Synthesis and in vivo evaluation of 3-substituted gababutins

Article abstract of DOI:10.1016/j.bmcl.2009.10.089

A range of 3-alkylated five-membered ring derivatives of Gabapentin were synthesized and several were found to have good levels of potency against the α2δ calcium subunit of a voltage-gated calcium channel. Two compounds were profiled in in vivo models of pain and anxiety.

Full text of DOI:10.1016/j.bmcl.2009.10.089

Bioorganic & Medicinal Chemistry Letters 20 (2010) 362–365
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and in vivo evaluation of 3-substituted gababutins
a
a
a
a
David C. Blakemore a, , Justin S. Bryans , Pauline Carnell , Nicola E. A. Chessum , Mark J. Field ,
*
Natasha Kinsella a, Jack K. Kinsora b, Simon A. Osborne a, Sophie C. Williams a
a Sandwich Laboratories, Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent, CT13 9NJ, UK
b Groton Laboratories, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 00340, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A range of 3-alkylated five-membered ring derivatives of Gabapentin were synthesized and several were
Received 11 October 2009
Revised 18 October 2009
Accepted 21 October 2009
Available online 25 October 2009
found to have good levels of potency against the
Two compounds were profiled in in vivo models of pain and anxiety.
a2d calcium subunit of a voltage-gated calcium channel.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Gababutin
Gabapentin
Pregabalin
Pain
Anxiety
Gabapentin (NeurontinÒ) (1)1 was launched as an add-on ther-
apy for epilepsy in 1994. Subsequently, positive effects in preclin-
ical models of neuropathic pain2,3 and anxiety4,5 have been
reported and its efficacy against neuropathic pain has been demon-
strated clinically in humans.6 Pregabalin7,8 (LyricaÒ) (2), has supe-
rior potency and pharmokinetics9 to gabapentin and has been
approved for the management of neuropathic pain associated with
diabetic peripheral neuropathy and post-herpetic neuralgia.
search for potent and efficacious analogues of gabapentin, we have
investigated a range of five-membered ring analogues which we
have called gababutins. Gababutin (3) has a binding affinity of
420 nM at the gabapentin binding site and we hypothesised that
the gababutin five-membered ring is not optimally filling the avail-
able space in the binding pocket. We reasoned that appending a
variety of alkyl groups to the gababutin ring would allow us to
probe this binding pocket, with the aim of generating compounds
of similar or better potency to gabapentin. For reasons of synthetic
expediency we chose to append substituents to the 3-position of
the gababutin ring.
HO2C
NH2
HO2C
NH2 HO2C
NH2
The synthesis of 3-alkyl gababutins presents a number of chal-
lenges with the key difficulty being controlling the stereochemistry
of the quaternary centre carrying the aminomethyl group.
NH2 CO2H
O
(1)
(2)
(3)
Control of key
quaternary centre
challenging
Gabapentin
IC50=140nM
Pregabalin
IC50=80nM
Gababutin
IC50=420nM
R
The binding of both (1) and (2) to a high affinity binding site,
Single enantiomer
required
located on the
a2d subunit of a voltage-gated calcium channel,
has been reported10 and it is thought that this site may be involved
in the mediation of the pharmacological actions of both gabapentin
Additionally, the 3-alkyl cyclopentanones required as starting
materials for the syntheses are not easy to obtain commercially:
few are available as racemates, and only 3-methyl cyclopentanone
is available in enantiomerically pure form as the (R)-(+)-isomer.
For this reason, initial work focused on the 3-methyl gababutin.
Our initial synthetic strategy is shown in Scheme 1. Ketone (4)
and pregabalin.11 Gabapentin and pregabalin bind to this
a2d sub-
unit with IC50 values of 140 nM and 80 nM, respectively. A number
of groups have carried out SAR studies around a
2d ligands.12 In our
was converted to
a,b-unsaturated ester (5) via a Horner–Emmons
* Corresponding author. Tel.: +44 1304 644573.
