A novel series of positive modulators of the AMPA receptor: Structure-based lead optimization

Article abstract of DOI:10.1016/j.bmcl.2010.08.063

Starting from lead compound 1, we demonstrate how X-ray structural data can be used to understand SAR and expediently optimize bioavailability in a novel series of AMPA receptor modulators, furnishing 5 with improved bioavailability and robust in vivo activity.

Full text of DOI:10.1016/j.bmcl.2010.08.063

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6072–6075
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
A novel series of positive modulators of the AMPA receptor: Structure-based
lead optimization
Craig Jamieson a, , Robert A. Campbell a, Iain A. Cumming a, Kevin J. Gillen a, Jonathan Gillespie a,
⇑
Bert Kazemier b, Michael Kiczun a, Yvonne Lamont a, Amanda J. Lyons a, John K. F. Maclean a,
Frederic Martin a, Elizabeth M. Moir a, John A. Morrow a, John Pantling a, Zoran Rankovic a, Lynn Smith a
a Merck Research Laboratories, MSD, Newhouse, Motherwell, Lanarkshire, ML1 5SH, UK
b Merck Research Laboratories, MSD, PO Box 20, Oss, 5340 BH, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
Starting from lead compound 1, we demonstrate how X-ray structural data can be used to understand
SAR and expediently optimize bioavailability in a novel series of AMPA receptor modulators, furnishing
5 with improved bioavailability and robust in vivo activity.
Received 29 June 2010
Revised 10 August 2010
Accepted 11 August 2010
Available online 15 August 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
AMPA receptor
Positive allosteric modulators
Ion channels
SBDD
In the preceding Letter,1 we illustrated how an HTS derived hit
could be rapidly optimized through application of biostructural
data to a lead compound (1), an AMPA receptor modulator worthy
of further optimization. Key data associated with compound 1 is
listed in Figure 1.
reasoned that a suitably sized solubilizing group could be ap-
pended either from the pyrazole region of the molecule or through
modification of the external amide. The strategy adopted is out-
lined in Figure 2.
We initially targeted the amide region as a means of rationaliz-
ing our template and sought to introduce an alcohol moiety to fur-
nish 2. Compared to the progenitor compound, 2 was equipotent
(pEC50 = 6.6) but lacked sufficient solubility (1 mg/L) and had unac-
ceptable microsomal stability (rat/human intrinsic clearance, Cli
During the hit-to-lead phase, the strategy adopted to improve
oral bioavailability was to introduce a conformational constraint
while maintaining acceptable levels of solubility. A hydrophilic
pocket on the receptor structure2 was targeted in order to meet
these two goals. However, as the summary data outlined in Figure
1 indicates, oral bioavailability remained sub-optimal despite the
compound having an acceptable half-life and CNS penetration.
Therefore, we sought an alternative approach to improving oral bio-
availability in the series. Additionally, we sought to further establish
SAR around the series, taking direction from our biostructural data.
As the mass of our lead compound (495 Da) was approaching
the upper limits of what is generally regarded as being useful for
an oral drug,3 we turned our attention to how we could produce
analogues of reduced complexity and molecular weight. A key
objective was to retain a high degree of solubility in these simpli-
fied analogues since we anticipated that this property would be of
benefit to bioavailability.4 From the available crystal structures, we
>270 lL/min/mg protein). Additionally, we attempted to introduce
amine functionality in this region, however, our attempts to do so
were hamstrung by the innate chemical instability of such species.
We next turned our attention to introduction of amine func-
tionality off the tetrahydroindazole species. Modeling based on
our crystal structures suggested that changes to the fused ring sys-
tem could be tolerated, and that main-chain atoms of Pro 494 and
Ser 729 provided potential interaction points. This was in contrast
to our previous experience1 with the tetrahydrobenzothiophene
moiety where preservation of the fused ring was essential to main-
tain modulatory activity at GluA1. As Table 1 illustrates, pendant
amines were generally well tolerated with acceptable measured ki-
netic solubilities in the range 20–100 mg/L.
Further profiling of selected analogues revealed that secondary
amines such as 4 and 5 displayed a good overall balance of po-
⇑
Corresponding author at present address: Dept. of Pure & Applied Chemistry,
University of Strathclyde, 295 Cathedral St, Glasgow G1 1XL, United Kingdom. Tel.:
+44 141 548 4830; fax: +44 141 548 5743.
tency, solubility and microsomal stability (e.g., 4 rat Cli <12
min/mg protein, human Cli = 33
ity = 65 mg/L; 5 rat Cli <19 L/min/mg protein, human Cli <12
l
L/
l
L/min/mg protein, solubil-
l
l
L/
E-mail address: craig.jamieson@strath.ac.uk (C. Jamieson).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.08.063

