The identification of 7-[(R)-2-((1S,2S)-2-benzyloxycyclopentylamino)-1- hydroxyethyl]-4-hydroxybenzothiazolone as an inhaled long-acting β2-adrenoceptor agonist

Article abstract of DOI:10.1016/j.bmcl.2014.06.014

The optimisation of two series of 4-hydroxybenzothiazolone derived β₂-adrenoceptor agonists, bearing α-substituted cyclopentyl and β-phenethyl amino-substituents, as inhaled long-acting bronchodilators is described. Analogues were selected for synthesis using a lipophilicity based hypothesis to achieve the targeted rapid onset of action in combination with a long duration of action. The profiling of the two series led to identification of the α-substituted cyclopentyl analogue 2 as the optimal compound with a comparable profile to the inhaled once-daily long-acting β₂-adrenoceptor agonist indacaterol. On the basis of these data 2 was promoted as the backup development candidate to indacaterol from the Novartis LABA project.

Full text of DOI:10.1016/j.bmcl.2014.06.014

Bioorganic & Medicinal Chemistry Letters 24 (2014) 4341–4347
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The identification of 7-[(R)-2-((1S,2S)-2-benzyloxycyclopentylamino)
-1-hydroxyethyl]-4-hydroxybenzothiazolone as an inhaled
long-acting b2-adrenoceptor agonist
Nicola Arnold, David Beattie, Michelle Bradley, Andrew Brearley, Lyndon Brown, Steven J. Charlton,
⇑
Robin A. Fairhurst , David Farr, John Fozard, Joe Fullerton, Martin Gosling, Julia Hatto, Diana Janus,
Darryl Jones, Lynne Jordan, Christine Lewis, Janet Maas, Clive McCarthy, Mark Mercer, Helen Oakman,
Neil Press, Rachel Profit, Friedrich Schuerch, David Sykes, Roger J. Taylor, Alexandre Trifilieff,
Andrew Tuffnell
Novartis Institutes for BioMedical Research, Respiratory Diseases Area, Horsham, United Kingdom
Novartis Institutes for BioMedical Research, Respiratory Diseases Area, Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
The optimisation of two series of 4-hydroxybenzothiazolone derived b2-adrenoceptor agonists, bearing
Received 26 April 2014
Revised 4 June 2014
Accepted 5 June 2014
Available online 26 June 2014
a
-substituted cyclopentyl and b-phenethyl amino-substituents, as inhaled long-acting bronchodilators
is described. Analogues were selected for synthesis using a lipophilicity based hypothesis to achieve
the targeted rapid onset of action in combination with a long duration of action. The profiling of the
two series led to identification of the
a-substituted cyclopentyl analogue 2 as the optimal compound
with a comparable profile to the inhaled once-daily long-acting b2-adrenoceptor agonist indacaterol.
On the basis of these data 2 was promoted as the backup development candidate to indacaterol from
the Novartis LABA project.
Keywords:
b2-Adrenoceptor agonist
Bronchodilator
Ó 2014 Elsevier Ltd. All rights reserved.
Chronic obstructive pulmonary disease
Following the efforts of several groups over the last 15 years a
number of inhaled long-acting b2-adrenoceptor agonists (LABAs)
have entered into development. The most advanced of these LABAs
have now received regulatory approval and are set to become the
basis for the next generation of once-daily treatments for asthma
and chronic obstructive pulmonary disease (COPD).1–3 The suc-
cessful LABAs from this group are anticipated to fulfil this role,
either alone, or in fixed-dose combination-products with other
once-daily inhaled therapies.4 Of these LABAs indacaterol was
the first to be approved as a single agent in 2008 for the treatment
of COPD in Europe (brand name: Onbrez), and is now approved for
this indication in several countries including the United States
(Arcapta).5 More recently the fixed-dose combination-product of
indacaterol with the muscarinic receptor antagonist glycopyrroni-
um bromide has been approved in Europe and Japan for the
treatment of COPD (Ultibro).6 To manage risk following the identi-
fication of indacaterol, and to ensure the highest probability of
success for the Novartis LABA project, activities were initiated at
an early stage to find a backup compound to indacaterol. As no
issues were apparent with indacaterol during the early stages of
development a backup compound had been sought with an
equivalent preclinical profile.7 In addition, to be as divergent as
possible from indacaterol a compound from a different structural
series to the 8-hydroxyquinolinone of the lead candidate was
preferred. In this letter we describe the identification of a backup
inhaled-LABA development candidate to indacaterol from
a
4-hydroxybenzothiazolone b2-adrenoceptor (b2AR) agonist series.
Previously we have described a lipophilicity-based library
approach to generate LABA leads within a 4-hydroxybenzothiazo-
lone series which was inspired by the natural product S1319.8
More recently others have also described LABA development can-
didates from the same 4-hydroxybenzothiazolone series, including
the analogue 1.3 The Novartis library approach had identified inter-
esting amino-substituents in the 4-hydroxybenzothiazolone series
including the
a-substituted cyclopentyl analogues 2, 3 and 4, as
shown in Figure 1.