E-mail address: david.blakemore@pfizer.com (D.C. Blakemore).
reaction. Subsequent nitromethane anion addition gave an
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.10.089

D. C. Blakemore et al. / Bioorg. Med. Chem. Lett. 20 (2010) 362–365
363
NO2 CO2Et
Ph
Ph
CO2H
CO2Et
O
(i)
(ii)
CO2H
CO2Me
CO2Me
(i)
(ii)
(4)
(5)
(6)
(12)
(13)
(14)
55:45 mixture of
diastereoisomers
(iii)
O
H
N
(iii)
MeO
(iv)
NH2
O
H
N
CO2H
CO2Me
NH2 CO2H
(iv)
(16)
(15)
(8)
(7)
Scheme 3. Reagents and conditions: (i) TMSCHN2, toluene (100%); (ii) RuCl3, NaIO4,
CCl4, MeCN, H2O (82%); (iii) DPPA, NEt3, toluene, reflux; MeOH, toluene, reflux (47%
from acid); (iv) 6 N HCl, 1,4-dioxane, reflux (85%).
Scheme 1. Reagents and conditions: (i) triethylphosphonoacetate, NaH, THF, 0 °C
to rt (95%); (ii) MeNO2, TBAF, THF, reflux (65%); (iii) H2, Ni, MeOH; (iv) 6 N HCl, 1,4-
dioxane, reflux (69% from nitroester).
was oxidised to a carboxylic acid with ruthenium tetroxide gener-
ated from catalytic ruthenium trichloride and sodium periodate
giving (14).13 The resulting acid was converted to a carbamate
via a Curtius rearrangement and hydrolysis of this carbamate gave
the amino acid (16) (the relative stereochemistry was confirmed
by NOE studies; chiral HPLC showed the enantiomeric excess to
be 98%).
To form the opposite diastereoisomer, the order of events was
changed (Scheme 4). In this case, benzyl acid (12) was converted
to t-butyl ester (17) followed by oxidation of the phenyl ring to
an acid (18). Conversion of the acid to methyl ester (19) was fol-
lowed by deprotection of the t-butyl ester to give acid (20). Curtius
rearrangement gave carbamate (21) and finally, acidic hydrolysis
gave the target amino acid (22) (the relative stereochemistry being
confirmed by NOE studies; enantiomeric excess of 98% was estab-
lished by chiral HPLC).
To investigate the impact of varying the 3-alkyl group on
potency of the gababutins, we needed access to a range of enanti-
omerically pure 3-alkyl cyclopentanones. To achieve this, we used
3-acetoxy cyclopent-2-en-1-one. Routes to both enantiomers of
this compound are known in the literature.14–16 Conjugate addition
of an alkyl zincate reagent to the acetoxy cyclopentenone (made
with enantiomeric excess of >99%) followed by elimination and
hydrogenation gave us 3-alkyl cyclopentanones without loss of
stereochemical integrity (Scheme 5).
approximately 1:1 mixture of diastereoisomeric nitroesters (6)
with the remote methyl group exerting no control over the face
of addition under the reaction conditions. Hydrogenation to the
lactam (7) and then acid hydrolysis gave a mixture of diastereoiso-
mers of the final amino acid (8).
To synthesise the individual diastereoisomers of 3-methyl gab-
abutin (8) we synthesized benzyl acid (12) as a single diastereoiso-
mer (Scheme 2).
Knoevenagel condensation of malonitrile onto ketone (4) gener-
ated the unsaturated cyanoester (9) with no apparent geometric
control. Low temperature addition of benzylmagnesium chloride
to the cyanoester gave a mixture of diastereoisomeric benzyl
cyanoesters (10) which on hydrolysis gave a 7:3 mixture of diaste-
reoisomers of benzyl acid (11) with the dominant diastereoisomer
forming as a result of the bulky benzyl group coming in on the
opposite face to the methyl group on the cyclopentyl ring. The
(S)-(À)-
a-methylbenzylamine salt of this acid could be recrystal-
lised to give the major product as a single diastereoisomer. Acid
treatment of the salt gave benzyl acid (12) as a single diastereoiso-
mer (the relative stereochemistry being confirmed by NOE stud-
ies). Benzyl acid (12) could be converted into either
diastereoisomer of the final 3-methyl gababutin by varying the
order of reactions.
