The discovery of adamantyl-derived, inhaled, long acting β2-adrenoreceptor agonists

Article abstract of DOI:10.1016/j.bmcl.2008.01.034

The design and profile of a series of adamantyl-containing long acting β₂-adrenoreceptor agonists are described. An optimal pharmacokinetic profile of low oral bioavailability was combined with a strong pharmacology profile when assessed using

Full text of DOI:10.1016/j.bmcl.2008.01.034

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1280–1283
The discovery of adamantyl-derived, inhaled, long acting
b2-adrenoreceptor agonists
Alan D. Brown,a Mark E. Bunnage,a Paul A. Glossop,a Kim James,a Rhys Jones,a
Charlotte A. L. Lane,a Russell A. Lewthwaite,a Simon Mantell,a Christelle Perros-Huguet,a
David A. Price,b,* Mike Trevethicka and Rob Webstera
aPfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Sandwich, Kent CT13 9NJ, UK
bPfizer Global Research and Development, Groton/New London Laboratories, Pfizer Inc., Eastern Point Road,
Groton, CT 06340, USA
Received 2 December 2007; revised 7 January 2008; accepted 9 January 2008
Available online 12 January 2008
Abstract—The design and profile of a series of adamantyl-containing long acting b2-adrenoreceptor agonists are described. An opti-
mal pharmacokinetic profile of low oral bioavailability was combined with a strong pharmacology profile when assessed using a
guinea pig trachea tissue model. A focus was then placed on developing a robust synthetic route to ensure rapid delivery of material
for clinical trials.
Ó 2008 Elsevier Ltd. All rights reserved.
Long acting b2-adrenoreceptor agonists are a highly
precedented drug class used for the treatment of asthma
and chronic obstructive pulmonary disease (COPD).1,2
OH
H
N
O
There are currently two marketed long acting b2-adreno-
HO
receptor agonists, salmeterol3 and formoterol,4,5 which
are approved for twice daily dosing, (Fig. 1).
HO
OH
H
N
There has been considerable interest in the discovery of
a once daily b2-adrenoreceptor agonist with a recent
flurry of presentations, patent applications and licensing
agreements from a number of institutions. This activity
prompted our initial disclosure6 of 1 as a long acting b2-
adrenoreceptor agonist with the potential for an inhaled,
once daily dosing regime. We described our medicinal
chemistry strategy in the design of 1 to ensure that
any oral fraction following inhalation does not contrib-
ute to the overall systemic exposure of 1.7 This was
achieved by ensuring the compound had low cell perme-
ability as measured using a CaCo-2 cell line to predict
for poor absorption across the gut wall.8 Combined with
high microsomal instability to ensure high first pass
metabolism in the liver of any absorbed fraction this en-
HO
OMe
NHCHO
Cl
Cl
OH
H
H
N
N
O
HO
HO
Figure 1. Structures of salmeterol, formoterol and 1.
sured low oral bioavailability. Coupled with this optimal
pharmacokinetic profile the compound possessed a
strong pharmacology profile, it was equipotent with
salmeterol and displayed a longer duration of action
than salmeterol when profiled in a guinea pig trachea
model.9 Considering the fully optimised preclinical
Keywords: b2; Adrenoreceptor; Adamantyl; Saligenin; Asthma;
Crystalline.
*
Corresponding author.
david.a.price@pfizer.com
Tel.:
+44
1034
644690; e-mail:
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.034