One optimisation strategy explored from this starting point was
to investigate the structure activity relationships (SAR) surrounding
⇑
Corresponding author at present address: Novartis Institutes for Biomedical
Research, Global Discovery Chemistry, Novartis Pharma AG, Werk Klybeck, Postfach,
CH-4002 Basel, Basel, Switzerland. Tel.: +41 61 69 64294; fax: +41 61 6966246.
E-mail address: robin.fairhurst@novartis.com (R.A. Fairhurst).
the a-substituent and the role of the cyclopentyl stereochemistry,
http://dx.doi.org/10.1016/j.bmcl.2014.06.014
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

4342
N. Arnold et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4341–4347
H
N
HN
HN
HN
OMe
HO
HO
HO
S
S
O
O
N
H
N
H
N
H
O
OH
OH
OH
indacaterol
S1319
1
HN
HN
HN
O
HO
HO
HO
S
S
S
O
O
O
N
H
N
H
N
H
OH
OH
OH
2
3*
4*
Figure 1. Structures of the b2AR agonists indacaterol, S1319, 1, 2, 3 and 4. ⁄1:1 Mixtures of diastereoisomers containing the racemic cis- and trans-cyclopentyl amino moieties
respectively.
leading to the targets of the general structure 5. In addition, b-phen-
ethyl amino-substituents had provided library members with
favourable profiles, as well as being established in a number of pre-
vious b2AR agonist structures, such as formoterol and 1, leading to
the targets of the general structure 6, as shown in Figure 2.
Additionally, the possibility to combine these two structural fea-
Scheme 2.10 In the key step, reductive amination with
a-phenyl-
ethylamine as the chiral auxiliary generated the cis-isomers with
a high level of facial selectivity. Absolute stereochemistry was
assigned based upon the reported observation that like generates
like between the
a-phenylethylamine centre and the induced
aminocyclopentyl centre.10
tures, by introducing an
moiety, was also seen as an interesting possibility (5, R = phenyl).
To prepare the desired -substituted cyclopentyl targets 5 an
a
-phenyl substituent into the cyclopentyl
To prepare the desired substituted b-phenethyl targets 6 an
alternative synthesis was employed in which tertiary amines were
used in the epoxide opening step, as shown in Scheme 3.11 For the
alkyl-substituted analogues: starting from either 3- or 4-bromo-
phenylacetic acid, conversion to the corresponding benzyl amide
was followed by the introduction of the desired alkyl groups
through Negishi couplings to give 14. Borane reduction then
gave the desire N-benyl b-phenylethylamines which cleanly
opened the epoxide 11, at the least-hindered position, to give the
protected target-compounds 15. Transfer hydrogenation under
acidic conditions then removed all three protecting groups to
provide the targeted 4-hydroxybenzothiazolones 6. The amines
to prepare the ether-substituted b-phenethyl containing analogues
were prepared from the corresponding hydroxyphenylacetic acids
in a closely related manner via the amides 16.11,12
a
improved version of our original synthesis was developed, as
shown in Scheme 1.8,9 As previously, a benzyne-mediated cyclisa-
tion-reaction of the 3-fluorophenyl thiocarbamate derivative 7
gave the 7-lithiated benzothiazole intermediate 8, which could
be acylated directly at ꢀ40 °C with the Weinreb amide of chloro-
acetic acid 9 to give 10. In contrast to the earlier route, under these
conditions no transmetalation step was required to modulate the
basicity of 8, and this transformation could be run reliably on a
multi-kilogram scale. Conversion of the chloroketone 10 to the
(R)-epoxide 11 followed the original sequence using a CBS reduc-
tion to generate the desired chirality.8 The
a-substituted cyclopen-
tylamines 12 were then monosilylated prior to reaction with the
epoxide 11. Epoxide opening occurred by addition to the least hin-
dered position to give the protected target-compounds 13. Acidic
deprotection then provided the targeted 4-hydroxybenzothiazol-
ones 5.
Following the above routes the analogues 17 to 35, shown in
Table 1, were prepared as single stereoisomers of the (R)-configu-
ration at the benzylic alcohol centre of the ethanolamine moiety.
We have previously described in detail the targeted LABA
preclinical profile that was put in place for the identification of
indacaterol, and the goal for the backup project was also to fulfil
The chiral
a-substituted cyclopentylamines 12 were prepared
based upon
a
previously described procedure, as shown in
R, OR
OMe
HN
HN
HN
R, OR
O
HO
HO
HO
S
S
O
O
N
H
N
H
N
H
H
OH
OH
OH
formoterol*
5
6
Figure 2. Structures of the b2AR agonists formoterol and the targeted series 5 and 6. ⁄Racemic mixture of the l-diastereoisomer.