To get the diastereoisomer where the aminomethyl group is on
the opposite face to the methyl group (16) (Scheme 3), acid (12)
was converted to the methyl ester (13) and then the phenyl ring
Ph
Ph
CO2H
(ii)
(i)
CO2H
CO2t-Bu
CO2t-Bu
NC
CO2Et
Ph CN
CO2Et
O
(i)
(ii)
(12)
CO2Me
(17)
CO2Me
(18)
(iii)
O
CO2Me
(4)
(9)
(10)
OMe
N
H
CO2H
CO2t-Bu
(v)
(iv)
(iii)
(21)
(20)
(19)
Ph
Ph CO2H
Ph COO-
Ph
(vi)
(v)
(iv)
CO2H
CO2H
NH3+
NH2
(12)
(11)
Single
diastereoisomer
70:30 mix of
diastereoisomers
(22)
Scheme 2. Reagents and conditions: (i) NCCH2CO2Et, NH4OAc, AcOH, toluene,
reflux (97%); (ii) BnMgCl, THF, À78 °C (94%); (iii) KOH, ethylene glycol, 160 °C
Scheme 4. Reagents and conditions: (i) oxalyl chloride, DMF, DCM; t-BuOH, DIPEA,
DCM (88%); (ii) RuCl3, NaIO4, CCl4, MeCN, H2O (73%); (iii) TMSCHN2, toluene (92%);
(iv) TFA, DCM (98%); (v) DPPA, NEt3, toluene, reflux; MeOH, toluene, reflux (63%);
(vi) 6 N HCl, 1,4-dioxane, reflux (60%).
(77%); (iv) (s)-(À)-
a-methylbenzylamine, EtOAc, 0 °C; 2 Â recrystallisation (40%);
(v) dilute HCl (91%).

364
D. C. Blakemore et al. / Bioorg. Med. Chem. Lett. 20 (2010) 362–365
affinities (a radioligand binding assay incorporating [3H]gabapen-
tin at the 2d subunit of a calcium channel was utilised as previ-
O
O
(24)
O
O
a
(i)
(ii)
ously described10) (Table 1). The 3-methyl, ethyl and n-propyl
substituents increased binding affinity relative to the starting gab-
abutin, while larger groups and dialkyl substitution were detri-
mental showing that space in the binding pocket is relatively tight.
The impact of the chirality of the 3-alkyl stereocentre on bind-
ing was now examined (Table 2). For ease of synthesis, while the 3-
alkyl stereocentre was set, the quaternary stereogenic centre was
undefined and compounds of type A and B were approximately
1:1 mixtures of diastereoisomers.
AcO
AcO
O
(23)
(25)
(iii)
O
(iv)
(27)
(26)
The stereochemistry at the 3-position was critical for activity
with compounds of type A, with the (R)-configuration at the 3-al-
kyl position, having good binding affinities and compounds of type
B, with an (S)-configuration at the 3-alkyl position, binding weakly.
To assess the impact of the quaternary stereogenic centre on
binding affinity, the single diastereoisomers of the 3-(R)-methyl-
gababutin were synthesized.
Scheme 5. Reagents and conditions: (i) n-PrMgCl, Me2Zn, THF, À78 °C; (ii) DBU,
Et2O, À40 °C (68% over two steps); (iii) H2, 10% Pd/C, EtOAc (88%); (iv) (2R,3R)-(À)-
2,3-butanediol, p-TSA, benzene, reflux (85%).
For example, conjugate addition of the n-propyl zincate (gener-
ated from n-propyl magnesium chloride and dimethylzinc) occurred
exclusively on the least hindered face of the cyclopentenone (the
acetoxy group forcing the n-propyl group onto the opposite face)
to generate propyl cyclopentanone (24). Low temperature elimina-
tion with DBU followed by hydrogenation occurred without any
epimerisation to give the 3-propyl cyclopentanone (26) in >98%
enantiomeric excess (as determined by chiral gas chromatography
or by derivatisation to acetals such as (27) and examination of 13C
NMR spectra).
