Synthesis and in vivo evaluation of bicyclic gababutins

Article abstract of DOI:10.1016/j.bmcl.2009.11.118

Synthesis of a number of bicyclic five-membered ring derivatives of gabapentin led to the identification of two compounds, (-)-(11A) and (20A) which both had an excellent level of potency against α₂δ and were profiled in an in vivo model of neu

Full text of DOI:10.1016/j.bmcl.2009.11.118

Bioorganic & Medicinal Chemistry Letters 20 (2010) 461–464
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and in vivo evaluation of bicyclic gababutins
*
David C. Blakemore , Justin S. Bryans, Pauline Carnell, Christopher L. Carr, Nicola E. A. Chessum,
Mark J. Field, Natasha Kinsella, Simon A. Osborne, Andrew N. Warren, Sophie C. Williams
Sandwich Laboratories, Pfizer Global Research and Development, Ramsgate Road, 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:
Synthesis of a number of bicyclic five-membered ring derivatives of gabapentin led to the identification of
Received 15 November 2009
Revised 22 November 2009
Accepted 24 November 2009
Available online 27 November 2009
two compounds, (À)-(11A) and (20A) which both had an excellent level of potency against
a2d and were
profiled in an in vivo model of neuropathic pain.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Gabapentin
Gababutin
Pregabalin
Pain
Gabapentin (NeurontinÒ) (1)1 and pregabalin (LyricaÒ) (2)2,3
have been shown preclinically to have utility against neuropathic
pain and anxiety; their efficacy against neuropathic pain has also
been demonstrated clinically in humans.4 Pregabalin has been ap-
proved for the management of neuropathic pain associated with
diabetic peripheral neuropathy and post-herpetic neuralgia and
achieved world-wide sales of over one billion dollars in 2007.
Gabapentin and pregabalin are thought to mediate their pharma-
recently reported our investigations into appending alkyl substit-
uents onto the 3-position which demonstrated that methyl, ethyl
and n-propyl groups enhanced binding affinity, when the stereo-
centre at the 3-position had the (R)-configuration. 3,4-Disubsti-
tution with methyl groups also led to enhanced binding
affinity but is again highly dependent on the stereochemistry
of the 3-position. Compounds (4) and (5) have greater binding
affinities than gabapentin but are indistinguishable from it in
the CCI (chronic constriction injury) weight-bearing model of
neuropathic pain.
cological actions through binding to the
a2d subunit of a voltage
gated calcium channel.5,6
We, and others, have disclosed SAR investigations around a range
2d ligands.7 Recently, our attention has focused on the five-mem-
bered ring analogue (3) which we have called gababutin.8,9
CO2H
CO2H
H2N
H2N
of
a
HO2C
NH2
HO2C
NH2 HO2C
NH2
(4)
IC50 = 107 nM
(5)
IC50 = 22nM
We now report the effect on binding and efficacy of building
conformational constraint into these molecules. We have synthes-
ised a number of bicyclic analogues of (4) where the ring junction
is cis and analogues of (5) where a trans ring junction is present.
With bicyclic gababutin analogues of (4) and (5), if the ring
junction is trans, then the stereochemistry of the quaternary centre
carrying the aminomethyl group is irrelevant as this centre is pseu-
do-asymmetric; the key problem is synthesis of the C2 symmetric
chiral ketone precursor. If the ring junction is cis, then the starting
ketone is achiral, or meso (as is the final amino acid), but introduc-
tion of the stereogenic quaternary centre carrying the amino-
methyl group must be controlled (by making use of the differing
(1)
Gabapentin
(2)
Pregabalin
(3)
Gababutin
Gababutin has a binding affinity of 420 nM at the
a2d binding
site compared to 140 nM for gabapentin. This observation sug-
gested to us that the gababutin five-membered ring is not opti-
mally filling the available space in the binding pocket. We have
* Corresponding author. Tel.: +44 1304 644573.
E-mail address: david.blakemore@pfizer.com (D.C. Blakemore).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.11.118