N. Arnold et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4341–4347
4343
O
O
F
X
S
S
i)
ii), iii)
S
N
O
O
N
N
H
O
O
O
Me
Cl
O
N
8, X = Li
MeO
9
7
11
10, X = C(O)CH2Cl
HN
R, OR
O
HO
v)
iv)
5
S
H2N
N
R, OR
O
12
13
Scheme 1. Preparation of the 4-hydroxybenzothiazolone b2-adrenoceptor agonists of the general structure 5. Reagents and conditions: (i) 1.1 equiv t-BuLi (15% in pentanes),
THF (0.35 M), ꢀ70 °C, then warmed to ꢀ30 °C, 1.4 equiv t-BuLi (15% in pentanes), 1.5 h, then cooled ꢀ70 °C, 1.3 equiv 9 in THF (6.8 M), 3 h, then quenched at ꢀ30 °C (60%); (ii)
0.1 equiv (1R,2S)-(+)-cis-1-amino-2-indanol, 1.0 equiv borane (1 M in THF), THF, 15 min, room temperature, then recrystallized from heptane (98% ee, 77%); (iii) 2.0 equiv
K2CO3, 9:1 acetone/water, 50 °C, 24 h, (99% ee, 99%); (iv) 1.2 equiv 12, 0.6 equiv N,O-bis-(trimethylsilyl)acetamide, DMF, 30 min, room temperature, then 11, 80 °C, 18 h
(43–89%); (v) HCO2H, room temperature, 24–48 h, or 2:1 iPrOH/1N HCl(aq), 80 °C, 24 h (43–83%).
these criteria as closely as possible.7,13 To achieve the desired onset
i), ii)
iii), iv)
and duration of action profiles the lipophilicity of the potential
targets were estimated prior to synthesis. To select suitable
R-substituents, to introduce into the series 5 and 6, calculated
logD7.4 values were used to make this assessment. As described
previously maintaining the lipophilicity close to that which had
been targeted during the lead-generation phase was implemented
(within the clogD7.4 window of 2.4 to 3.5).8,12,13 To assess the bio-
logical profiles of the compounds: affinities for the human b1AR
and b2AR were initially assessed using radio-ligand binding assays.
In terms of the functional selectivity for the b2AR over the b1AR the
binding assays, which measure the interaction with the low-affin-
ity state of the receptors, provided a good selectivity ranking, but
were found to underestimate the functional selectivity.14 The
4-hydroxybenzothiazolone catechol mimetic was first described
in a series of dopamine receptor (DR) agonists,15 and DR activity
was identified to be the most significant off-target activity for
the series 5 and 6 following screening versus an internal panel of
enzymes and receptors.16 For the most interesting analogues
functional b2AR potency and efficacy, onset of action (OoA) and
duration of action (DoA) were then assessed in an electrically
stimulated superfused guinea-pig tracheal-strip assay as described
previously.14 These data for the reference compounds and ana-
logues 2, 3, 4 and 17–35 are shown in Table 1.
H2N
12
-substituted cyclopentylamines of the
O
O
R, OR
R, OR
Scheme 2. Preparation of the alkyl and aryl
a
general structure 12. Reagents and conditions: (i) 1.0 equiv R-MgX (2.0 M in
diethylether, where X = Cl, Br), 0.1 equiv CuI, 10 min room temperature, then drop
wise cyclopentene oxide (exothermic), room temperature, 4 h (46–70%); (ii)
3.0 equiv C6H6N+ClCrOꢀ3 (PCC, 20% on neutral alumina), toluene, room temperature,
18 h (27–66%); (iii) 1.2 equiv
a-phenylethylamine (either (R)- or (S)-enantiomer:
see text), 1.5 equiv AcOH, 1.5 equiv NaBH(OAc)3, 1,2-dichloroethane, room tem-
perature, 48 h (33–70%); (iv) 5 atm H2, 3 equiv AcOH, MeOH, room temperature,
18 h (81–93%).
R
Br
i), ii)
O
O
N
H
Ph
Ph
HO
HO
14
OR
OH
i), iii)
O
O
N
H
16
For the a-substituted cyclopentyl series 5 the observation from
R, OR
the library members was for little difference between the cis- and
trans-substituted analogues 3 and 4 in terms of b2AR activity and
selectivity. However, the ready access to the optically pure cis-
Ph
N
HO
iv), v)
vi)
a-substituted cyclopentylamines described above and the trend
14 or 16
6
for greater activity and selectivity for the cyclopentyl cis-analogues
3, versus the trans-analogues 4, led to focusing exclusively on the
cis-analogues going forward. In terms of the cyclopentyl stereo-
chemistry: comparing 17 and 19 with 18 and 20 indicated the
(R,R)-diastereoisomers to be more active than the (S,S)-diastereoi-
somers. This relationship was evident throughout the series, when
both diastereoisomers were prepared, and the emphasis was direc-
ted towards the more active (R,R)-diastereoisomers. However, the
likelihood of an increased level of technical complexity being
required to produce low dose formulations for inhalation resulted
in an upper limit on potency. As discussed previously, compounds
with b2AR Ki values less than formoterol were deprioritised to
S
N
O
O
15
Scheme 3. Preparation of the 4-hydroxybenzothiazolone b2AR agonists of the
general structure 6. Reagents and conditions: (i) 1.0 equiv BnNH2, 1.2 equiv EDCI,
1.2 equiv HOBt, CH2Cl2, water, room temperature, 18–24 h (35–75%); (ii) 1.5 equiv
R-ZnBr, 0.1 equiv Pd(dppf)Cl2, THF, room temperature, 24–48 h (68–93%); (iii);
1.2 equiv nBuBr, 2.0 equiv Cs2CO3, 0.3 equiv KI, AcCN, 16 h, reflux (86–87%); (iv)
2 equiv DIBALH, toluene, 50 °C, 4 h (39–54%); (v) 0.9 equiv 11, nBuOH, 110 °C, 24 h
(56–67%); (vi) Pd black, HCO2H, room temperature, 24 h (38–46%).

