The identification of orally bioavailable thrombopoietin agonists

Article abstract of DOI:10.1016/j.bmcl.2009.01.032

Recently, we disclosed a series of potent pyrimidine benzamide-based thrombopoietin receptor agonists. Unfortunately, the structural features required for the desired activity conferred physicochemical properties that were not favorable for the developmen

Full text of DOI:10.1016/j.bmcl.2009.01.032

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1428–1430
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The identification of orally bioavailable thrombopoietin agonists
*
Michael J. Munchhof , Amy S. Antipas, Laura C. Blumberg, William H. Brissette, Matthew F. Brown,
Jeffrey M. Casavant, Jonathan L. Doty, James Driscoll, Thomas M. Harris, Lilli A. Wolf-Gouveia,
Christopher S. Jones, Qifang Li, Robert G. Linde, Paul D. Lira, Anthony Marfat, Eric McElroy,
Mark Mitton-Fry, Sandra P. McCurdy, Lawrence A. Reiter, Sharon L. Ripp, Andrei Shavnya,
Lisa M. Thomasco, Kristen A. Trevena
Pfizer Global Research and Development, Groton Laboratories, MS8220-2405, Eastern Point Road, Groton, CT 06340, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Recently, we disclosed a series of potent pyrimidine benzamide-based thrombopoietin receptor agonists.
Unfortunately, the structural features required for the desired activity conferred physicochemical prop-
erties that were not favorable for the development of an oral agent. The physical properties of the series
were improved by replacing the aminopyrimidinyl group with a piperidine-4-carboxylic acid moiety. The
resulting compounds possessed favorable in vivo pharmacokinetic properties, including good
bioavailability.
Received 1 December 2008
Revised 6 January 2009
Accepted 12 January 2009
Available online 15 January 2009
Keywords:
TPO
Ó 2009 Elsevier Ltd. All rights reserved.
Thrombopoietin
Agonist
The discovery of agonists of the thrombopoietin receptor (TPOr)
is an active area of research within the pharmaceutical industry for
the treatment of a number of conditions resulting in suppressed
platelet levels.1 The FDA recently approved Eltrombopag, shown
in Figure 1, as the first thrombopoietin agonist for the treatment
of idiopathic thrombocytopenic purpura (ITP).2
that displayed the required hydrogen bond acceptor. Additionally,
we replaced the phenyl benzoic acid moiety with the correspond-
ing nicotinic acid to help reduce lipophilicity and facilitate analog
preparation. Combined, these changes were pursued in an attempt
to move to a more drug-like physical property space, including bet-
ter solubility and more polar compounds (Fig. 2).
We recently reported a novel series of potent TPOr agonists
exemplified by 1.3 Unfortunately, the structural features required
for the desired activity conferred physicochemical properties that
were not favorable for the development of an oral agent.4 The poor
solubility of the series made permeability, as well as other in vitro
properties such as metabolic stability, difficult to evaluate.
Based on these criteria a variety of replacements were screened,
resulting in the identification of the pyrrolidine-3- and piperidine-
4-carboxylic acids 2 and 3.6,7 Acid 2 had excellent agonist activity
in the BaF3 reporter assay (EC50 = 33 nM), but showed only a slight
improvement in solubility as compared to 1.8 Compound 3 was not
as potent as 2, but displayed improved aqueous solubility relative
Earlier analoging efforts suggested that one way to obtain the
desired agonist activity was to maintain planarity from the amide
to the thiazole and through the pendent 2-fluoro-3-(trifluorometh-
oxy)phenyl ring of 1.3,5 Efforts to modify this region of the pharma-
cophore while preserving the planarity had previously met with
little success.5 As a result, we turned our attention to replacements
for the diaminopyrimidine portion of the lead. A pharmacophore
model of TPO agonism was also used, which suggested that a
hydrogen bond acceptor was favored in the terminus of the phe-
nylpyrimidine-4,6-diamine, portion of the molecule.5 We therefore
sought to identify structural groups that would remove the planar
diaminopyrimidine and replace it with a heterocycloalkyl group
to 1 (167 vs <1 lg/mL). Unfortunately, improved solubility of 3,
coupled with moderate permeability in the CACO-2 assay, did
not translate to improved oral bioavailability upon administration
HN
OH
N
CO2H
N
N
O
Eltrombopag
* Corresponding author. Tel.: +1 860 441 6189.