462
D. C. Blakemore et al. / Bioorg. Med. Chem. Lett. 20 (2010) 461–464
levels of steric hindrance on attack of the exo and endo faces of the
bicyclic system).
amethyldisilazide and this was quenched at low temperature with
dimethylallyl bromide to give cyanoalkene (17). Attack occurred
on the less hindered exo side of the bicyclic system (as confirmed
by NOE studies) to give (17) as a single diastereoisomer (99% dia-
stereoisomeric excess as confirmed by HPLC). Ozonolysis of alkene
(17) in the presence of sodium hydroxide and methanol allowed
the generation of cyanoester (18). Reduction of the cyanide group
of (18) yielded the lactam (19) which was hydrolysed to the amino
acid (20B) with hydrochloric acid.
CO2H
O
Control of key
quaternary centre
unnecessary
NH2
(ketone is C2 symmetric)
Single enantiomer
O
Analogue (20A), where the aminomethyl group is on the less hin-
dered exo face of the bicyclic system, can be formed via an interme-
diate ketone and this is illustrated in Scheme 3. Commercially
available anhydride (21) was reduced to cis-diol (22). Following its
conversion to di-iodide (24), ring-closure to generate the cis-cyclo-
butyl ring was carried out using tert-butyllithium (or, alternatively,
with n-butyllithium). Oxidative cleavage of the cyclohexene ring
gave a diacid which was converted to diester (26). Dieckmann cycli-
sation followed by decarboxylation gave the desired meso ketone
(27). This could be converted to amino acid (20A) using the nitro-
methane addition method used in Scheme 1 (although the nitro-
methane addition gave 10–15% of the undesired diastereoisomeric
product) or via the allyl addition route shown below. Knoevenagel
condensation on ketone (27) gave unsaturated cyanoester (28).
Allylmagnesium bromide added to the unsaturated cyanoester in a
conjugate manner with addition occurring exclusively from the less
hindered exo face of the molecule (a diastereoisomeric excess of 99%
was determined by HPLC). Hydrolysis of the addition product gave
acid (29) as a single diastereoisomer. Following conversion of acid
(29) to ester (30), oxidative cleavage of the allyl group gave acid
(31) which was Curtius rearranged and hydrolysed to give desired
amino acid (20A) as a single diastereoisomer (NOE studies con-
firmed that the aminomethyl group was on the exo face of the bicy-
clic system).
CO2H
Control of key
quaternary centre
provided by steric
impact of second ring
NH2
Meso ketone
Bicyclic analogues without a symmetry element required both
an enantiopure ketone precursor and control of the stereogenic
quaternary centre. This is illustrated in the synthesis of (À)-(11A)
from the known chiral ketone (7) (Scheme 1).
Enantiopure bicyclo[3.2.0]hept-2-en-6-one10 (6) was hydroge-
nated to give ketone (7) which was transformed in good yield to
the a,b-unsaturated ester (8) via a Horner–Emmons reaction. Con-
jugate addition of the anion of nitromethane to the unsaturated
ester (8) resulted in the formation of the nitroester (9). The nitro-
methane anion attacks the bicyclic system exclusively from the
less hindered exo face (as confirmed by NOE analysis) to give (9)
as a single diastereoisomer. Reduction of (9) by hydrogenation in
the presence of a nickel catalyst gave lactam (10) as a result of
the in situ cyclisation of the intermediate amino ester generated.
Hydrolysis of lactam (10) with 6 N HCl produced the target amino
acid (À)-(11A) as a hydrochloride salt (chiral HPLC showed the
enantiomeric excess to be 98%).
The diastereoisomeric bicyclic cis analogues (20A) and (20B)
both possess a plane of symmetry and so are achiral compounds.
Analogue (20B), where the aminomethyl group is on the more hin-
dered endo face of the bicyclic system, can be formed via an inter-
mediate cyanide and this is illustrated in Scheme 2.
Reduction of the commercially available cyclobutane-1,2-dicar-
boxylic acid (12) gave a diol which was converted to the bis-mes-
ylate (13). Displacement of the mesyl groups with lithium bromide
in acetone occurred in high yield to give dibromide (14). Ring clo-
sure of dibromide (14) with ethyl cyanoacetate gave cyanoester
(15) as a mixture of diastereoisomers. Krapcho reaction of the
cyanoester with lithium chloride, water and dimethylsulfoxide
yielded cyanide (16) via hydrolysis and decarboxylation. The anion
of cyanide (16) was generated by deprotonation with lithium hex-
In the case of the trans-bicyclic analogues, the target molecules
were made as racemates from the precursor ketone using the
nitromethane addition route.
It has been a recurring theme with the gababutins that small
changes in structure can dramatically change binding affinity at
the receptor and that the receptor is very sensitive to chirality.
Binding affinity data for the trans-bicyclic gababutins are shown
in Table 1 (a radioligand binding assay incorporating [3H]gabapen-
tin at the
a2d subunit of a calcium channel was utilised as previ-
ously described5).
CO2H
OMs
OMs
Br
Br
a
c
b
CO2H
(12)
(13)
(14)
a
b
d
(-)
(-)
(6)
(7)
(8)
O
O
CN
CN
f
e
EtO2C
c
CN
CO2Et
(17)
(16)
(15)
g
H
e
d
CN
h
N
i
NH2
CO2Me
O
CO2H
NO2
CO2Et
(-)-(11A)
(10)
(9)
NH2
CO2H
N
(18)
(19)
(20B)
O
Scheme 2. Reagents and conditions: (a) LiAlH4, Et2O, 0 °C to rt (98%); (b) MsCl,
NEt3, DCM, À40 °C to rt (73%); (c) LiBr, acetone, reflux (91%); (d) NCCH2CO2Et,
K2CO3, DMF (99%); (e) LiCl, H2O, DMSO, 150 °C (44%); (f) LHMDS, THF, À40 °C;
dimethylallylbromide, À78 °C (72%); (g) O3, NaOH, MeOH, DCM (71%); (h) H2, Ni,
MeOH; (i) 6 N HCl, 1,4-dioxane (80% from cyanoester).
Scheme 1. Reagents and conditions: (a) H2, Pd/C, EtOAc (b) (EtO)2P(O)CH2CO2Et,
NaH, THF (72% over 2 steps); (c) MeNO2, Bu4N+FÀ, THF, 70 °C (87%); (d) Ni sponge,
H2, MeOH (54%); (e) 6 N HCl (72%).

