2,5-Dihydropyrazolo[4,3-c]pyridin-3-ones: Functionally selective benzodiazepine binding site ligands on the GABAA receptor

Article abstract of DOI:10.1016/j.bmcl.2004.04.085

2,5-Dihydropyrazolo[4,3-c]pyridin-3-ones are GABA_A receptor benzodiazepine binding site ligands with functional selectivity for the α3 subtype over the α1 subtype. SAR studies to optimise this functional selectivity are described.

Full text of DOI:10.1016/j.bmcl.2004.04.085

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3441–3444
2,5-Dihydropyrazolo[4,3-c]pyridin-3-ones: functionally selective
benzodiazepine binding site ligands on the GABAA receptor
Andrew Mitchinson,* John R. Atack, Peter Blurton, Robert W. Carling,
ꢀ
Jose L. Castro, Karen S. Curley, Michael G. N. Russell, George Marshall,
Ruth M. McKernan, Kevin W. Moore, Robert Narquizian, Alison Smith,
Leslie J. Street, Sally-Anne Thompson and Keith Wafford
Merck Sharp & Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Eastwick Road, Harlow CM20 2QR, UK
Received 26 December 2003; revised 8 April 2004; accepted 22 April 2004
Available online
Abstract—2,5-Dihydropyrazolo[4,3-c]pyridin-3-ones are GABAA receptor benzodiazepine binding site ligands with functional
selectivity for the a3 subtype over the a1 subtype. SAR studies to optimise this functional selectivity are described.
Ó 2004 Published by Elsevier Ltd.
The major inhibitory neurotransmitter in the central
nervous system is c-aminobutyric acid (GABA). GABA
receptors are subdivided into three pharmacological
lytics has been spurred on by the recent identification of
subtypes of GABAA receptors (a1–6, b1–3, c1–3, d, e, p
and h). The BZ site is formed in receptors containing b,
c2 and either a1, a2, a3 or a5 subunits. It is thought that
subtype selective ligands will discriminate between the
behaviours mediated by GABAA channels. Studies in
transgenic mice have demonstrated that sedation is
mediated through the a1-containing GABAA recep-
tors.3;4 Our laboratory has reported the discovery of
triazolo[4,3-b]pyridazines as functionally selective a2/3
partial agonists,5 which act as non-sedating anxiolytics
in animal models.6 In this communication, we report
that 2,5-dihydropyrazolo[4,3-c]pyridin-3-ones are also
functionally selective a2/3 ligands, and describe the
medicinal chemistry that led to their optimisation.
1;2
classes of which GABAA
and GABAC are ligand
gated ion channels, whilst GABAB receptors are coupled
to G-proteins. The binding of GABA to GABAA ion
channels causes them to open, allowing chloride flow
into the neuron, which hyperpolarises the membrane
and inhibits further neuronal activity.
Allosteric modulatory sites on GABAA receptors exist
for various ligands, including benzodiazepines (BZs),
barbiturates, ethanol, avermectins, picrotoxin, zinc ca-
tions and loreclezole, of which the BZ site has attracted
most attention. Positive allosteric modulators (agonists)
at the BZ site enhance GABAergic inhibition, whilst
negative allosteric modulators (inverse agonists) reduce
inhibition. Antagonists at the BZ site are also known,
which elicit no functional response.
In 1987 researchers from Ciba–Geigy described CGS
17867A, a tetrahydropyrazoloquinoline, which was a
potent inhibitor of [3H]flunitrazepam binding to the
central BZ site (IC50 ¼ 1.0 nM).7 In animal models, CGS
17867A showed anxiolytic activity, but without many of
the limiting effects observed for classical BZs. We deci-
ded to screen this compound for in vitro activity against
the a1, a2 and a3 GABAA subtypes. Affinity was mea-
sured at cloned human a1b3c2, a2b3c2 and a3b3c2
GABAA receptors stably expressed in L(tkꢀ) cells by
displacement of [3H]Ro15-1788 binding.8 Efficacy was
determined at the same receptors transiently expressed
in Xenopus oocytes by measurement of the modulatory
Full agonists acting at the BZ site have long been used
as anxiolytics, but exhibit side effects such as cognitive
impairment, sedation, ataxia and potentiation of etha-
nol. Chronic use can result in tolerance, dependence and
withdrawal symptoms. The search for improved anxio-
* Corresponding author. Fax: +44-1279-440390; e-mail: andrew_
mitchinson@merck.com
0960-894X/$ - see front matter Ó 2004 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2004.04.085