A. D. Brown et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1280–1283
1281
pharmacokinetic/pharmacology package that we had
achieved with 1 it was going to be a significant challenge
to deliver a compound with a superior profile; indeed we
decided that a compound with an equivalent pharmacol-
ogy/pharmacokinetic profile but which addressed poten-
tial risks with 1 would be of value to the portfolio as a
back up strategy.
man microsomal preparation and possessed poor cell
permeability as measured in the CaCo-2 cell line. This
in vitro pharmacokinetic data suggested the profile we
desired of poor gut wall permeability coupled with high
first pass metabolism was achievable in this alternative
expression. The duration of action as measured in a gui-
nea pig trachea model was shorter than the lead 1 which
was attributed to the reduction in lipophilicity. In previ-
ous work we had found a correlation between lipophil-
icity and duration of action in this tissue based model
and this has also been observed by other groups.6,10
The inhaled b2 agonist market is a highly competitive
arena and it was imperative that we obtained clinical
data as rapidly as possible to validate our screening cas-
cade and biological models. A key goal to expedite pro-
gression to the clinic is ensuring the synthesis of the
clinical candidate is as short as possible to maximise syn-
thetic accessibility of bulk quantities of the active phar-
maceutical ingredient (API). A parameter often used in
the measurement of synthetic complexity is the number
of steps contained in the longest linear sequence. With
the intrinsic structural complexity of many b2-adrenore-
ceptor agonists it was accepted that the synthesis of
these compounds would be relatively complex and re-
quire numerous steps. However, to try and counter this
it was decided to ensure as large a number of crystalline
intermediates as possible and a crystalline active phar-
maceutical ingredient (API) so that purification of bulk
material was as simple as possible. While compound 1
itself was crystalline as the L-tartrate salt, in the route
used there were no crystalline intermediates and this pre-
sented challenges in the efficiency of scaling up the
chemistry to deliver bulk quantities for any clinical stud-
ies (Scheme 1).
Previous experience in the project showed an excellent
correlation between the potencies of compounds using
the cell based assay and potency in guinea pig tissue
and to reduce ex vivo tissue work potencies were only
run on compounds that were of sufficient interest to
the project.6 A duration of action was measured as the
time taken for the muscle tone at an Emax concentration
of the compound to recover by 50% of the inhibition in-
duced where the Emax is the maximum inhibition achiev-
able by that compound. The Emax concentration was
estimated from the potency measured in the cell based
assay. The compounds were described as long (L) if they
possessed a duration of action similar to that of lead 1
(>6 h) and medium (M) if they possessed a duration of
action shorter than 1 (<6 h). After this period of time
tissue fade would begin and further measure of duration
of action would be unreliable. This contracted tissue
screening paradigm was essential for the rapid prosecu-
tion of the project and selection of a candidate for clin-
ical studies.
After our early extensive efforts using benzylamines as
coupling partners in this template it was proving difficult
to deliver a compound with the required pharmacologi-
cal/pharmacokinetic properties that offered differentia-
tion over 1. Attention now turned to examining
whether alternative amines could be used and in partic-
ular cyclic saturated amines. Initial replacement of the
dichlorophenyl expression in 1 with a cyclohexyl group
generated 2 and we were delighted to find that this com-
pound retained many of the attractive features of the
lead series. The compound was essentially equipotent
at the recombinant human b2 receptor expressed in
CHO cells and still retained excellent selectivity over
b1. Even though the compound was significantly less
lipophilic than 1 it was still rapidly turned over in a hu-
To increase the lipophilicity of the cyclohexyl lead 2 the
homologated analogue cycloheptane analogue 3 was
prepared which gave a compound that was less potent
than 1 but showed a long duration of action in the gui-
nea pig tissue model. This gave us confidence that within
these cycloalkyl expressions we should be able to deliver
the required pharmacological profile.
Truncation and concomitant ring expansion to generate
the isolipophilic 4 gave the required duration of action
in a compound that was still highly potent in its effect
on the b2 adrenoreceptor and highly selective over the
b1 receptor. However, we were confident that within this
expression we could deliver a more potent, selective
H
H
H
N
H
Ph
N
OMe
a
Ph
N
OH
b
Ph
N
O
O
O
9
11
10
c
OTBDMS
H
N
H
H
N
H2N
N
e,f
d
L-tartrate
O
O
BnO
HO
7
13
12
Scheme 1. Reagents and conditions: (a) aq LiOH, THF, 95%; (b) 1-adamantylamine, DIP, THF, 82%; (c) Pd/C, NH4HCO2, EtOH, 90%; (d) melt
reaction with saligenin head group, 72%; (e) Pd/C, NH4HCO2, EtOH, 100%; (f) L-tartaric acid, EtOH, 55%.

