Part 2: Design, synthesis and evaluation of hydroxyproline-derived α2δ ligands

Article abstract of DOI:10.1016/j.bmcl.2011.04.056

Conformational constraint has been used to design a potent series of α₂δ ligands derived from the readily available starting material (2S,4R)-hydroxy-l-proline. The ligands have improved physicochemistry and potency compared to their linear counterparts (described in our earlier publication) and the lead compound has been progressed to clinical development.

Full text of DOI:10.1016/j.bmcl.2011.04.056

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3767–3770
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Part 2: Design, synthesis and evaluation of hydroxyproline-derived
a2d ligands
⇑
David J. Rawson , Delphine Brugier, Anthony Harrison, Jo Hough , Julie Newman, Joe Otterburn,
Graham N. Maw, Jenny Price, Lisa R. Thompson, Paul Turnpenny, Andrew N. Warren
Sandwich Laboratories, Ramsgate Road, Pfizer Global Research and Development, Sandwich, Kent CT13 9NJ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 20 April 2011
Conformational constraint has been used to design a potent series of a2d ligands derived from the readily
available starting material (2S,4R)-hydroxy-L-proline. The ligands have improved physicochemistry and
potency compared to their linear counterparts (described in our earlier publication) and the lead com-
pound has been progressed to clinical development.
Keywords:
Gabapentin
Ó 2011 Elsevier Ltd. All rights reserved.
a2d
Renal clearance
Glomerular filtration
Gabapentin (NeurontinÒ) (1)1 was launched in 1993 as an add-
on therapy for epilepsy. Positive effects in preclinical models of
neuropathic pain were reported in 1997 and gabapentin was
approved for use in post-herpetic neuralgia (PHN) in 2001. Pregab-
alin (LyricaÒ)(2)2,3 has been approved for the management of neu-
ropathic pain associated with diabetic peripheral neuropathy and
post-herpetic neuralgia and is now a leading medicine in the treat-
ment of neuropathic pain. It has the advantage over gabapentin of
increased potency and linear pharmacokinetics over the efficacious
dose range.4
H2N
CO H
H2N
2
CO2H
H2N
CO2H
NH2
CO2H
Gabapentin
Pregabalin
(2)
Gababutin
(3)
(1)
(4)
CO2H
S
N
CO2H
H
S
NH2
Cl
(6)
(5)
Br
The binding of both gabapentin (1) and pregabalin (2) to a high
affinity binding site, located on the a2d subunit of a voltage-gated
calcium channel, has been reported5 and it is thought that this site
may be involved in the mediation of the pharmacological actions of
both gabapentin and pregabalin.6 Gabapentin and pregabalin bind
The most potent glycine derivatives contain a thioether; how-
ever the compounds suffered from moderate solubility and short
half-lives (as described in the previous paper). In the phenoxy gly-
cine series, improvements in potency could be achieved through
to this
a
2d subunit with IC50 values of 140 and 80 nM, respectively.
2d ligands
-amino
There have been several recent disclosures on novel
a
alkylation
a to the heteroatom (8 and 9 in Table 1). As an extension
including gababutin (3),7 bicyclic gababutins (4)8 and
a
to this observation, further conformational restraint was imposed
through cyclisation to give a phenoxyproline derivative (10). Initial
analogues were derived from the naturally occurring, (2S,4R)-hy-
acids (5).9 Recently, we have focused our attention on improving
the properties of glycine-derived ligands (6).
droxy-L-proline to give final products with the (2S,4S) absolute
configuration. Significant increases in potency were observed and
encouragingly, the phenoxy ethers had improved or equivalent
potency compared to the thioether derivatives (e.g., 10 and 11).
We also examined the effect of the proline stereochemistry and
found that final compounds derived from the unnatural (2R,4R)-
prolinol had significantly reduced potency (12).
A simple synthetic route was used to rapidly access a large num-
ber of analogues (Scheme 1). The amine and acid functionalities of
(2S,4R)-hydroxy-L-proline (13) were protected using standard
⇑
Corresponding author. Tel.: +44 1304 648733.
E-mail address: david.rawson@pfizer.com (D.J. Rawson).
Present address: Centre for Therapeutics Discovery, MRC Technology, 1-3
Burtonhole Lane, London NW7 1AD, UK.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.04.056