C. Jamieson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6072–6075
6073
min/mg protein, solubility = 107 mg/L). Within the series, tertiary
amine derivatives generally exhibited poor microsomal stability
(e.g. 9 rat/human Cli >270 lL/min/mg protein).
Having confirmed that introduction of polar functionality was
tolerated in the pyrazole region, we then sought to establish addi-
tional SAR in this area. From inspection of the crystal structure of 5
bound to the S1S2 LBD of GluA2 (Fig. 3), we concluded that the
hydrophobic interaction made by the trifluoromethyl group could
be approximated by other moieties such as tert-butyl. To this
end, we prepared a series of 3-t-butyl functionalized pyrazolyl
amines shown in Table 2.
Again further profiling of selected analogues indicated that com-
pounds such as 12 had a reasonable combined profile of GluA1 activ-
ity, solubility (79 mg/L) and microsomal stability (rat Cli = 15
lL/
min/mg protein; human Cli = 15 L/min/mg protein). The X-ray
l
structure of 12 in complex with the S1S2 ligand binding domain
confirmed the binding mode was as anticipated with the tert-butyl
system creating a hydrophobic interaction with the receptor
(Fig. 4).
Previously, we had not examined the role of the central methy-
lene spacer, and in particular, determinedtheeffect of extending this
linker system on potency at GluA1. Although relatively close hydro-
phobic contacts were made by the distal ends of our modulators
with the receptor, we hypothesized that this could be further opti-
mized through increasing the length of the central linker system. A
NH
GluR1 pEC50 = 6.6
Solubility = 20 mg/L
T1/2 = 3.8 h
F% = 3.8
B : P ratio = 1.03
HN
O
H
N
F
N
N
F
S
O
F
1
Figure 1. Lead compound 1, summary property data & structure in complex with
S1S2 LBD. As the binding site spans an intramolecular two-axis, two orientations of
1 are observed in the crystal structure, but only one (pyrazole moiety on LHS, as
shown in preceding Letter) is shown here for clarity.
HO
H
N
F
N
N
F
NH
S
O
F
2
HN
O
H
N
H2N
O
F
N
N
F
S
O
H
N
F
F
N
N
F
S
1
O
F
R2 3 - 11
N
Figure 3. X-ray structure of compound 5 in complex with the S1S2 J LBD of GluA2
(Alternate orientation to that in Figure 1, with pyrazole moiety on RHS). The
pendant amine function sits between Pro 494 from one monomer and Ser 729 from
another, and forms a hydrogen bond with the backbone carbonyl of Ser 729.
Residues 727–730 were omitted from the surface depiction for clarity. Here the
main-chain atoms of those residues are shown as sticks to illustrate the hydrogen
bond between 5 and Ser 729.
R1
Figure 2. Simplification of lead compound 1.
Table 1
Pyrazole substituted amines
F
H2N
S
F
Table 2
N
H
N
O
N
tert-Butylpyrazole substituted amines
F
O
O
NH2
H
N
N
R2
N
R1
N
S
O
R1
R2
pEC50
a
Compds
R2
N
3
4
5
6
7
8
9
10
11
H
H
H
Me
H
H
Et
H
H
Me
Et
Me
iPr
CH2CH2OMe
6.1
6.9
6.1
6.3
5.5
5.8
6.2
6.2
6.0
R1
Compds
R1
H
R2
pEC50
a
12
13
14
15
16
Me
Me
cPr
Et
6.5
6.5
5.8
6.1
5.8
Me
H
Et
cPr
H
H
–CH2CH2–CH2–
iPr
a
a
Values are means of two experiments performed in duplicate.
Values are means of two experiments performed in duplicate.