E-mail address: Michael.j.munchhof@pfizer.com (M.J. Munchhof).
Figure 1. Eltrombopag.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.01.032

M. J. Munchhof et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1428–1430
1429
OH
N
F
F
CF3
OCF3
O
S
O
S
Cl
N
N
N
N
N
H
H
N
N
N
N
1
2
H
O
BaF3 EC50 = 7 nM
LogD = 4.61
MW = 530
BaF3 EC50 = 33 nM
LogD = 2.94
MW = 528
OH
F
CF3
O
S
N
N
H
N
N
3
BaF3 EC50 = 3300 nM
LogD = 1.21
MW = 480
O
OH
Figure 2. Lead compounds.
in rats (F < 3%). We incorrectly concluded at the time that a further
improvement in solubility was required.9
Table 1
O
O
S
S
Because of the superior potency of the 3-chloro group and
piperidine-4-carboxylic acid combination, we chose to focus fur-
ther work on this series. Thus, we next investigated modification
of the piperidine-4-carboxylic acid moiety. Analogs were prepared
by addition of amines to dichloropyridyl amide 4 (Table 1).
Replacement of the piperidine-4-carboxylic acid with carboxam-
ides, sulfonamides, sulfonic acids, and carboxylic acid isosteres
such as tetrazoles all yielded compounds that maintained a range
of agonist activities. Acyclic acids such as 8, as well as the pipera-
zine-3- and piperidine-3-carboxylic acids 9 and 10, were also mod-
erately potent. Removing the hydrogen bond acceptor of 2 to afford
R
R1
R2
1
R2
N
N
N
N
F
F
F
F
F
F
N
N
Cl
F
F
4
R1
Cl
R2
BaF3 EC50
0.028
(
lM)
logD
4.92
MW
527
a
Compound ID
H2N
O
5
N
O
2
6
Cl
Cl
0.033
0.034
2.94
4.86
528
563
N
piperidine 13 resulted in loss of activity (EC50 > 10 lM). These re-
O
sults are consistent with our hypothesis that a hydrogen bond
acceptor is important for activity and demonstrated that some
flexibility with regard to its general spatial placement can be
tolerated.
O
S
H2N
N
O
We next turned our attention to the pyridine moiety. Replace-
ment of the pyridine ring with less basic pyrimidine and pyrazine
moieties led to inactive compounds (data not shown). These re-
sults suggested that the basicity of the pyridine nitrogen may be
important for activity. To explore this possibility, several 3-substi-
tuted pyridine analogs were prepared with both electron donating
and withdrawing substituents. However, the rank order of potency
of these substituted analogs, Br > Cl > CF3 > CN > Me > H > F/MeO
does not correlate with the calculated basicity of the pyridine
nitrogen (Table 2). Further evidence that pyridine basicity does
not play a critical role was provided by benzene analogs of 2. Com-
pound 16, the direct comparator to 2, is only slightly less potent.
Addition of a second chloro group to the benzene ring (14) im-
proved the potency to 10 nM. Although the benzene analogs pro-
vided useful insight into the SAR, they were compromised by
increased lipophilicity. Unfortunately, none of these changes
showed a significant improvement in activity or solubility over 2.10
During the course of our studies, Smith Kline Beecham disclosed
thrombopoietin agonists of a different chemotype (Fig. 3).11 Of par-
ticular interest was their disclosure of the use of a bis-ethanol-
amine salt to improve the aqueous solubility of this carboxylic
acid-containing compound (22, Eltrombopag) from <0.001 to
14 mg/mL.
