standard deviations are
conditions.
25–50%. See Ref. 1 for
11. The hERG IC50 values are reported as averages of at least
two determinations and were acquired by binding compe-
tition experiments using membrane preparations from
human embryonic kidney cells that constitutively express
hERG.
12. Sinhababu, A. K.; Thakker, D. R. Adv. Drug Delivery
Rev. 1996, 19, 241.
13. Silverberg, L. J.; Dillon, J. L.; Vemishetti, P. Tetrahedron
Lett. 1996, 37, 771.
In conclusion, 2,5-dihydropyrroles consisting of a
30-phenol at C2 were synthesized and found to display
enhanced potency against the mitotic kinesin KSP ver-
sus their C2-phenyl counterparts. In combination with
preferred neutral N1 acyl groups, these phenols showed
an improved hERG binding profile, but limited aqueous
solubility. A phosphate prodrug strategy was employed
that engendered >40 mg/mL aqueous solubility to the
lead phenols, and these phenolic phosphates more
14. Jones, S.; Selitsianos, D.; Thomspson, K. J.; Toms, S. M.
J. Org. Chem. 2003, 68, 5211.