Potent and selective α1A adrenoceptor partial agonists-Novel imidazole frameworks

Article abstract of DOI:10.1016/j.bmcl.2009.03.162

Novel imidazole frameworks have been identified as potent partial agonists of the α_1A adrenergic receptor, with good selectivity over the α_1B, α_1D and α_2A receptor sub-types. Nitrile 28 possessed attractive CNS

Full text of DOI:10.1016/j.bmcl.2009.03.162

Bioorganic & Medicinal Chemistry Letters 19 (2009) 3118–3121
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Potent and selective a1A adrenoceptor partial agonists—Novel
imidazole frameworks
*
Gavin A. Whitlock , Paul E. Brennan, Lee R. Roberts, Alan Stobie
Pfizer Global Research and Development, Sandwich Labs, 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:
Novel imidazole frameworks have been identified as potent partial agonists of the a1A adrenergic recep-
tor, with good selectivity over the a1B, a1D and a2A receptor sub-types. Nitrile 28 possessed attractive CNS
drug-like properties with good membrane permeability and no P-pg mediated efflux. 28 also possessed
excellent solubility, metabolic stability and wide ligand selectivity.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 26 February 2009
Revised 27 March 2009
Accepted 29 March 2009
Available online 7 April 2009
Keywords:
Partial agonist
a1A Adrenergic receptor
a
1-Adrenoceptorsaremembersofthe7TMsuperfamilyofG-pro-
The a1A pharmacology of these new templates was then
compared to the original indanyl framework 3 (Table 1), and a
key criteria for progression for new templates was a significant
increase in activity over 3.
tein-coupled receptors, and three subtypes of
have been cloned (a1A
Recent disclosures have suggested that selective partial agonists of
a1-adrenoceptors
, a1B, a
1D), expressed and characterized.1
the
a
1A receptor may have clinical utility in the treatment of stress
The isomerised 2-indanyl 2-imidazole 4 showed weaker activity
and was not taken further. In contrast, the tetrahydrobenzimidaz-
ole 5 showed encouraging a1A activity with excellent selectivity
when screened as a racemic mixture. Screening of single enantio-
mers 5a and 5b showed that both exhibited a1A pharmacology
but 5a was slightly more potent. Absolute stereochemistry was
not determined at this time. N-Alkylated imidazole structures 6
and 7 were particularly interesting as they lacked the N–H of most
urinary incontinence (SUI).2 In addition, a1A partial agonists may
have increased selectivity for SUI efficacy over undesired cardiovas-
cular effects when compared to full a1A agonists.2 Recently, we
described a novel series of 2-imidazole a1A partial agonists with
excellent selectivity over a1B, a1D and a2A and attractive drug-like
properties.3 However, as part of our work on this mechanism, biol-
ogy studies indicated that a1A partial agonists impart their in vivo
efficacy in models of SUI via a centrally mediated pathway,4 rather
than a direct effect on urethral smooth muscle. Compounds such
as our initial lead compound 1 were shown to have poor BBB pene-
tration, which was linked to high P-pg mediated efflux when
assessed in MDCK MDR-1 cells. Further work was then undertaken
to improve BBB penetration by reducing TPSA and compound 2
was identified (Fig. 1) which showed excellent BBB penetration.4,5
As part of a multi-series approach to a1A partial agonists, we then
sought to discover further imidazole templates which could deliver
imidazole or imidazoline
a
-agonists.6 The 5,5-system 6 showed
good a1A activity whereas the 6,5 template 7 had slightly reduced
activity. The single enantiomers 6a and 6b showed a very different
activity profile, with essentially all the a1A pharmacology residing
in enantiomer 6a.
N
MeSO2NH
N
MeO
potent and selective a1A partial agonist activity with physicochem-
N
H
N
H
ical properties aligned with good BBB penetration (low TPSA, low
mwt, few hydrogen bonding groups), and ideally improved meta-
bolic stability when compared to 2. We now wish to report the
results of that work in this Letter and the article following this one.
The initial medicinal chemistry strategy focused on simple tem-
plates linking an aryl ring to an imidazole in a variety of conform-
ationally constrained systems 4–7 (Fig. 2).
Cl
Cl
2
1
α1A EC50 27nM (58%)
α1A EC50 43nM (61%)
Selective over α1B, α1D, α2A
Selective over α1B, α1D, α2A
HLM Clint 30 μl/min/mg
MDCK MDR-1 40/36
Good BBB penetration
HLM Clint <7 μl/min/mg
MDCK MDR-1 7/29
Poor BBB penetration
* Corresponding author. Tel.: +44 1304 649174; fax +44 1304 651987.
E-mail address: gavin.whitlock@pfizer.com (G.A. Whitlock).
Figure 1. Structures, pharmacology and ADME properties of imidazole a1A partial
agonists 1 and 2.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.03.162