CO2H
CO2H
H2N
H2N
(16)
IC50=30nM
(22)
IC50=71nM
As well as using 3-acetoxy cyclopent-2-en-1-one to synthesise
the chiral 3-propyl cyclopentanone, we tested out the use of the
chiral cyclopentadienone synthon, ketodicyclopentadiene (28)17
with the aim of obtaining benzyl acid derivative (34) as a single
diastereoisomer (Scheme 6). Conjugate addition of propylzincate
to ketone (28) gave the desired propyl ketone (29) as a single dia-
stereoisomer. Knoevenagel condensation was followed by benzyl
Grignard addition to unsaturated cyanoester (30). Following basic
hydrolysis, benzyl acid (32) was obtained as a single diastereoiso-
mer. The retro Diels–Alder reaction occurred without epimerisa-
tion to give (33) which was hydrogenated to give (34) as a single
diastereoisomer.
Both compounds (16) and (22) had good binding affinities at the
gabapentin binding site and so were evaluated in in vivo models
of pain and anxiety.
Firstly, both compounds were evaluated in the carageenan in-
duced thermal hyperalgesia model of pain (Fig. 1). In this model,
the pro-inflammatory agent carrageenan was administered via
intraplantar injection into the paw of a rat. Within two hours of
injection, the paw became inflamed and hyperalgesia had been in-
duced. An infra-red light source was shone on the paw and the
time it took the animal to withdraw its paw from the heat-source
was measured. Before treatment, baseline measurements were ta-
ken giving a reading of 10 s to paw withdrawal. After 2 h, once the
Initially, we synthesized the alkylated gababutins as equimolar
mixtures of all four possible isomers and determined their binding
Table 1
H
(i)
(ii)
HO2C
HO2C
NH2
H
H
H
H
H
O
NC
NC
O
(28)
(29)
(30)
(iii)
R
CO2Et
NH2
H
(v)
(iv)
H
H
H
HO2C
(33)
Ph
EtO2C
Ph
HO2C
Ph
IC50=3180nM
(32)
(31)
(vi)
R
IC50 (nM)
H
420
Me
82
Et
55
n-Pr
n-Bu
i-Bu
t-Bu
Ph
69
690
760
>10,000
5060
9280
6230
HO2C
Ph
(34)
Scheme 6. Reagents and conditions: (i) n-PrMgCl, Me2Zn, THF, À78 °C (98%); (ii)
NCCH2CO2Et, NH4OAc, AcOH, benzene, reflux (75%); (iii) BnMgCl, Me2Zn, THF,
À78 °C; (iv) KOH, ethylene glycol, 160 °C (75% over two steps); (v) Ph2O, Et2O,
215 °C (58%); (vi) H2, PtO2, EtOAc, 30 °C, 60 psi (99%).
PhCH2
PhCH2CH2

D. C. Blakemore et al. / Bioorg. Med. Chem. Lett. 20 (2010) 362–365
365
Table 2
CO2H
CO2H
H2N
H2N
R
R
A
B
R
Compound
IC50 (nM)
Me
Et
n-Pr
Me
Et
A
A
A
B
B
B
88
53
29
1300
706
1200
n-Pr
*
Figure 2. = P <0.05, significantly different compared to vehicle.
sured at different doses. The more anxiety that the animals feel,
the less they drink and the less shocks they receive. For an effective
compound, as the dose of compound increases then the number of
shocks received by the animals should also increase. The minimum
effective dose (MED) for activity for pregabalin in this model is
10 mg/Kg. Compound (16) had an MED of 3 mg/Kg while com-
pound (22) proved to have poor efficacy in this model having an
MED of 30 mg/Kg.
The difference between the two compounds is surprising and
shows that binding affinity alone is not a good indicator of
in vivo activity. The fact that (16) is only effective in the CITH
model on intrathecal dosing but effective in the Vogel conflict
model on oral dosing shows that there are some subtle effects
at work and that pain and anxiety effects may be mediated differ-
ently. Notwithstanding, we have identified one compound with
similar efficacy to pregabalin in an in vivo model of pain and
one compound with superior efficacy to pregabalin in a model
of anxiety.