4344
N. Arnold et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4341–4347
Table 1
Structures of the amino substituents of the analogues 17–35, human b2AR, b1AR and D3R Ki values, calculated logD7.4 values and guinea-pig electrically stimulated superfused
tracheal strip data for the reference compounds and the 4-hydroxybenzothiazolone analogues 2, 3, 4 and 17–35
17
18
19
20
21
22
23
24
29
31
O
30
25
26
27
28
O
O
32
33
34
35
Compound
b2 Ki (nM)
b1 Ki (nM)
D3 Ki (nM)
clogD7.4
Guinea-pig tracheal strip
OoA (min)a
IC50 (nM)
DoA (h)b
Indacaterol
Formoterol
Salmeterol
2
3
20.6 2.8
2.6 0.1
3.2 1.0
2.5 0.1
0.8
91
4
1048
1626
1049
895
782
1141
842
745
1100
64
512
120
512
575
279
502
477
—
3.0
1.3
3.1
2.4
3.0
3.0
3.0
3.0
3.6
3.6
2.5
3.5
2.6
3.0
3.1
3.1
3.6
3.0
3.2
3.2
3.6
3.7
3.1
3.1
2.6
7.9
0.4
3.5
2.2
—
35
28
120
36
—
>10
1.2
>10
>10
—
315 162
784 452
70 27
6.9
4
0.6
18
—
—
—
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
0.8 0.3
6.7 3.1
2.0 1.6
97 51
4.2 2.6
0.4 0.1
5.5 2.9
7.1 1.8
15 1.4
149 19
12 1.7
266 63
19 5.4
3.2 1.1
6.8 4.3
1.4 0.6
3.9 2.0
1.7 0.7
32 18
17 3.3
66 19
40 16
391 271
62 26
0.8 0.3
30 8.6
58 24
77 35
343 138
33 10
438 74
28 13
5.5 4.0
6.1 1.7
7.9 1.8
13 0.8
0.3
23
5.7
—
<0.3
—
12
—
22
—
20
—
14
12
—
6.4
—
8.5
—
87
66
61
—
47
—
33
—
32
—
68
—
36
32
—
54
—
33
—
>10
>10
>10
—
>10
—
5.5
—
3.7
—
8.6
—
6.4
8.9
—
9.0
—
9.6
—
149
152
115
73
129
144
—
4.3
21 8.0
0
a
Onset of action: time to maximal effect at the IC50 dose level.
Duration of action: time for response to fall to 50% of maximal response during washout phase at the IC50 dose level.
b
facilitate formulation for inhalation.8,13 Based upon the targeted
profile, the most interesting a-substituents with respect to the tar-
geted b2AR potency were the (S,S)-cyclopentyl 18, (R,R)-cyclohexyl
19, (R,R)-isopropyl 21 and (R,R)-phenyl 23. The analogues 17 and
22 were considered to be too highly potent based upon the above
criteria and analogue 20 was deemed not to be sufficiently potent.
For the b-phenethyl series 6: propyl, butyl and n-butoxy resi-
dues in the meta- and para-positions, as exemplified by com-
pounds 29–34, were well tolerated and provided analogues in
the desired lipophilicity range. These analogues exhibited b2AR
potencies at the higher potency limit of the targeted profile, and
remained within 4-fold of the previously reported unsubstituted
parent compound.8 However, introducing two substituents at both
the meta- and para-positions, as exemplified by the 3,4-dimethyl
analogue 35, led to 27-fold decrease in potency relative to the
unsubstituted parent compound.
Exploring the possibility for substitution in the a-phenyl substitu-
ent of 23 to further increase lipophilicity within the targeted
range: methyl and ethyl groups in the ortho and meta-positions
were well tolerated as exemplified by 24, 25 and 27. In contrast
introducing a methyl substituent into the para-position, 26, led
to a 27-fold decrease in potency, and as a result substituents in this
position were not pursued further. Similarly meta-alkoxy substitu-
ents led to an even greater decrease in potency as exemplified by
the ethoxy analogue 28.
In terms of selectivity for the b2AR over the b1AR, the a-substi-
tuted cyclopentyl analogues 5 exhibited a clear trend for greater
selectivity in comparison to the b-phenethyl analogues 6 (mean
selectivities of 11-fold versus 2.3-fold for 5 verses 6). However,
some overlap in selectivity was observed between the series. When

N. Arnold et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4341–4347
4345
compared to the reference compounds the majority of the
4-hydroxybenzothiazolone analogues exhibited a greater selectiv-
ity than determined for indacaterol, and this was used as the basis
for further progressing compounds. Employing these criteria the
analogues 27, 29–31, 33 and 34 were not pursued further due to
concerns over the increased potential for producing b1AR mediated
side effects.
undesirable b2AR mediated systemic side-effects.8 Measuring the
in vitro glucuronidation rates for the series 5 and 6 showed high
intrinsic clearance values for the majority of examples in incuba-
tions in microsomes with the uridine 50-diphospho-glucuronic acid
(UDPGA) cofactor. However, assessing in more detail the overall
metabolic profiles of selected compounds revealed complex pat-
terns, thus making prediction of the likely exposure to parent
and b2AR active metabolites more difficult. As a result the decision
was made to assess the tolerability of the compounds in vivo at the
earliest opportunity, rather than extensive early in vitro pharmaco-
kinetic profiling, to select compounds with the most favourable
side effect profiles.