3442
A. Mitchinson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3441–3444
Table 1. In vitro data for CGS 17867A
GABAA subtype Ki (nM)a
Oocyte efficacy Patch-clamp
(%)b
efficacy (%)a
a1
a2
a3
0.19 ( 0.07)
0.77 ( 0.28)
1.03 ( 0.28)
12 ( 4) 30
61 ( 9) 100
93 ( 10) 100
17 ( 2)
43 ( 14)
70 ( 3)
a Values are means ( SEM) for n ¼ 3–7.
b Values are means ( SEM) for n ¼ 3–7, concentration of test com-
pound (nM).
effect on the GABA EC20 ion current using two elec-
trode voltage-clamp electrophysiology;9 these experi-
ments were performed using a single concentration of
the test compound. Efficacy was also measured by
whole-cell patch-clamp recording from L(tkꢀ) cells
expressing the receptors as above;10 full concentration-
effect curves were generated using this method. The re-
sults are shown in Table 1.
Scheme 1. (i) NH3, EtOH, 100%; (ii) ketone, cat. 4-toluenesulfonic
acid, toluene, reflux; (iii) Dowtherm A, reflux; (iv) POCl3, reflux; (v) 4-
R–C6H4NHNH2, n-BuOH, reflux.
zolo[4,3-c]pyridin-3-one core. The general chemistry
employed is shown in Scheme 1. Diethyl ethoxymeth-
ylenemalonate 2 was converted to its amino analogue 3
by treatment with ethanolic ammonia, and then 3 was
condensed with ketones under Dean–Stark conditions to
afford enamines 4. Pyrolysis of 4 yielded the pyridones 5,
which were converted into 4-chloropyridines 6 with
phosphorus oxychloride. The target compounds 7–18
were then prepared by condensation of intermediates 6
with various aryl hydrazines.
As expected, CGS 17867A had high affinity for all the
subtypes measured, but most intriguingly it was found
to have functional selectivity, with significant efficacy at
a3 and a2, but much reduced efficacy at a1. The reduced
sedation observed in animals dosed with CGS 17867A is
attributable with hindsight to the low efficacy of this
compound at the a1 subtype, with the anxiolytic
behaviour being mediated through a2/a3. 7,8,9,10-Tetra-
hydroimidazo[1,2-c]pyrido[3,4-e]-pyrimidin-5(6H)-ones 1
have recently been reported as functionally selective
ligands for the GABAA receptor BZ site,11 but this is the
first time that the structurally related 2,5-dihydropy-
razolo[4,3-c]pyridin-3-ones have been disclosed as hav-
ing a similar profile.
For the compounds incorporating a fused piperidine
ring, Scheme 2 shows how the piperidine nitrogen was
elaborated. The ethoxycarbonyl protected compound 18
was converted to the secondary amine 19 by treatment
with hydrogen bromide in acetic acid. This was then
€
alkylated with alkyl halides in the presence of Hunig’s
Using CGS 17867A as a lead, we pursued a medicinal
chemistry programme to optimise the series for a3
functional selectivity. The desired efficacy profile was to
have partial to full agonism at a3, but antagonism at a1.
In vitro screens were set up as described above; oocyte
data was used as the primary efficacy screen, with whole-
cell patch-clamp recording used as a more robust sec-
ondary assay for selected compounds.
Studies concentrated initially on varying the chlorine
substituent on the phenyl ring, and on making changes
to the saturated ring fused to the 2,5-dihydropyra-
Scheme 2. (i) HBr–AcOH, 80%; (ii) RX (X ¼ Cl, Br), DIPEA, DMF;
(iii) 3-pyridine acetic acid, WSCDI, HOBt, DIPEA, DCM, 47%.