G. A. Whitlock et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3118–3121
3119
N
N
N
N
N
N
N
N
H
N
N
H
H
3
5
7
4
6
Figure 2. Structures of proposed new imidazole templates 4–7.
Table 1
In vitro functional a1A, a1B, a1D and a2A agonist activity for compounds 3–7
a,b
a,b
a,b
Compds
a1A EC50 (nM)a,b
a1A Emax (%)
a1B EC50 (nM) (Emax
)
a1D EC50 (nM) (Emax
)
a2F EC50 (nM) (Emax)
3
4
5
5a
5b
6
6a
6b
7
473
1650
234
123
315
83
73
>10,000
721
54
26
78
75
76
80
74
25
68
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
7360 (30%)
>10,000
>10,000
NT
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
NT denotes not tested.
a
See Ref. 3 for description of assay conditions.
Values are geometric means of at least three experiments.
b
Br
N
O
(a)
(b)
(c)
+
N
H
O
O
O
Br
O
R1
8
R1
9
R1
R1
10-16
NO2
CO2Et
H
N
N
N
CO2Et
(d)
(e)
(f), (g)
R1
R1
17
R1
18
19
R1
20-29
Scheme 1. Reagents and conditions: (a) either (i) ArBr, Mg, CuI, TMS-Cl or (ii) ArB(OH)2, cat. Pd(OAc)2, cat. SbCl3, AcOH, 50 °C; (b) NBS, Amberlyst-15, EtOAc, rt; (c)
Formamidine acetate, Hunigs base, DMSO, 80 °C; (d) MeNO2, DBU, 0 °C to rt; (e) either (i) RaNi, 100 psi H2, 100 °C or (ii) SnCl2, EtOAc, rt; (f) Me3OBF4, CH2Cl2, rt; (g) (i)
H2NCH2CH(OEt)2, HCl, EtOH, rt; (ii) HCl/dioxane, H2O, 100 °C.
With two new templates 5 and 6 in hand, synthetic routes were
devised which would give access to aryl-substituted analogues
(Scheme 1). The tetrahydrobenzimidazoles were accessed via a
three step route from cyclohexenone. Conjugate addition of either
an aryl cuprate or aryl boronic acid gave the 3-aryl cyclohexanones
8. Bromination with NBS proceeded in a non-selective fashion to
In silico modeling was also carried out, to increase our under-
standing of how 5 and 6 overlapped with the 1-indanyl 2-imidaz-
ole 3. It was hoped this modeling would then guide substitution off
the aryl ring of 5 and 6. The equatorial conformation of 5 over-
lapped poorly with the highly constrained 1-indanyl template 3.
We then speculated that 5 could adopt an axial conformation with
only a small energy penalty (0.8 kcal/mol),8 and this overlapped
much better, with both imidazole and aryl rings in close alignment
(Fig. 3). This suggested that substitution in the 2-position of the
aryl ring of 5 should overlap with the key methoxymethyl group
of compound 2. Similar modeling was carried out for the pyrrolo-
imidazole framework 6. In this case the alignment of the aromatic
rings appeared to be less optimal (Fig. 3), however the modeling
again suggested the 2-position to have the best overlap. A small
range of analogues were then prepared to test these hypotheses.
To rapidly assess potency, selectivity and drug-like properties,
pharmacology and in vitro ADME data (logD, human liver micro-
some stability (HLM) and Pampa membrane permeability) was
generated in parallel.
give regioisomeric
a-bromoketones 9. These could not be sepa-
rated and so were reacted as a crude mixture with formamidine
acetate to afford the desired compounds 10–16. Reactions (b)
and (c) were poor yielding, but were sufficiently robust to allow
initial SAR investigations to be carried out. The pyrrolo-imidazole
series was accessed using an efficient four step procedure. Conju-
gate addition of nitromethane to the unsaturated esters 17
proceeded in high yield (80–100%), and was followed by reduction
of nitro compounds 18 to the primary amines which underwent
spontaneous cyclisation to the lactam 19.7 Conversion of the
lactam to the cyclic imino-ether was effected with Meerweins
reagent. This was followed by displacement with
a glycine
aldehyde equivalent, and cyclisation afforded the desired target
compounds, which were then separated into the individual enanti-
omers by preparative chiral HPLC.
In the tetrahydrobenzimidazoles 10–16 we were pleased to see
that a 2-substituent did give better a1F potency (Table 2, examples