Assessing the data for series 5 and 6 up to this point led to com-
pounds 2 and 34 being identified as the most promising examples,
and prior to in vivo profiling both compounds were further
screened more broadly for off-targets and to generate further com-
parative data with indacaterol. These data showed 2 to have an
acceptable off-target profile when screened against the broad panel
of assays and in follow up functional screens. In contrast, compound
34 produced a weakly positive result in a micronucleus screen
which stopped the further progression of this compound.19 Evaluat-
ing the remaining compound, 2 was further assessed in comparison
with indacaterol in physicochemical and pharmacokinetic studies,
selected data are shown in Table 2. Overall these data showed
compound 2 to possess a very similar profile to indacaterol. In a
functional cAMP assay in A431 cells, endogenously expressing the
b2AR, 2 was found to be 6-fold more potent than indacaterol and
of slightly higher intrinsic efficacy. High selectivities in functional
b1AR and D2R assays were observed for 2, which were comparable
In terms of selectivity for the b2AR over the DR family, again the
a-substituted cyclopentyl analogues 5 exhibited a clear trend for
greater selectivity in comparison to the b-phenethyl analogues 6
(mean selectivities of 214-fold versus 40-fold for 5 vs 6). When
compared to the reference compounds the more selective exam-
ples from the 4-hydroxybenzothiazolone analogues such as 2,
17–19, 21–24, 32 and 34 were comparable to, or better than,
indacaterol in terms of relative affinities for the b2AR verses the
dopamine D3 receptor (D3R). To assess the functional conse-
quences, a dopamine D2 receptor (D2R) cellular assay revealed
some of the analogues 6 to be agonists of intermediate intrinsic
efficacy at this receptor, as exemplified by 34, for which an EC50
of 123 nM was determined with a maximal effect of 43% relative
to the reference agonist dopamine.17 In terms of applying criteria
for progressing compounds, a cut off of greater than 30-fold selec-
tivity was implemented. This lead to the analogues 20, 25, 26, 29,
and 31 being deprioritised due to concerns over the potential for
producing DR mediated side effects.
In the next stage of evaluation the most promising compounds
were tested in an electrically-stimulated guinea-pig tracheal strip
assay. A number of the
a-substituted cyclopentyl analogues 5
exhibited a longer time to maximal effect which extended well
beyond the end of the 30 min drug administration phase. As a
result the analogues 17–19, 21 and 27 were deprioritised due to
the risk of a delayed onset of action. This differentiation was based
upon the delayed OoA observed in man with the twice-daily LABA
salmeterol and the corresponding delayed OoA profile measured in
to those observed with indacaterol. One property where
a
difference was seen for 2 was the 100-fold higher solubility when
compared to indacaterol. This increased solubility was considered
to be potentially beneficial, and seen as a way to accelerate dissolu-
tion to ensure a rapid OoA following dry-powder delivery. LogD7.4
and affinity for immobilised artificial membrane (KIAM) measure-
ments showed 2 to be slightly less lipophilic than indacaterol.
Comparison of the calculated with the measured logD7.4 values
shows the prediction to be greater than a log unit higher in both
cases. Similar ionisation constants for both compounds also indi-
cated 2 to be predominantly zwitterionic at physiological pH, a
property which has been linked to a favourable colocalisation of
indacaterol with the b2AR in lipid raft micro domains.20 In terms
of the pharmacokinetic profiles, comparable plasma protein
binding and rat clearance and volume of distribution data provided
further confidence that 2 would exhibit a comparable LABA profile
to indacaterol, based upon the previously hypothesised roles of
these parameters in satisfying the targeted profile.13
The similar profile of 2 in comparison with indacaterol led to
the compound entering into in vivo profiling to assess DoA and tol-
erability. Previously we have shown an anti-bronchoconstrictor
model in the rhesus monkey to be a good predictor of intrinsic
DoA and tolerability using clinically established b2AR agonists,
and also as a key model in the selection of indacaterol.14,21 Com-
pound 2 was initially tested in a dose response study from which
the above tracheal strip assay.18 In contrast the
a-phenyl cyclopen-
tyl analogues 23 and 25 produced times to maximal effect consis-
tent with a rapid OoA. However, the DoA for both these compounds
were shorter than for indacaterol and suggested they would not
satisfy the targeted once-daily profile. Therefore, the only analogue
from the series 5 satisfying the targeted OoA and DoA profile was
the original library member 2, which showed a very similar profile
to indacaterol. From the b-phenethyl analogues 6 the less lipophilic
analogues 29, 30 and 34 produced times to maximal effect consis-
tent with a rapid OoA. In contrast the more lipophilic analogue 32
resulted in an extended time to maximal effect and was depriori-
tised due to OoA concerns. Of the series 6 with satisfactory OoA
profiles, compounds 30 and 34 also showed DoA profiles approach-
ing indacaterol and were selected as the most promising com-
pounds. Although lipophilicity had proven to be a good predictor
of OoA and DoA previously in the tracheal-strip assay, we have also
discussed the role of potency and intrinsic efficacy on these param-
eters.18 In that context all the analogues 5 and 6 tested in the
tracheal-strip assay were able to function as full agonists at the
highest concentrations tested and demonstrated high levels of
potency. Thus, although a clear trend is evident for increased
OoA and DoA with increased lipophilicity for the series 5 and 6,
the reason for the relatively low success rate in identifying
compounds with an overlapping rapid OoA and suitably long
DoA are unclear at the present time when compared to previous
applications of this approach.