A. Mitchinson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3441–3444
3443
base, affording compounds 20 and 21. The amide 22 was
prepared by carbodiimide-mediated coupling with 3-
pyridine acetic acid.
methoxy group to give difluoro- and trifluoromethoxy
substituents gave no obvious trend in the a3 oocyte
data, but the difluorinated analogue 10 was found to be
an antagonist at a1 when measured in the patch-clamp
assay. More compounds were prepared incorporating
the difluoromethoxy group for this reason (see Table 3).
Incorporating a cyclobutyloxy group (11) significantly
increased the efficacy at a1, an effect possibly related to
the size and lipophilicity of the substituent. Adding the
smaller cyclopropyloxy group gave the best compound
in this series (12) other than CGS 17867A, with the
highest a3 efficacy recorded in both functional assays,
Table 2 shows the effect on affinity and efficacy of
varying the 4-substituent on the phenyl ring. In general,
affinity for the two subtypes was not greatly affected by
changes at this position. Disappointingly, none of the
substituents gave a3 efficacy as high as that of CGS
17867A. A fluorine substituent (7) reduced the a3 effi-
cacy to 33%, although complete antagonism at a1 was
observed in the oocytes. Increasing fluorination of a
Table 2. Affinity and efficacy at a1 and a3 GABAA BZ receptor subtypes for 2,5-dihydropyrazolo[4,3-c]pyridin-3-ones varying the 4-phenyl sub-
stituent
Compds
R
X
Ki a1/nMa
Ki a3/nMa
Oocyte efficacy Oocyte efficacy Patch-clamp
a1 (%)b
Patch-clamp
efficacy a1 (%)a efficacy a3 (%)a
a3 (%)b
CGS 17867A
7
Cl
F
CH
CH
CH
CH
CH
CH
CH
N
0.19 ( 0.07)
0.29 ( 0.06)
0.31 ( 0.07)
7.33 ( 3.13)
1.00 ( 0.00)
1.30 ( 0.10)
1.13 ( 0.46)
0.57 ( 0.03)
1.03 ( 0.28)
0.65 ( 0.13)
0.61 ( 0.12)
4.70 ( 1.40)
2.25 ( 0.65)
2.55 ( 0.20)
2.19 ( 0.57)
1.14 ( 0.14)
12 ( 4) 30
)2 ( 4) 30
15 ( 4) 30
93 ( 10) 100
33 ( 5) 30
55 ( 6) 30
17 ( 2)
––
––
70 ( 3)
––
––
812
OMe
OCF3
OCHF2
O–cyBu
O–cyPr
H
9
14 ( 20) 300
21 ( 4) 100
49 ( 9) 100
19 ( 2) 100
)39 ( 3) 100
35 ( 2) 300
56 ( 7) 300
36 ( 10) 300
62 ( 9) 100
)7 ( 2) 100
––
2.0 ( 2)
––
––
34 ( 6)
––
10
11
12
13
)4.3 ( 4)
––
54 ( 4)
––
a Values are means ( SEM) for n ¼ 3–7.
b Values are means ( SEM) for n ¼ 3–7, concentration of test compound (nM).
Table 3. Affinity and efficacy at a1 and a3 GABAA BZ receptor subtypes for 2,5-dihydropyrazolo[4,3-c]pyridin-3-ones varying the saturated ring
Compds
R
X
Ki a1/nMa
Ki a3/nMa
Oocyte efficacy a1(%)b
Oocyte efficacy a3(%)b
CGS 17867A
Cl
Cl
CH2
(CH2)0
(CH2)2
O
0.19 ( 0.07)
0.44 ( 0.01)
1.06 ( 0.41)
1.55 ( 0.05)
0.38 ( 0.09)
0.45 ( 0.00)
1.03 ( 0.28)
1.12 ( 0.36)
4.27 ( 1.62)
6.40 ( 1.20)
0.89 ( 0.21)
7.45 ( 1.39)
12 ( 4) 30
19 ( 1) 10,000
23 ( 3) 100
11 ( 2) 300
9 ( 2) 30
93 ( 10) 100
51 ( 3) 3000
87 ( 7) 300
14
15
16
17
20
Cl
OCHF2
OCHF2
Cl
32 ( 3) 1000
25 ( 5) 100
123 ( 4) 1000
S
111 ( 14) 30
21
22
Cl
Cl
0.18 ( 0.04)
3.15 ( 0.15)
0.56 ( 0.05)
5.40 ( 0.20)
131 ( 5) 100
49 ( 3) 300
199 ( 21) 100
2.3 ( 9) 1000
a Values are means ( SEM) for n ¼ 3–7.
b Values are means ( SEM) for n ¼ 3–7, concentration of test compound (nM).