6074
C. Jamieson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6072–6075
HN
O
H2N
O
H
H
N
N
F
N
N
F
N
F
F
N
S
S
O
O
F
F
HN
N
H
20
21
Figure 5.
this parameter could enhance bioavailability, which led to the de-
sign of 21. Although 20 had acceptable potency (pEC50 = 6.0) it did
not show sufficient microsomal stability (rat Cli 33
tein, human Cli 57 L/min/mg protein) to merit further progres-
sion. While compound 21 had similar potency to both 5 and 20
(pEC50 = 5.8) and good microsomal stability (rat Cli 19 L/min/mg
protein, human Cli 16 L/min/mg protein), it displayed much high-
lL/min/mg pro-
l
Figure 4. X-ray structure of compound 12 in complex with the S1S2 J LBD of GluA2.
The orientation of 12 is as shown in Figure 3. In this case the pendant amine forms a
hydrogen bond with Pro 494.
l
l
er clearance in vivo (Clp = 43 mL/min/kg), therefore, it was not fur-
ther progressed.
Returning to compound 5, further in vitro profiling indicated
the compound was a selective AMPA positive modulator, with no
activity at kainate, NMDA or GABA. Wider selectivity determina-
tion5 indicated no off-target activity against a panel of receptors
and channels. Determination of brain to plasma ratio in Wistar
BRL rats gave a value of 0.05 which indicated modest exposure in
the central compartment. However, testing the compound in vivo
showed that 5 gave robust potentiation of AMPA evoked single unit
activity as measured by electrophysiology in the hippocampus of
anesthetized rats,6 with an MED of 0.3 mg/kg when dosed intrave-
nously. This mechanistic data gave us confidence that 5 was capa-
ble of modulating AMPA activity in vivo despite its apparently
modest brain penetration.
Syntheses of the compounds described above are outlined in
Schemes 1–3. Compound 2 was prepared in a 3 step manner from
the carboxylic acid 22 (prepared as described in the preceding pub-
lication1). Copper mediated decarboxylation gave the amide 23 fol-
lowed by Villsmier formylation and subsequent borohydride
reduction furnished 2.
Compounds 3–16 and compound were accessed as illustrated in
Scheme 2. Gewald cyclisation7 provided the aminothiophene
derivative 24 which could then be acetylated with using bromo-
acetyl bromide in the presence of base. Alkylation of the appropri-
ate pyrazole system with 25 followed by reductive amination gave
3–11 and 12–16, respectively. Compound 21 was accessed using
chemistry analogous to that depicted in Scheme 2 replacing cyano-
acetamide with N-methylcyanoacetamide in the Gewald step.
The homologated analogues 17–19 were prepared using similar
methodology shown also in Scheme 2.
Compound 20 was realized utilizing the synthetic sequence
shown in Scheme 3. The requisite pyrrolopiperidine system 29
was prepared according to literature methods.8 Starting from
Boc-piperidinone, enamine 28 was prepared. Subsequent reaction
with trifluoracetic anhydride and trapping of the resulting b-dike-
tone with hydrazine gave 29. Alkylation of 29 with intermediate 25
(Scheme 3) followed by acidolysis of the Boc group provided com-
pound 20.
selected number of analogues were prepared and evaluated against
GluA1 for their modulatory activity as summarized in Table 3.
The resulting activity data suggested the homologated species
were up to ten fold less potent than their progenitor compounds.
In addition, microsomal stability was eroded, presumably due to
the higher lipophilicity associated with the compounds (e.g. 17
rat Cli 49 lL/min/mg protein, human Cli 89 lL/min/mg protein).
From the amine derivatives prepared from 3 to 16, compounds
4, 5, and 12 were advanced to in vivo DMPK studies to assess oral
bioavailability. Table 4 shows a summary of the key properties for
each compound.
This subset of compounds showed encouraging oral bioavail-
ability compared to our initial lead 1, with 5 in particular being
most promising. The bioavailability data prompted us to prepare
a limited number of further analogues of 4 and 5 in order to deter-
mine if absorption could be further enhanced. The crystal structure
of compound 5 prompted us to consider conformational constraint
(20, Fig. 5) as we believed a heteroatom could successfully be
incorporated into the tetrahydroindazole system. Another option
explored was to reduce the H-bond donor count to determine if
Table 3
Homologated central linker analogues
F
NH2
F
O
F
N
H
N
N
S
N
O
R2
R1
a
Compds
R1
R2
pEC50
17
18
19
H
H
Et
Me
Et
Et
5.8
5.6
5.7
a
Values are means of two experiments performed in duplicate.
Table 4
In summary, this work has demonstrated how Structure-Based
Drug Design (SBDD) can be exploited within target families such as
ion channels where previously its application had been hindered.
Throughout our optimization trajectory, a central component has
been the application of SBDD in directing our template modifica-
tions as well as offering key insights into SAR within the series.
Starting from a lead compound 1 which exhibited poor oral
bioavailability, we have demonstrated how this can be optimized
to yield a rationalized entity (5) that has significantly enhanced
In vivo pharmacokinetic properties of selected compounds
a
a
a
Compds
T1/2 (h)
Vss (L/kg)
Clp (mL/min/kg)
F%b
1
4
5
3.8
1.5
1.1
1.9
0.5
2.4
0.6
1.1
2.3
23
15
9
4
15
43
16
12
a
2 mg/kg iv dose Wistar BRL rats.
10 mg/kg po dose Wistar BRL rats.
b