an ED80 of 2.6 lg/kg was determined for the inhibition of a meth-
acholine-challenge administered 5 min after the end of the drug
administration period. To have the highest confidence in the key
time course arm of the study, 2 was dosed head-to-head with ind-
acaterol. Both compounds were administered at the ED80 dose
level, the indacaterol dose being taken from the previously
To achieve the tolerability component of the targeted profile,
the 4-hydroxybenzothiazolone series had in part been initially
selected due to exhibiting a high susceptibility towards glucuron-
idation. Such a clearance pathway was anticipated to provide a
means to rapidly eliminate the systemic component of the dose
in the form of a b2AR inactive metabolite, so as to minimise
reported dose response study (12.4 l
g/kg).14 Data from the time
course profiles for the inhibition of repeated methacholine chal-
lenges and for the changes in heart rate (HR) in the rhesus monkey
are shown in Figure 3.
From this rhesus monkey study, comparison of the indacaterol
data with the previously reported data for this compound shows

4346
N. Arnold et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4341–4347
Table 2
Comparison of the data for 2 with indacaterol
Assay/measurement
Compound 2
Indacaterol
Cellular b2AR: A50 (nM), % efficacya
0.25 0.05, 103.4 8.2%
2.2 1.3 ꢁ 10ꢀ6
>10ꢀ5
924
8.5, 7.5
1.25, 1.21
96.1%
79.8 3.6, 6.2 0.6
1.5, 91%
Functional b1AR: EC50
(
l
M)b
3.5 2.2 ꢁ 10ꢀ5
Functional D2R: A50 (M)17
>10ꢀ5
8
8.3, 6.7
1.77, 1.90
95.1–96.2%
66.9 7.5, 4.1 1.0
Solubility pH6.8 (
l
M)c
pKa: phenol, amined
e
LogD7.4, logKIAM
Plasma protein bindingf
Rat PK: Cl (l/kg), Vss (ml/min/kg)g
a
b
c
d
e
f
Elevation of cAMP measured in A431 cells endogenously expressing the b2AR, % efficacy relative to maximum response to formoterol.8
Inhibition of the electrically stimulated contraction of guinea-pig left atria.14
Thermodynamic solubility was measured using the shake vial method.14
pKa values were determined by potentiometric titration.14
Lipophilicity values were measured using chromatographic methods.14
Human protein binding was measured for 2 using an ultrafiltration method, and for indacaterol as previously reported.13
g
Pharmacokinetic data were generated in separate experiments following the i.v. administration of a 2.1 and 5.0 lmol/kg dose of 2 and indacaterol, respectively.
Figure 3. Comparison of 2 and indacaterol at the ED80 dose level in the methacholine-induced bronchoconstriction model in the rhesus monkey and the associated changes in
heart rate. Results are expressed as mean standard error of the mean of the number of animals shown in parentheses. ⁄P <0.05 and ⁄⁄P <0.001 indicates that the value differs
significantly from the equivalent value in the vehicle treated animals. The following parameters were also measured with no significant differences observed versus the
vehicle treated controls at each time point: systolic and diastolic arterial blood pressure; serum potassium.
an excellent level of reproducibility with this model.14 When
administered at equipotent bronchodilating doses 2 and indacater-
ol showed very similar levels of antibronchoconstrictor activity at
all time points, indicating both compounds to have very similar
intrinsic DoA and for 2 to be 5-fold more potent. In contrast, the
associated b2AR systemic side effects, as assessed by increased
HR, remained significantly higher for 2 than for indacaterol. How-
ever, when compared to the historical data generated in the rhesus
monkey with the well-tolerated twice-daily b2AR formoterol at the
same dose level (ED80):20 2 shows an improved side effect profile
and greater intrinsic duration of action. Based upon these data 2
was considered to be capable of delivering an equivalent DoA to
indacaterol and to have a good tolerability profile with respect to
the associated systemic b2AR mediated side effects.
Taking the above data set into consideration 2 was considered
to possess a very similar profile to indacaterol and to satisfy all
aspects of the targeted product profile. As a result 2 was promoted
as the second compound to progress into development from the
project, as the backup candidate to indacaterol. To prepare the
material to support the early development activities the synthesis
described in Scheme 1 was used with the commercially available
primary amine (1S,2S)-2-benzyloxycyclopentylamine.22
In summary, by following up on the two most promising leads
from a 4-hydroxybenzothiazolone b2AR agonist series a number of
analogues satisfying the activity and selectivity criteria, whilst
falling within the targeted physical property space, have been
identified. Upon further profiling 7-[(R)-2-((1S,2S)-2-benzyloxy-
cyclopentylamino)-1-hydroxyethyl]-4-hydroxybenzothiazolone,
compound 2, a member of the original N-substituent screening
library, was identified as the optimal compound from the series.