3444
A. Mitchinson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3441–3444
combined with complete antagonism as measured in the
patch-clamp assay. Finally, introduction of a 2-pyridyl
group (13) led to a dramatic reduction in efficacy at both
subtypes, resulting in inverse agonism.
dihydropyrazolo[4,3-c]pyridin-3-ones has been opti-
mised for functional selectivity, leading to the synthesis
of 12. This compound has an excellent in vitro profile,
combining good affinity at GABAA receptor subtypes
with good partial agonism at a3 and complete antago-
nism at a1. The estimated receptor occupancy of 12 in
the mouse is comparable to that of CGS 17867A, whilst
its pharmacokinetic profile in the rat is superior. How-
ever, the low receptor occupancy of 12 in the rat pre-
cluded its further in vivo characterisation.
Table 3 shows the SAR for compounds with changes
made to the saturated ring of CGS 17867A. Contraction
to a five-membered ring (14) afforded a high affinity
ligand for both subtypes, but exhibited reduced efficacy
at a3. Expansion to a seven-membered ring (15) signif-
icantly increased the efficacy at a1 compared to CGS
17867A. Incorporating an oxygen (16) or sulfur atom
(17) into the saturated ring led to much reduced efficacy
at a3, thus removing most of the functional selectivity.
Good affinity was maintained when amines were intro-
duced to the ring (20, 21), indicating a potentially useful
space that could be further explored in this region of the
molecule. However, functional selectivity had been lost
for these compounds, with full agonism observed at
both subtypes. Incorporating an amide (22) gave
acceptable affinity, but the efficacy profile was now the
reverse of that required, with agonism at a1 and
antagonism at a3.
Acknowledgements
We thank Samantha Clark, Fahd Choudhry and Ben
Satelle for additional oocyte efficacy measurements, and
Desmond O’Connor for pharmacokinetic data.
References and notes
1. Stephenson, F. A. Biochem. J. 1995, 310, 1.
2. Mehta, A. K.; Ticku, M. K. Brain Res. Rev. 1999, 29, 196.
€
3. Rudolph, U.; Crestani, F.; Benke, D.; Brunig, I.; Benson,
J. A.; Fritschy, J.-M.; Martin, J. R.; Bluethmann, H.;
From the in vitro data, compound 12 looked to have the
best efficacy profile in the patch-clamp assay, combining
antagonism at a1 with partial agonism at a3. The
compound was therefore selected for further profiling
in vivo. Compound 12 was found to have excellent
bioavailability of 91% in the rat, compared to only 21%
(in our hands) for CGS 17867A. The improved bio-
availability of our compound may be attributed to its
reduced clearance (19 mL/min/kg for 12 compared to
30 mL/min/kg for CGS 17867A). Occupancy at GABAA
receptors in the mouse brain was determined by inhi-
bition of in vivo [3H]Ro15-1788 binding.13 Dosed i.p. at
10 mg/kg in 70% aqueous PEG, the occupancy for 12
after 30 min was 87% in the spinal cord (rich in a2/3
subtypes) and 96% in the cerebellum (rich in a1 sub-
types). This compared favourably with CGS 17867A,
for which the occupancy values were 81% and 95% in
the spinal cord and cerebellum, respectively, when dosed
as above. However, in the rat, occupancy for 12 deter-
mined 30 min after oral dosing at 10 mg/kg was only
18% (measured for the whole brain), which was con-
sidered insufficient to elicit an anxiolytic effect in vivo.
For this reason, the compound was not progressed
further into behavioural studies.
€
Mohler, H. Nature 1999, 401, 796.
4. Sieghart, W. Trends Pharmacol. Sci. 2000, 21, 411.
5. Broughton, H. B.; Carling, R. W.; Castro Pineiro, J. L.;
Guiblin, A. R.; Madin, A.; Moore, K. W.; Russell, M. G.;
Street, L. J.; Patent Application WO 98/04559.
6. McKernan, R. M.; Rosahl, T. W.; Reynolds, D. S.; Sur,
C.; Wafford, K. A.; Atack, J. R.; Farrar, S.; Myers, J.;
Cook, G.; Ferris, P.; Garrett, L.; Bristow, L.; Marshall,
G.; Macauley, A.; Brown, N.; Howell, O.; Moore, K. W.;
Carling, R. W.; Street, L. J.; Castro, J. L.; Ragan, C. I.;
Dawson, G. R.; Whiting, P. J. Nat. Neurosci. 2000, 3, 587.
7. Bennet, D. A.; Amrick, C. L.; Wilson, D. E.; Boast, C. A.;
Loo, P.; Bernard, P. S.; Schmutz, M.; Gerhardt, S. C.;
Braunwalder, A.; Klebs, K.; Yokoyama, N.; Liebman, J.
M. Drug Dev. Res. 1987, 11, 219.
8. Hadingham, K. L.; Wingrove, P.; Le Bourdelles, B.;
Palmer, K. J.; Ragan, C. I.; Whiting, P. J. Mol. Pharma-
col. 1993, 43, 970.
9. Wafford, K. A.; Whiting, P. J.; Kemp, J. A. Mol.
Pharmacol. 1993, 43, 240.
10. Brown, N.; Kerby, J.; Bonnert, T. P.; Whiting, P. J.;
Wafford, K. A. Br. J. Pharmacol. 2002, 136, 965.
11. Albaugh, P. A.; Marshall, L.; Gregory, J.; White, G.;
Hutchison, A.; Ross, P. C.; Gallagher, D. W.; Tallman,
J. F.; Crago, M.; Cassella, J. V. J. Med. Chem. 2002, 45,
5043.
12. Yokoyama, N. Patent application EP 168350.
13. Atack, J. R.; Smith, A. J.; Emms, F.; McKernan, R. M.
Neuropsychopharmacology 1999, 20, 255.
In conclusion, we have determined that CGS 17867A is
a functionally selective BZ binding site ligand on the
GABAA receptor, and this may explain its atypical
behavioural profile in vivo compared to classical BZs.
Taking CGS 17867A as a starting point, a series of 2,5-