3768
D. J. Rawson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3767–3770
Table 1
Table 2
X
Cl
Cl
CO2H
N
H
X
Y
a
R
Compound
X
O
R
a
2d IC50 (nM)
CO2H
CO2H
N
H
N
H
18
4
Compound
X
Y
R
IC50 (nM)
Cl
7
8
9
S
—
—
—
S
O
O
H
H
Me
—
—
—
195
>8000
523
72
32
O
O
—
—
—
11
O
32
10 (2S, 4S)
11 (2S,4S)
12 (2R,4S)
>10,000
Cl
Cl
22
23
24
OCH2
48
HO
HO
HO
CO2Me
i
ii
N
CO2H
CO2Me
.HCl
N
H
13
N
H
O
O
102
399
O
15
14
Cl
Cl
Cl
N
O
O
O
O
iii
iv
Cl
v
CO2H
CO2Me
CO2H
. HCl
N
N
N
H
O
25
26
O
O
11
30
O
O
O
18
16
17
O
Scheme 1. Reagents and conditions: (i) AcCl, MeOH, 95%; (ii) BOC2O, DCM, NEt3,
97%; (iii) diisopropyl azidodicarboxylate, PPh3, m-chlorophenol, THF, 74%,; (iv)
LiOH, aq THF, 78%; (v) 4 M HCl, dioxane, 95%.
methods. A Mitsunobu reaction using the desired phenol resulted in
clean inversion of the hydroxy stereochemistry10 and conversion of
the protected 4R-hydroxy proline 15 to the desired (4S)-pheno-
xyproline 16. This was doubly deprotected to yield 18 in 51% overall
yield from an inexpensive starting material.
The majority of compounds in Table 2 were made using this
methodology to explore the effect of the aromatic substituent on
potency and pharmacokinetics.
27
28
O
O
1680
17
OCF3
A Mitsunobu reaction was also used to access compounds such
as 22 (Scheme 2). Benzoic acid, PPh3 and DEAD were used to invert
the configuration at the 4-position. Hydrolysis of the resulting ben-
zoate ester yielded the inverted alcohol 19. Benzylation followed
by di-deprotection yielded the diasteromerically pure 4-ben-
zyloxyproline derivative 2210 (following purification on Dowex™
50WX8-200 resin eluting with water followed by 9:1 MeoH/aq
NH3).
Small lipophilic substituents (such as F and Cl), at the meta and
para position are most preferred with larger groups such as naph-
thyl (26) and trifluoromethoxy (27) showing reduced potency. Hal-
ogen substitution at both meta and para positions (33) results in a
fall in potency showing that potency is not solely driven by lipophil-
icity and the conformational rigidity of the proline template does
not allow large groups to be accepted so readily. Ortho substitution
is detrimental to potency (29) and polar groups are not tolerated in
any position on the aromatic ring (31 and 32). The thioethers (X = S)
offer no potency advantage over the phenoxyethers and have infe-
rior physicochemistry.
F
Cl
29
30
31
O
O
O
52
269
CN
OH
>10,000
32
O
94