D. C. Blakemore et al. / Bioorg. Med. Chem. Lett. 20 (2010) 461–464
463
Table 2
O
a
b
OH
OH
OMs
OMs
O
CO2H
NH2
NH2 CO2H
(21)
(22)
(25)
(23)
O
c
( )n
( )n
B
e, f
d
MeO2C
MeO2C
(26)
I
I
A
(24)
n
Compound
IC50 (nM)
0
0
1
1
2
2
(20A)
(20B)
(35A)
(35B)
(36A)
(36B)
38
332
2600
>10,000
720
g, h
CN
i
j, k
O
CO2Et
CO2H
>10,000
(27)
(28)
(29)
l
(11A) has good binding affinity at the receptor. Again, increasing
the size of either ring of the bicyclic system reduced binding affin-
ity (Table 3) reinforcing our belief that the bicyclo[3.2.0]heptyl sys-
tem was optimally sized.
n, o
m
NH2
CO2H
CO2H
CO2Me
CO2Me
(20A)
(31)
(30)
Scheme 3. Reagents and conditions: (a) LiAlH4, THF, reflux (80%); (b) MsCl, NEt3,
DCM, À40 °C to rt (80%); (c) NaI, acetone, reflux (70%); (d) t-BuLi, pentane–ether
(3:2), À25 °C; (e) NaIO4, RuCl3ÁH2O, MeCN, EtOAc, H2O; (f) MeOH, concd H2SO4 (85%
from di-iodide); (g) KOt-Bu, THF, reflux (95%); (h) DMSO, H2O, 155 °C (97%); (i)
ethyl cyanoacetate, NH4OAc, AcOH, toluene, reflux (78%); (j) allylmagnesium
bromide, dimethylzinc, THF, À78 °C; (k) KOH, ethylene glycol, 165 °C (89% over the
two steps); (l) TMSCHN2, toluene (97% from cyanoester); (m) RuCl3, NaIO4, CCl4,
MeCN, H2O (83%); (n) DPPA, NEt3, toluene, reflux; (o) 6 N HCl, reflux (72% over the
two steps).
Table 3
( )m
( )n
( )m
( )n
CO2H
NH2
CO2H
NH2
(+/-)
(+/-)
The effect of constraint is to dramatically reduce binding affin-
ity at a2d as compared to the trans-dimethyl gababutin (5). While
A
B
this can be rationalised for compounds (33) and (34) where the
extra methylene groups might make the molecules too large to
fit in the receptor it is surprising that the direct bicyclic analogue
of (5), compound (32), has poor binding affinity.
The tight SAR around these bicyclic gababutins is further illus-
trated with the cis compounds (Table 2). All the compounds with
the aminomethyl group on the more hindered endo face of the mol-
ecule have poor binding affinity. Of the compounds where the ami-
nomethyl is on the exo face of the molecule, only compound (20A)
m
n
Compound
IC50 (nM)
1
1
1
2
2
0
0
1
0
0
(11A)
(11B)
(37A)
(38A)
(38B)
25
2800
7181
600
3088
has high affinity for a2d.
Due to the limited scope for change in the size of the bicyclic
skeleton, we hypothesised that (20A) fitted the binding pocket
optimally. Keeping the bicyclic skeleton the same, we synthesised
regioisomers of (20A) with the amino acid portion of the molecule
directly attached to the four-membered ring; two diastereoisomers
(11A) and (11B) are possible, and we were pleased to find that
Table 1
NH2 CO2H
(+/-)
( )n
n
Compound
IC50 (nM)
0
1
2
(32)
(33)
(34)
2900
>10,000
>10,000
Figure 1. (* = P <0.05, *** = P <0.01, significantly different compared to vehicle.