N
7
8
Cl
Cl
0.043
0.057
2.79
1.57
528
488
HO
HO
O
HN
O
HN
HO
HO
N
9
Cl
Cl
0.129
0.294
1.88
2.94
545
528
O
N
10
O
O
O
11
12
H
0.430
0.485
>10.0
1.83
3.24
7.1
494
552
484
N
N
N
N
Cl
Cl
N
HN
We prepared the ethanolamine salt of 2 and observed a sim-
ilar dramatic increase. While the parent acid had no measurable
aqueous solubility, the solubility of the ethanolamine salt im-
proved to 15.6 mg/mL. The ethanolamine salt of 2 was pro-
13
N
a
All reported EC50 values were calculated by plotting stimulation index ratio of
drug, against stimulation index ratio of hTPO control.

1430
M. J. Munchhof et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1428–1430
Table 2
Having achieved good oral exposure with 2, we further assessed
R1
the attributes of the compound. PK studies in dogs also showed
good bioavailability, low clearance, and low volume of distribu-
tion.15 Incubation in human liver microsomes resulted in no met-
abolic turnover, and no reactive metabolites were detected. As
expected for a lipophilic carboxylic acid, 2 was highly bound
(>99.5%) to rat, dog, mouse, and human plasma protein (Table 3).
In summary, the replacement of the diaminopyrimidine portion
of 1 with a heterocycloalkyl moiety led to the identification of
compounds with much improved in vivo pharmacokinetic profiles.
In turn, these advances allowed the program to evaluate the in vivo
efficacy of this series in transgenic mouse models.16
O
N
O
N
S
N
F
HO
X
R2
CF3
Compound ID
R1
X
R2
Cl
BaF3 EC50
M)
Calculated Pkb of
pyridyl nitrogen
(
l
14
2
Cl
Cl
Br
Cl
CF3
CN
Me
H
C
0.01
NA
N
N
C
N
N
N
N
N
N
0.033
0.050
0.056
0.134
0.138
0.253
0.430
5.21
5.14
NA
5.07
4.83
7.79
7.50
5.08
6.94
15
16
17
18
19
11
20
21
H
References and notes
1. For recent related publications and background information on thrombopoietin
and thrombopoietin agonists see (a) Revill, P.; Serradell, N.; Bolos, J. Drugs
F
>1.00
>1.00
*
Future 2006, 31, 767; (b) Alper, Phil B.; Marsilje , Thomas H.; Mutnick, Daniel;
MeO
Lu, Wenshuo; Chatterjee, Arnab; Roberts, Michael J.; He, Yun; Karanewsky,
Donald S.; Chow, Donald; Lao, Jianmin; Gerken, Andrea; Tuntland, Tove; Liu,
Bo; Chang, Jonathan; Gordon, Perry; MartinSeidel, H.; Tian, Shin-Shay Bioorg.
*
Med. Chem. Lett. 2008, 18, 5255; (c) Marsilje , Thomas H.; Alper, Phil B.; Lu,
N
N
Wenshuo; Mutnick, Daniel; Michellys, Pierre-Yves; He, Yun; Karanewsky,
Donald S.; Chow, Donald; Gerken, Andrea; Lao, Jianmin; Kim, Min-Ju; Martin
Seidel, H.; Tian, Shin-Shay Bioorg. Med. Chem. Lett. 2008, 18, 5259.
OH
HO
H
N
N
O
O
2. For a recent review of TPO agonists in clinical development for ITP see Stasi,
Roberto; Evangelista, Maria L.; Amadori, Sergio Drugs 2008, 68, 901.
*
HO
3. Reiter, Lawrence A.; Subramanyam, Chakrapani; Mangual, Emilio J.; Jones ,
NH2
Christopher S.; Smeets, Marc I.; Brissette, William H.; McCurdy, Sandra P.; Lira,
Paul D.; Linde, Robert G.; Li, Qifang; Zhang, Fangning; Antipas, Amy S.;
Blumberg, Laura C.; Doty, Jonathan L.; Driscoll, James P.; Munchhof, Michael J.;
Ripp, Sharon L.; Shavnya, Andrei; Shepard, Richard M.; Sperger, Diana;
Thomasco, Lisa M.; Trevena, Kristen A.; Wolf-Gouveia, Lilli A.; Zhang, Liling
Bioorg. Med. Chem. Lett. 2007, 17, 5447.