3120
G. A. Whitlock et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3118–3121
Figure 3. In silico overlap of (a) axial conformations of 5 (green) with 3 (purple) and (b) low energy conformation of 6 (blue) with 3 (purple).
Table 2
In vitro functional a1A
,
a1B and a1D agonist activity, in vitro ADME data and logD values for compounds 5, 6, 10–16 and 20–29
N
N
R1
R1
N
N
H
10-16
20-29
a,b
a,b
Compds R1
a1A EC50 (nM)a,b a1A Emax (%) a1B EC50 (nM) (Emax
)
a1D EC50 (nM) (Emax
)
LogD HLM Clint (
l
l/min/mg) PAMPA Papp (10À6 cm/s)
5a
6a
10
11
12
13
14
15
16
16a
16b
20
21
22
23
24
25
26
27
28
29
H
H
2-Cl
3-Cl
123
73
276
514
>10,000
144
1050
>10,000
75
74
60
71
—
95
60
—
87
83
85
56
45
82
56
27
63
75
87
74
77
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
1100 (20%)
>10,000
NT
>10,000
NT
1770 (36%)
NT
NT
>10,000
>10,000
>10,000
7360 (30%)
>10,000
>10,000
>10,000
3500 (25%)
>10,000
>10,000
>10,000
>10,000
>10,000
364 (50%)
>10,000
NT
>10,000
NT
301 (58%)
NT
NT
>10,000
>10,000
2.6
1.9
3.1
3.3
3.3
2.8
2.6
2.6
2.2
2.2
2.2
2.6
2.6
1.9
2.0
2.0
2.0
2.3
NT
1.3
1.8
7
47
51
34
39
NT
NT
NT
NT
41
30
27
51
34
57
47
57
NT
35
NT
23
25
15
10
21
16
NT
NT
16
<7
<7
<7
18
32
37
23
30
NT
22
NT
<7
<7
4-Cl
2-OMe
3-OMe
4-OMe
2-CH2OMe 76
2-CH2OMe 48
2-CH2OMe 70
2-Cl
4-Cl
2-OMe
3-OMe
4-OMe
2-F
19
3460
22
578
939
42
2-OCHF2
11
2-C(O)NH2 371
2-CN 51
2-CH2OMe 82
NT denotes not tested.
a
Values are geometric means of at least two experiments.
See Ref. 3 for description of assay conditions.
b
10 and 13) when compared to 3 and 4-substittution (examples 11,
12, 14 and 15). The preferred methoxymethyl group was then
introduced and this led to an increase in a1A activity in example
16. Enantiomer separation again showed both enantiomers were
active, with 16a the slightly more potent enantiomer. This series
did combine good a1A potency, selectivity and attractive ADME
properties (good HLM stability and passive membrane permeabil-
ity). However, this series generally had high Emax values, indicative
of full agonism and more complex analogues proved too challeng-
ing to make via the synthetic route outlined, and so this series was
halted.
Attention was then turned to the pyrrolo-imidazole template.
For clarity only the analogues (20–29) from the more potent enan-
tiomeric series are reported, although the absolute stereochemis-
try was not determined at this time. Chloro and methoxy
analogues 20–24 suggested that 2-substitution increased potency
over 6a, however the 2-Cl analogue 20 had poor selectivity over
expand the SAR. Small lipophilic groups were well tolerated, such
as fluoro (25) and OCHF2 (26) whereas the amide 27 lost potency.
However, 25 still had poor selectivity over a1B and a1D. The nitrile
28 had acceptable potency, good selectivity, excellent metabolic
stability and passive membrane permeability. Interestingly the
methoxymethyl group (29), which was preferred in the indanyl
2-imidazole lead 2, did not increase potency in this case.
The nitrile 28 had the best balance of pharmacology and in vitro
ADME properties, and was assessed further. Compound 28 had no
activity against the hERG channel (up to 22 lM), and no Ki < 1 lM
were found when the compound was screened against a panel of
150 receptors, ion channels and enzymes at Cerep. The aqueous sol-
ubility of highly crystalline 28 free base was also measured at
>0.5 mg/ml across a pH range of 1–9. Additionally, 28 had excellent
fluxwhenassessed inMDCKMDR-1cells(A-Bflux = 34 Â 10À6 cm/s,
B-A flux = 41 Â 10À6 cm/s) which was indicative of excellent passive
permeability combined with no evidence of P-pg mediated efflux.
This solubility and flux data should be predictive of both good oral
a1B and
a1D. Further 2-substituents were then introduced to