1282
A. D. Brown et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1280–1283
compound to ensure we delivered a high quality alterna-
tive to 1. It was clear that if we were to combine the re-
quired lipophilicity to drive duration of action with the
exquisite potency we could obtain with cyclohexyl de-
rived substituents we would need to be more creative
than ring homologation as in 4.
our series. Adamantyl is a chemically robust expression
which would be amenable to a variety of reaction condi-
tions and so opened up the possibilities of being able to
extensively reorder the synthetic route to give alternative
crystalline intermediates.
Incorporation of an adamantyl group into the homolo-
gated series generated 5 which displayed a reduction in
potency, however, in the direct linked series 7 possessed
an exciting profile with salmeterol-like potency and
excellent b1 selectivity. It is interesting to note that the
alternative 2-adamantyl regioisomer 6 showed 3-fold
lower potency in the primary assay suggesting how
specific the binding is in this region of the receptor
(Table 1).
A review of the structure of marketed drugs containing
cycloalkyl substituents was undertaken to design alter-
native expressions to the cycloalkyl substituent in 2.
From the anti-infectives literature amantidine
8
(Fig. 2) caused interest as an antiviral agent which spe-
cifically inhibits the replication of influenza A viruses
at low concentrations. It was also encouraging that as
a marketed agent amantidine would have been exten-
sively reviewed for safety issues and this gave us confi-
dence in the incorporation of an adamantyl group into
Progression of 7 into the guinea pig tissue assay con-
firmed an extended duration of action comparable to
1. Guinea pig in vivo studies showed that 7 had an iden-
tical pharmacological profile to that of the lead com-
pound 1 following intra-tracheal dosing in terms of its
potency and duration of action. With this data 7 was
prioritised for full pharmacokinetic profiling and the ini-
tial work was to assess cell permeability as measured by
H
N
O
Figure 2. Amantidine 8.
Table 1. LogD, cell potency and tissue duration of action for analogues
OH
H
H
N
R
N
O
HO
HO
Compounds
logD
Potencya (pM)
30
b1/b2
HLM (min)
4
Duration of actionb (h)
Cl
1
2.6
6339
L
*
Cl
2
1.5
2.0
1.9
2.5
2.1
2.1
2.1
20
118
200
250
190
60
6787
1886
192
3
M
L
L
L
L
L
L
*
3
Not done
*
4
2
*
5
547
2
*
*
6
451
2
7
2211
7885
2
*
Salmeterol
—
70
—
a Potency and efficacy at human recombinant b adrenoreceptors expressed in CHO cells assessed as elevations in cyclic AMP (n = 3). In this assay all
compounds appeared to be full agonists.
b A duration of action is measured as the time taken for the muscle tone at an Emax concentration of the compound to recover by 50% of the
inhibition induced where the Emax is the maximum inhibition achievable by that compound (n = 2). The compounds were described as long (L) if
they possessed a lead compound 1-like duration and medium (M) if they possessed a duration of action shorter than 1.