464
D. C. Blakemore et al. / Bioorg. Med. Chem. Lett. 20 (2010) 461–464
Table 4
It is noteworthy that (+)-(11A) was completely inactive in this
model unless dosed intrathecally. It would, therefore, appear that
the compound is not penetrating the blood–brain barrier. Gabapen-
*
Compound
System L binding IC50 (lM)
tin itself is a substrate for the Large amino acid transporter (System-
L
(4)
7
amino acid transporter) and it has been assumed that it is this trans-
porter that conveys the compounds into the CNS.7 Interestingly,
while (20A) binds to the transporter, neither (+)-(11A) or (À)-
(11A) appear to have any affinity for it (Table 4). It is, however, pos-
sible that (À)-(11A) is entering the CNS via another mechanism.
(À)-(11A) and (20A) have very similar rat pharmacokinetic pro-
files with high bioavailabilities and a short half-lives (Table 5). As
expected for such polar molecules (log D7.4 = À1.0 in both cases),
the compounds are not metabolised but are renally cleared at glo-
merular filtration rate.
(5)
10
38
>10,000
6931
(20A)
(À)-(11A)
(+)-(11A)
Prevention of uptake of 3H-Leucine into CHO cells.11
*
Table 5
Compound
F (%)
Elimination half-life (h)
Clearance (mL/min/kg)
Compounds (À)-(11A) and (20A) have both been progressed for
(À)-(11A)
(20A)
100
84
2.3
2.4
4.5
4.3
further studies.
Acknowledgements
NH2 CO2H
Ti-Zhi Su is thanked for the System L binding data. Alison Betts
is thanked for the pharmacokinetic data. We thank the staff of the
Structural and Analytical Services Department, Sandwich, for ana-
lytical data.
CO2H
(+)
CO2H
(-)
NH2
NH2
References and notes
(20A)
IC50 = 38 nM
(+)-(11A)
IC50 = 19 nM
(-)-(11A)
IC50 = 46 nM
1. Bryans, J. S.; Wustrow, D. J. Med. Res. Rev. 1999, 19, 149.
2. Dworkin, R. H.; Kirkpatrick, P. Nat. Rev. Drug Disc. 2005, 4, 455.
3. Zareba, G. Drugs Today 2005, 41, 509.
4. Rice, A. S. C.; Maton, S. Pain 2001, 94, 215.
5. Gee, N. S.; Brown, J. P.; Dissanayake, V. U. K.; Offord, J.; Thurlow, R.; Woodruff,
G. N. J. Biol. Chem. 1996, 271, 5768.
6. Stahl, S. M. J. Clin. Psychiatry 2004, 65, 1033. and references cited therein.
7. Field, M. J.; Li, Z.; Schwarz, J. B. J. Med. Chem. 2007, 50, 2569. and references
cited therein.
8. Blakemore, D. C.; Bryans, J. S.; Carnell, P.; Chessum, N. E. A.; Field, M. J.; Kinsella,
N.; Kinsora, J. K.; Osborne, S. A.; Williams, S. C.; Bioorg. Med. Chem. Lett., 2009.
doi:10.1016/j.bmcl.2009.10.089.
9. Blakemore, D. C.; Bryans, J. S.; Carnell, P.; Field, M. J.; Kinsella, N.; Kinsora, J. K.;
Meltzer, L. T.; Osborne, S. A.; Thompson, L. R.; Williams, S. C.; Bioorg. Med. Chem.
Lett., 2009. doi:10.1016/j.bmcl.2009.10.121.
10. Klempier, N.; Geymayer, P.; Stadler, P.; Faber, K.; Griengl, H. Tetrahedron:
Asymmetry 1990, 1, 111.
Both enantiomers of (11A) have good binding affinities at
a2d
and these two compounds and (20A) were dosed orally in an
in vivo model of neuropathic pain, the CCI (chronic constriction in-
jury) weight-bearing model. In this model, loose ligatures are
placed around the sciatic nerve of a rat causing hypersensitivity
in one hind paw. The ability of the rat to place weight on each hind
paw is measured and the differential between ipsilateral (affected)
and contralateral (unaffected) paw is calculated.
All potent gababutins (IC50 values of less than 150 nM at
a2d)
tested in this model to date have achieved similar efficacy to gaba-
pentin. Both (20A) and (À)-(11A) proved to be more efficacious
than gabapentin in this model, with (À)-(11A) having the most
effective profile yet seen in this model giving a complete blockade
of the pain effect (Fig. 1).
11. Belliotti, T. R.; Caparis, T.; Ekhato, I. V.; Kinsora, J. J.; Field, M. J.; Heffner, T. J.;
Meltzer, L. T.; Schwarz, J. B.; Taylor, C. P.; Thorpe, A. J.; Vartanian, M. G.; Wise, L.
D.; Su, T.-Z.; Weber, M. L.; Wustrow, D. J. J. Med. Chem. 2005, 48, 2294.