2
22
Figure 3. SKB lead.
4. For
a review on physiochemical properties and their relationship to oral
exposure see Wenlock Mark, C.; Austin, Rupert P.; Patrick, Barton; Davis,
Andrew M.; Leeson, Paul D. J. Med. Chem. 2003, 46, 1250.
Table 3
Compound 2 PK data
5. For a detailed study of the agonism-inducing conformation of this series see
*
Reiter, Lawrence A.; Jones , Christopher S.; Brissette, William H.; McCurdy,
Sodium salt
Ethanolamine salt
Sandra P.; Abramov, Yuriy A.; Bordner, Jon; DiCapua, Frank; Munchhof, Michael
J.; Rescek, Diane M.; Samardjiev, Ivan J.; Withka, Jane M. Bioorg. Med. Chem. Lett.
2008, 18, 3000.
Rat PK
Clp (mL/min/kg)
6.87
0.831
1.34
66.8
6.34
0.579
1.29
53.4
Vdss (L/kg)
6. Full experimental details are described in WO2007036769.
7. Kalgutkar, Amit S.; Driscoll, James; Zhao, Sabrina X.; Walker, Gregory S.;
Shepard, Richard M.; Soglia, John R.; Atherton, James; Yu, Linning; Mutlib,
Abdul E.; Munchhof, Michael J.; Reiter, Lawrence A.; Jones, Christopher S.; Doty,
Johnathan L.; Trevena, Kristen A.; Shaffer, Christopher L.; Ripp, Sharon L. Chem.
Res. Toxicol. 2007, 20, 1954.
½
T
(h)
F (%)
Dog PK
Clp (mL/min/kg)
Vdss (L/kg)
0.94
0.228
2.31
50
8. The BaF3 reporter assay conditions are described in WO2007036769.
9. This conclusion was based on the IV pharmacokinetic data for 3, that showed
the compound had low in vivo clearance (2.5 mL/min/kg), in addition to the
fact the compound showed moderate permeability in the CACO-2
(4.5 Â 10À6 cm/s) assay.
½
T
(h)
F (%)
10. All of the compounds in Tables 1 and 2 had aqueous solubilities of <10
11. See WO2003098992 for full details.
lg/mL.
gressed into rat PK studies in which the bioavailability was
found to be 54%, which appeared to substantiate our hypothesis
regarding the importance of solubility.12 Interestingly, however,
the sodium salt of 2 was subsequently tested and shown to have
comparable oral exposure and clearance properties, even though
it had no measurable solubility.13 This trend held for several re-
lated acids but was not true in every case.14 Clearly, although
the replacement of diaminopyrimidine portion of 1 had not led
to the desired increase in solubility, it did confer improved prop-
erties relative to the lead, that resulted in favorable in vivo phar-
macokinetic properties.
12. Compound 2 was dosed orally in rats at 5 mg/kg in 0.5% methyl cellulose (aq)
(0.5 mg/mL, 10 mL/kg).
13. Earlier in the program we had prepared sodium salts of several acids and seen
no improvement in solubility.
14. For example, the ethanol amine salt of 3 had greater than 2 mg/mL solubility,
but no detectable bioavailability. The SAR suggested that the piperidine acids
may be substrates for active transporters, but we have no experimental
evidence to support this.
15. Compound 2 was dosed orally in dogs at 5 mg/kg in 0.5% methyl cellulose (aq)
(0.5 mg/mL, 10 mL/kg).
16. Publication in preparation describing the in vivo efficacy of this series. William
H. Brissette, Paul D. Lira, Sandra P. McCurdy, Robin T. Nelson, Kuldeep Neote,
Jeffrey L. Stock, James P. Driscoll, Kristen A. Trevena, Richard M. Shepard,
Christopher S. Jones, Michael J. Munchhof, Lawrence A. Reiter, Sharon L. Ripp.