G. A. Whitlock et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3118–3121
3121
absorption and good BBB penetration.9 Compound 28 was also pro-
gressedtorati.v. pharmacokinetics, andshowedmoderateclearance
(Cl 34 ml/min/kg) and low volume of distribution (Vd 1.5 L/kg) after
bolus solution dosing (0.3 mg/kg).
The remaining concern with the pyrrolo-imidazoles was a1A
Emax, which was viewed to be too high at 70–80%.10 The Letter
immediately following this one11 will detail how a1A Emax was
reduced and EC50 activity was increased, whilst retaining the
desired ADME and CNS drug-like properties of 28.
2. Musselman, D. M.; Ford, A. P. D. W.; Gennevois, D. J.; Harbison, M. L.; Laurent, A.
L.; Mokatrin, A. S.; Stoltz, R. R.; Blue, D. R. BJU Int. 2004, 93, 78.
3. Whitlock, G. A.; Conlon, K.; McMurray, G.; Roberts, L. R.; Stobie, A.; Thurlow, R.
J. Bioorg. Med. Chem. Lett. 2008, 18, 2930.
4. Conlon, K.; Christy, C.; Whitlock, G. A.; Roberts, L. R.; Stobie, A.; Thurlow, R. J.;
McMurray, G. J. Pharmacol. Exp. Ther., submitted for publication.
5. Roberts, L. R.; Bryans, J.; Conlon, K.; McMurray, G.; Stobie, A.; Whitlock, G. A.
Bioorg. Med. Chem. Lett. 2008, 18, 6437.
6. In a recent review of
a1-receptor agonists, no N-alkylated imidazolines or
imidazoles were described. Bishop, M. J. Curr. Topics Med. Chem. 2007, 7, 135.
7. CAUTION should be exercised in the handling of nitro compounds. DSC analysis
was determined on all alkyl nitro analogues 18. If an exotherm was observed at
>200 °C then the RaNi conditions at 100 °C were deemed to be safe. For
analogues with DSC exotherms at <200 °C the more ambient SnCl2 conditions
were employed.
Acknowledgements
8. In silico molecular modeling calculated
a 0.8 kcal/mol energy difference
We would like to thank Alison Bridgeland for screening data;
Debbie Lovering, Edward Pegden, Danny Ho, Katherine England,
Rachel Osborne, Simon Wheeler and Helen Mason for compound
synthesis. We also thank Peter Bungay for ADME assessment of
compounds 10–29.
between axial and equatorial conformations of 5. This energy gap is well
within the typical 5 kcal/mol difference observed for bound versus low energy
conformations reported in the literature, see: Perola, E.; Charifson, P. S. J. Med.
Chem. 2004, 47, 2499.
9. Wang, Q.; Rager, J. D.; Weinstein, K.; Kardos, P. S.; Dobson, G. L.; Li, J.; Hidalgo, I.
J. Int. J. Pharm. 2005, 288, 349.
10. The Roche clinical agent Ro 115-1240 had an Emax of 60% in our hands, which
was in line with published data, see Blue, D. R.; Daniels, D. V.; Gever, J., R.; Jett,
M. F.; O’Yang, C.; Tang, H. M.; Williams, T. J.; Ford, A. P. D. W. BJU Int. 2004, 93,
162. Ro 115-1240 was reported not to cause cardiovascular effects at the dose
tested in the clinic, see Ref. 2.
References and notes
1. (a) Hieble, J. P.; Bylund, D. B.; Clarke, D. E.; Eikenburg, D. C.; Langer, S. Z.;
Lefzowitz, R. J.; Minneman, K. P.; Ruffolo, R. R., Jr. Pharmacol. Rev. 1995, 47, 267;
(b) Zhong, H.; Minneman, K. P. Eur. J. Pharmacol. 1999, 375, 261.
11. Roberts, L. R.; Brennan, P. E.; Fish, P. V.; Storer, R. I.; Whitlock, G. A. Bioorg. Med.
Chem. Lett. 2009, 19, 3113.