A. D. Brown et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1280–1283
1283
apical to basal flux rate through a monolayer of CaCo-2
cells. This assay suggested that not only did 7 possess
intrinsically poor cell permeability but it was also highly
effluxed by transporter proteins. This in vitro data com-
bined with the low microsomal stability demonstrated
that 7 should have low oral bioavailability due to poor
absorption through the gut wall and high first pass
metabolism. Analogue 7 was progressed into rat
in vivo studies and following intra venous administration
the volume of distribution was high (Vd = 14 L/kg),
however, with a clearance of greater than liver blood
flow (103 ml/min/kg) the measured half life was 1.6 h.
An oral cross over was conducted and confirmed that
in rat the oral bioavailability was <5%. This pharmaco-
kinetic profile added to the confidence that any swal-
lowed fraction of the inhaled dose in man would not
contribute to the systemic exposure of 7. Compound 7
was also screened for off target pharmacology and
showed no significant affinity (<100 nM) for other
receptors, enzymes or ion channels.
ery of high quality material for subsequent clinical
studies.
In conclusion, we have described our efforts to deliver
a b2-adrenoreceptor agonist that has a longer duration
of action in pre-clinical models than salmeterol. Addi-
tionally a key design feature was to ensure that com-
pounds would have low oral bioavailability compared
to salmeterol to reduce systemic effects through the
swallowed fraction after inhalation. This was achieved
through introducing amide functionality with high
hydrogen bonding potential to limit absorption while
also ensuring high first pass metabolism by the liver.
Furthermore, through collaboration with experts in
large scale chemistry 10 was designed with the goal
of delivering an optimised synthetic route at time of
nomination to expedite delivery of bulk material for
clinical trials.
Acknowledgments
Attention now turned to reordering the synthetic route
to maximise the number of crystalline intermediates
and minimise synthetic complexity. Within the early dis-
covery phase of the project the amide coupling was per-
formed as the final step to allow rapid variation of the
amine expression. However, there were no crystalline
intermediates in the route which presented difficulties
in purification when scaling up the final optimised com-
pound. It was decided that with such a chemically ro-
bust adamantyl expression the amide coupling could
be performed earlier in the route. This had the advan-
tage that the new intermediate 10 required was highly
crystalline and precipitated from the reaction solvents
as the hydrochloride salt and could be isolated by filtra-
tion. Amide coupling of 1-adamantylamine and cleavage
of the benzylamine chiral auxiliary generated 12 which
could again be purified by recrystallisation. The synthe-
sis was completed with installation of the saligenin
expression, hydrogenation and cleavage of the benzyl
protecting group and finally removal of the silicon pro-
tecting group. This could be achieved using L-tartaric
acid which performed the protodesilylation and salt for-
mation to give crystalline, high purity final material.
This focus on route optimisation, collaboration with
large scale chemistry experts and selection of a clinical
candidate using synthetic expedience as a key criterion
was a highly successful approach to ensure a rapid deliv-
The authors recognise Michele Coghlan, Susan Sum-
merhill and Karen Wright, for their expertise in generat-
ing and analysing the biological data.
References and notes
1. Waldeck, B. Eur. J. Pharmacol. 2002, 445, 1.
2. Siafakas, N. M.; Vermeire, P.; Price, N. B. Eur. Respir. J.
1995, 8, 1398.
3. Johnson, M. Med. Res. Rev. 1995, 153, 225.
4. Anderson, G. P. Life Sci. 1993, 52, 2145.
5. Palmqvist, M.; Persson, G.; Lazer, L.; Rosenborg, J.;
Larsson, P.; Lotvall, J. Eur. Respir. J. 1997, 10, 2484.
6. Brown, A. D.; Bunnage, M. E.; Glossop, P. A.; James, K.;
Jones, R.; Lane, C. A. L.; Lewthwaite, R. A.; Mantell, S.;
Perros-Huguet, C.; Price, D. A.; Trevethick, M.; Webster,
R. Bioorg. Med. Chem. Lett. 2007, 17, 4012.
7. Bennett, J. A.; Harrison, T. W.; Tattersfield, A. E. Eur.
Respir. J. 1999, 13, 445.
8. Artursson, P.; Palm, K.; Luthman, K. Adv. Drug Delivery
Rev. 1996, 12, 67.
9. Nials, A. T.; Sumner, M. J.; Johnson, M.; Coleman, R. A.
Br. J. Pharmacol. 1993, 108, 507.
10. Austin, R. P.; Barton, P.; Bonnert, R. V.; Brown, R. C.;
Cage, P. A.; Cheshire, D. R.; Davis, A. M.; Dougall, I. G.;
Ince, F.; Pairaudeau, G.; Young, A. J. Med. Chem. 2003,
46, 3210.