The data supported 2 to be: 5- to 6-fold more potent than indaca-
terol; to possess a good selectivity profile; to combine a rapid OoA,
with a comparable intrinsic DoA to indacaterol; and to have a
favourable tolerability profile with respect to the systemic b2AR
mediated side effects that are associated with the targeted levels
of bronchodilation. On the basis of these results 2 was progressed
into development as the backup LABA candidate to indacaterol.
References and notes
1. (a) Jacobsen, J. R. Future Med. Chem. 2011, 3, 1607; (b) Cazzola, M.; Calzetta, L.;
Matera, M. G. Br. J. Pharmacol. 2011, 163, 4.
2. More recent disclosures of new LABAs: (a) Procopiou, P. A.; Barrett, V. J.;
Biggadike, K.; Butchers, P. R.; Craven, A.; Ford, A. J.; Guntrip, S. B.; Holmes, D. S.;
Hughes, S. C.; Jones, A. E.; Looker, B. E.; Mutch, P. J.; Ruston, M.; Needham, D.;

N. Arnold et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4341–4347
4347
Smith, C. E. J. Med. Chem. 2014, 57, 159; (b) Kempsford, R.; Norris, V.; Siederer,
S. Pulm. Pharmacol. Ther. 2013, 14, 89; (c) Jacobsen, J. R.; Choi, S. K.; Combs, J.;
Fourmier, E. J. L.; Klein, U.; Pfeiffer, J. W.; Thomas, G. R.; Yu, C.; Moran, E. J.
Bioorg. Med. Chem. Lett. 2012, 22, 1213; (d) Procopiou, P. A.; Barrett, V. J.; Ford,
A. J.; Looker, B. E.; Lunniss, G. E.; Needham, D.; Smith, C. E.; Somers, G. Bioorg.
Med. Chem. 2011, 19, 6026; (e) Procopiou, P. A.; Barrett, V. J.; Bevan, N. J.;
Butchers, P. R.; Conroy, R.; Emmons, A.; Ford, A. J.; Jeulin, S.; Looker, B. E.;
Lunniss, G. E.; Morrison, V. S.; Mutch, P. J.; Perciaccante, R.; Ruston, M.; Smith,
C. E.; Somers, G. Bioorg. Med. Chem. 2011, 19, 4192.
12. Alikhani, V.; Beer, D.; Bentley, D.; Bruce, I.; Cuenoud, B. M.; Fairhurst, R. A.;
Gedeck, P.; Haberthuer, S.; Hayden, C.; Janus, D.; Jordan, L.; Smithies, K.;
Wissler, E. Bioorg. Med. Chem. Lett. 2004, 14, 4705.
13. Baur, F.; Beattie, D.; Beer, D.; Bentley, D.; Bradley, M.; Bruce, I.; Charlton, S. J.;
Cuenoud, B.; Ernst, R.; Fairhurst, R. A.; Faller, B.; Farr, D.; Keller, T.; Fozard, J. R.;
Fullerton, J.; Garman, S.; Hatto, J.; Hayden, C.; He, H.; Howes, C.; Janus, D.; Jiang,
Z.; Lewis, C.; Loeuillet-Ritzler, F.; Moser, H.; Reilly, J.; Steward, A.; Sykes, D.;
Tedaldi, L.; Trifilieff, A.; Tweed, M.; Watson, S.; Wissler, E.; Wyss, D. J. Med.
Chem. 2010, 53, 3675.
3. Recent disclosures of 4-hydroxybenzothiazolone based LABAs: (a) Stocks, M. J.;
Alcaraz, L.; Bailey, A.; Bonnert, R.; Cadogan, E.; Christie, J.; Dixon, J.; Connolly,
S.; Cook, A.; Fisher, A.; Flaherty, A.; Humphries, A.; Ingall, A.; Jordan, S.; Lawson,
M.; Mullen, A.; Nicholls, D.; Paine, S.; Pairaudeau, G.; Young, A. Med. Chem. Lett.
2014, 5, 416; (b) Alcaraz, L.; Bailey, A.; Cadogan, E.; Connolly, S.; Jewell, R.;
Jordan, S.; Kindon, N.; Lister, A.; Lawson, M.; Mullen, A.; Dainty, I.; Nicholls, D.;
Paine, S.; Pairaudeau, G.; Stocks, M. J.; Thorne, P.; Young, A. Bioorg. Med. Chem.
Lett. 2012, 22, 689; (c) Nicholls, D. J.; Jordan, S.; Cadogan, E.; Lawson, M.; Austin,
R. P.; Paine, S. W.; Gardiner, P.; Bonnert, R.; Connolly, S.; Young, A. Pulm.
Pharmacol. Ther. 2012, 25, 293; (d) Connolly, S.; Alcaraz, L.; Bailey, A.; Cadogan,
E.; Christie, J.; Cook, A. R.; Fisher, A. J.; Hill, S.; Humphries, A.; Ingall, A. H.; Kane,
Z.; Paine, S.; Pairaudeau, G.; Stocks, M. J.; Young, A. Bioorg. Med. Chem. Lett.