D. J. Rawson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3767–3770
3769
Table 3
Table 2 (continued)
a
Compound
X
O
R
a
2d IC50 (nM)
Cl
O
R2
33
26
O
F
R
N
R1
Cl
O
34
35
OCH2
14
7
Compound
R
R1
R2
IC50 (nM)
Cl
Cl
36
37
38
39
40
H
H
t-Bu
H
H
H
H
H
Me
Me
H
H
H
1260
>10,000
325
>1000
>1000
S
t-BuOC(O)
nPr
9
S
72
Table 4
Pharmacokinetics of compound in rat and dog
Cl
Compound
Species
Log D
Clu (ml/min/kg)
hPPB (%)
Rat T1/2 (h)
a
IC50 for displacement of tritiated gabapentin from porcine cerebral cortex
membranes as described in Ref. 6
18
18
35
25
33
22
Rat
Dog
Rat
Rat
Rat
Rat
À0.1
À0.1
0.3
41
49
56
111
108
26
93
93
84
89
95
83
0.4
1.5
0.3
1
1.5
0.4
À0.1
0.1
Cl
À0.2
HO
HO
O
i
ii
CO2Me
Table 5
Human pharmacokinetic parameters for compound 18
CO2Me
N
N
CO2Me
N
O
O
O
Dose (mg)
10
Cmax (mM)
Tmax (h)
0.5
T1/2 (h)
O
O
19
20
15
Cl
O
2.4
2.3
Cl
iv
iii
As there is currently a very poor understanding of the translation
of active renal processes between species and compound 18 had an
unprecedented level of a2d potency, good absorption and excellent
physicochemistry, the compound was progressed to first in human
studies to assess safety and human pharmacokinetics. The pharma-
cokinetic parameters in man are summarised in Table 5.
Compound 18 showed good absorption in man and linear in-
creases in AUC over the clinical dose range (10–600 mg), however
a short half-life was seen suggesting that active renal secretion is
occurring. The half-life was not considered commensurate with
b.i.d. administration and the compound was not progressed further
as a standard oral agent for chronic use in the treatment of neuro-
pathic pain. The potency, high solubility and good oral absorption
do, however, offer other opportunities for alternative formulations
(such as controlled release) or parenteral administration.
O
O
CO2H
22
21
CO2Me
N
H
N
H
Scheme 2. Reagents and conditions: (i) (a) diisopropyl azidodicarboxylate, PhCO2H,
PPh3, THF; (b) LiOH, aq THF/MeOH; (c) K2CO3, acetone, 81%; (ii) NaH, 3-Cl-benzyl
bromide, DMF, 40%; (iii) LiOH, aq THF, 94%; (iv) trifluoroacetic acid, DCM, purified
on Dowex™ 50WX8-200 resin, 5%.
We also examined the effect of stereochemistry and alkylation on
the proline template through the synthesis of both a-alkylated dia-
stereomers (36 and 37) through methylation of 16 and separation of
the diastereomers. The methods outlined in Scheme 1 were used to
synthesis 38 and 39; 40 was made throughremoval of the carbamate
protecting group of (16) followed by a standard reductive alkylation
protocol and deprotection. All derivatives showed a marked reduc-
tion in potency and were not progressed further (Table 3).
Rat and dog pharmacokinetic studies were carried out on lead
compounds from the initial SAR analysis (Table 4). In both species
the clearance of free drug (Clu) was found to be significantly higher
than glomerular filtration rate and unchanged parent was recov-
ered in the urine.11 This suggested an active renal secretion process
was the dominant clearance pathway.
Within the 4-phenoxy analogues, the compounds with in-
creased lipophilicity (e.g., 33) showed the highest PPB and longest
half-lives consistent with a reduced total clearance, however, po-
tency could not be maintained and compound 18 had the best
overall profile.
Acknowledgment
We thank the staff of the Structural and Analytical Services
Department, Sandwich, for analytical data.
References and notes
1. Bryans, J. S.; Wustrow, D. J. Med. Res. Rev. 1999, 19, 149.
2. Rice, A. S. C.; Maton, S. Pain 2001, 94, 215.
3. Dworkin, R. H.; Kirkpatrick, P. Nat. Rev. Drug Disc. 2005, 4, 455.
4. Zareba, G. Drugs Today 2005, 41, 509.
5. Stahl, S. M. J. Clin. Psychiatry 2004, 65, 1033. and references therein.
6. Gee, N. S.; Brown, J. P.; Dissanayake, V. U. K.; Offord, J.; Thurlow, R.; Woodruff,
G. N. J. Biol. Chem. 1996, 271, 5768.

3770
D. J. Rawson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3767–3770
7. Blakemore, D. C.; Bryans, J. S.; Carnell, P.; Carr, C. L.; Chessum, N. E. A.; Field, M.
J.; Kinsella, N.; Osborne, S. A.; Warren, A. N.; Williams, S. C. Bioorg. Med. Chem.
Lett. 2010, 20(1), 248.
8. Blakemore, D. C.; Bryans, J. S.; Carnell, P.; Carr, C. L.; Chessum, N. E. A.; Field, M.
J.; Kinsella, N.; Osborne, S. A.; Warren, A. N.; Williams, S. C. Bioorg. Med. Chem.
Lett. 2010, 20(2), 461.
9. Mortell, K. H.; Anderson, D. J.; Lynch, J. J., III; Nelson, S.; Sarris, S.; McDonald, H.;
Sabet, R.; Baker, S.; Honore, P.; Lee, C.-H.; Jarvis, M. F.; Gopalakrishnan, M.
Bioorg. Med. Chem. Lett. 2006, 1138.
10. As determined by 400 MHz 1H NMR.
11. 85% of 16 recovered in rat urine over 24 h.