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