2011, 21, 4612; (e) Stocks, M. J.; Alcaraz, L.; Bailey, A.; Bonnert, R.; Cadogan, E.;
Christie, J.; Connolly, S.; Cook, A.; Fisher, A.; Flaherty, A.; Hill, S.; Humphries, A.;
Ingall, A.; Jordan, S.; Lawson, M.; Mullen, A.; Nicholls, D.; Paine, S.; Pairaudeau,
G.; St-Gallay, S.; Young, A. Bioorg. Med. Chem. Lett. 2011, 21, 4027.
14. Battram, C.; Charlton, S. J.; Cuenoud, B.; Dowling, M. R.; Fairhurst, R. A.; Farr, D.;
Fozard, J. R.; Leighton-Davies, J. R.; Lewis, C. A.; McEvoy, L.; Turner, R. J.;
Trifilieff, A. J. Pharmacol. Exp. 2006, 371, 762.
15. Weinstock, J.; Gaitanopoulos, D. E.; Stinger, O. D.; Franz, R. G.; Hieble, J. P.;
Kinter, L. B.; Mann, W. A.; Flaim, K. E.; Gessner, G. J. Med. Chem. 1987, 30, 1166.
16. DR affinity was determined by the ability of the test compounds to displace
radiolabelled (RS)-trans-7-hydroxy-2-[N-propyl-N-(30-iodo-20-propenyl)amino]tetralin
(
125I-7-HO-PIPAT) from membranes derived from HEK cells over-expressing
the human D3R. All the data in Table
determinations.
1 are means from duplicate
17. Functional DR activity was measured using a radiolabelled [35S]GTP
c
S assay
with membranes expressing the human D2R derived from Chinese Hamster
Ovary cells. Data are expressed as the % response relative to 100
minus the basal activity.
lM dopamine
18. Trifilieff, A.; Charlton, S. J.; Fairhurst, R. A. The Preclinical Pharmacology of
Indacaterol. In Indacaterol: the First Once-daily Long-acting Beta2 Agonist for
COPD. Milestones in Drug Therapy series; Trifilief, A., Ed.; Springer: Basel, 2014.
Chapter 2.
4. (a) Norman, P. Expert Opin. Investig Drugs 2013, 22, 1569; (b) Cazzola, M.;
Calzetta, L.; Matera, M. G. Br. J. Pharmacol. 2011, 163, 4.
5. http://www.novartis.com/newsroom/media-releases/en/2011/1528013.shtml.
6. http://www.novartis.intra/news/media-releases/en/2013/09/2013-09-23_1730580.
shtml.
7. Fairhurst, R. A.; Charlton, S. J.; Trifilieff, A. The Design of the Indacaterol
Molecule. In Indacaterol: the First Once-daily Long-acting Beta2 Agonist for COPD.
Milestones in Drug Therapy series; Trifilief, A., Ed.; Springer: Basel, 2014.
8. Beattie, D.; Bradley, M.; Brearley, A.; Charlton, S. J.; Cuenoud, B. M.; Fairhurst, R.
A.; Gedeck, P.; Gosling, M.; Janus, D.; Jones, D.; Lewis, C.; McCarthy, C.; Oakman,
H.; Stringer, R.; Taylor, R. J.; Tuffnell, A. Bioorg. Med. Chem. Lett. 2010, 20, 5302.
9. Fairhurst, R. A.; Janus, D.; Lawrence, A. Org. Lett. 2005, 7, 4697.
19. Compounds were incubated for 20 h with a mouse lymphoma cell line
(L5178Y) with and without metabolic activation with rat S9. An increase in the
number of cells containing micronuclei, occuring below the limit of
cytotoxicity or up to the limit of solubility, was considered a positive result
for clastogenic and/or aneugenic behaviour. Further examples from the b-
phenethyl series 6 also produced positive results in this assay.
20. Lombardi, D.; Cuenoud, B.; Krämer, S. D. Eur. J. Pharm. Sci. 2009, 38, 533.
21. (a) Fozard, J. R.; Buescher, H. Drug Dev. Res. 2002, 57, 1; (b) Fozard, J. R.;
Buescher, H. Pulm. Pharmacol. Ther. 2001, 14, 289.
22. (1S,2S)-2-Benzyloxycyclopentylamine is available commercially in high optical
purity (>98.5% ee) as part of the BASF ChiPros range (CAS No: 181657-57-8)
Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Kebler, M.; Stürmer, R.; Zelinski,
T. Angew. Chem., Int. Ed. 2004, 43, 788.
10. Hartmann, S.; Ullrich, S.; Hupfer, C.; Frahm, A. W. Eur. J. Med. Chem. 2000, 35,
377.
11. Beattie, D.; Cuenoud, B.; Fairhurst, R. A.; Taylor, R. J. Benzothiazolone
derivatives having b2-adrenoceptor activity. WO 2004/016601, Feb. 26, 2004.