Acknowledgements
*
**
Figure 1. = P <0.05, = P <0.01, significantly different compared to vehicle.
We would like to thank Jane McGuffog for GC/HPLC analysis
(including enantiopurity analysis) and Giles Ratcliffe for NMR stud-
ies (including diastereoisomer analysis).
hyperalgesia had fully formed the time to paw withdrawal had
reduced to 2.5 s. At this point the compound to be assessed was
dosed orally and any reversal of the hyperalgesia was assessed
by measuring the time to paw withdrawal.
References and notes
Compound (22) showed good levels of efficacy in this model
with a 30 mg/kg dose reversing the hyperalgesia and giving a sim-
ilar effect to that of pregabalin. Compound (16) showed no effect in
this model on oral dosing but did reverse the hyperalgesia on intra-
thecal dosing (direct into the spinal cord). This data suggested that
compound (16) did not penetrate the blood–brain barrier. Gaba-
pentin itself is a substrate for the Large amino acid transporter
1. Bryans, J. S.; Wustrow, D. J. Med. Res. Rev. 1999, 19, 149.
2. Pan, H. L.; Eisenach, J. C.; Chen, S. R. J. Pharmacol. Exp. Ther. 1999, 288, 1026.
3. Hunter, J. C.; Gogas, K. R.; Hedley, L. R.; Jacobsen, L. O.; Kassotakis, L.;
Thompson, J.; Fontana, D. J. Eur. J. Pharmacol. 1997, 324, 153.
4. Singh, L.; Field, M. J.; Ferris, P.; Hunter, J. C.; Oles, R. J.; Williams, R. G.;
Woodruff, G. N. Psychopharmacology 1997, 127, 1.
5. de-Paris, F.; Busnello, J. V.; Vianna, M. R. M.; Salgueiro, J. B.; Quevedo, J.;
Izquierdo, I.; Kapczinski, F. Behav. Pharmacol. 2000, 11, 169.
6. Rice, A. S. C.; Maton, S. Pain 2001, 94, 215.
7. Dworkin, R. H.; Kirkpatrick, P. Nat. Rev. Drug Disc. 2005, 4, 455.
8. Zareba, G. Drugs Today 2005, 41, 509.
9. Guay, D. R. P. Am. J. Geriatr. Pharmacother. 2005, 3, 274.
10. Gee, N. S.; Brown, J. P.; Dissanayake, V. U. K.; Offord, J.; Thurlow, R.; Woodruff,
G. N. J .Biol. Chem. 1996, 271, 5768.
11. Stahl, S. M. J. Clin. Psychiatry 2004, 65, 1033. and references cited therein.
12. Field, M. J.; Li, Z.; Schwarz, J. B. J. Med. Chem. 2007, 50, 2569. and references
cited therein.
13. Carlsen, Per H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J. Org. Chem. 1981, 46,
3936.
14. Deardorff, D. R.; Myles, D. C. Org. Synth. 1989, 67, 114.
15. Deardorff, D. R.; Windham, C. Q.; Craney, C. L. Org. Synth. 1995, 73, 25.
16. Johnson, C. R.; Bis, S. J. Tetrahedron Lett. 1992, 33, 7287.
(System-L
amino acid transporter)11 and it has been assumed that
it is this transporter that conveys these highly polar compounds
into the CNS. It is possible that different diastereoisomers have dif-
ferent affinities for the transporter and this could explain the dif-
ference seen between the two compounds in the CITH model.
However, this hypothesis was undermined by the data for com-
pound (16) from the water-lick (Vogel) conflict model of anxiety
(Fig. 2). In this model, a group of rats were dosed orally with com-
pound and were then placed in a cage containing a drinking tube
from which they received a shock each time they drank. The mean
number of shock episodes received per rat over 10 min was mea-
17.
Sato, M.; Hattori, H.; Murakami, M.; Kaneko, C. Chem. Lett. 1993, 1919.