C. Jamieson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6072–6075
6075
O
OH
HO
HN
a
b, c
HN
O
HN
O
S
S
S
N
O
N
N
N
F
F
N
F
F
N
F
F
F
F
F
22
2
23
Scheme 1. Reagents and conditions: (a) Cu, quinoline 200 °C, lW, 95%; (b) POCl3, DMF, DCE, reflux, 86%; (c) NaBH4, THF, rt, 97%.
O
NH2
O
NH2
O
NH2
H
N
H
N
b, c
N
H2N
a
N
R
Br
S
O
S
S
R1
O
N
25
24
3-11
(R = CF3)
(R = tBu)
R2
12-16
d
F
F
F
O
NH2
O
NH2
b, c
N
N
H
H
N
N
Cl
R1
N
R2
O
S
S
O
17-19
26
Scheme 2. Reagents and conditions: (a) bromoacetyl bromide, Et3N, CH2Cl2, rt, 71%; (b) 3-trifluoromethyl pyrazole or 3-tert-butylpyrazole derivative, NaH, DMF, 0 °C to rt,
27–96%; (c) NaBH(OAc)3, AcOH, CH2Cl2, rt, 12–50%; (d) 3-chloropropionyl chloride, Et3N, CH2Cl2, rt, 79%.
O
H2N
O
O
N
N
N
N
HN
N
H
N
F
b, c
d, e
a
F
N
F
N
F
F
S
O
F
O
O
O
O
O
O
N
H
20
29
28
27
Scheme 3. Reagents and conditions: (a) morpholine, cat. PTSA, PhMe, reflux, quant.; (b) trifluoroacetic anhydride, Et3N, CH2Cl2, 5 °C to rt, quant.; (c) hydrazine hydrate, EtOH,
5 °C to rt, 43%; (d) 25, K2CO3, DMF, 50 °C, 36%; (e) TFA/CH2Cl2, rt, 96%.
2. Armstrong, N.; Sun, Y.; Chen, G. Q.; Gouaux, E. Nature 1998, 395, 913.
bioavailability with retention of receptor potency as well as prom-
ising in vivo activity.
Coordinates and structure factors have been deposited in the Protein Data
Bank for complexes of compounds 1 (3O6G), 5 (3O6H) and 12 (3O6I).
We believe that 5 provides us with a valuable asset in order to
delineate the role of AMPA receptors in neurological disorders.
Additional pharmacological profiling of 5 is on-going and will be
the subject of a separate publication.
3. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv. Drug Delivery Rev.
1997, 23, 3.
4. Lipinski, C. A. J. Pharmacol. Toxicol. Methods 2001, 44, 235.
5. Selectivity profiling against a panel of 80 receptors and channels carried out by
CaliperLifeSciences, 68ElmStreet, Hopkinton,MA01748 USA;www.caliperls.com.
6. Vandergriff, J.; Huff, K.; Bond, A.; Lodge, D. Neuropharmacology 2001, 40, 1003.
7. Gewald, K.; Schinke, E.; Boettcher, H. Chem. Ber. 1966, 99, 94.
8. Pennell, A. M. K.; Aggen, J. B.; Wright, J. J. K.; Sen, S.; McMaster, B. E.; Dairaghi, D.
J.; Chen, W.; Zhang, P. U. S. Patent Appl. Publ. 2006106218, 2006.
References and notes
1. Preceding Letter.