SAR of a series of inhaled A2A agonists and comparison of inhaled pharmacokinetics in a preclinical model with clinical pharmacokinetic data

Article abstract of DOI:10.1016/j.bmcl.2009.05.027

COPD is a major cause of mortality in the western world. A_2A agonists are postulated to reduce the lung inflammation that causes COPD. The cardiovascular effects of A_2A agonists dictate that a compound needs to be delivered by inhala

Full text of DOI:10.1016/j.bmcl.2009.05.027

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4471–4475
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
SAR of a series of inhaled A2A agonists and comparison of inhaled
pharmacokinetics in a preclinical model with clinical pharmacokinetic data
*
Simon J. Mantell , Peter T. Stephenson, Sandra M. Monaghan, Graham N. Maw, Michael A. Trevethick,
Michael Yeadon, Don K. Walker, Matthew D. Selby, David V. Batchelor, Stuart Rozze, Helene Chavaroche,
Arnaud Lemaitre, Karen N. Wright, Lynsey Whitlock, Emilio F. Stuart, Patricia A. Wright, Fiona Macintyre
Pfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Kent, CT13 9NJ, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 26 February 2009
Revised 7 May 2009
Accepted 8 May 2009
Available online 12 May 2009
COPD is a major cause of mortality in the western world. A2A agonists are postulated to reduce the lung
inflammation that causes COPD. The cardiovascular effects of A2A agonists dictate that a compound needs
to be delivered by inhalation to be therapeutically useful. The pharmacological and pharmacokinetic SAR
of a series of inhaled A2A agonists is described leading through to human pharmacokinetic data for a clin-
ical candidate.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
A2A
COPD
Pharmacokinetics
Inhalation
The administration by inhalation of therapeutic agents that
have the lung as their target organ is a well-accepted clinical prac-
tice. For example, the majority of currently prescribed asthma
treatments, for example, b-agonists and inhaled corticosteriods,
are all delivered by the inhaled route.1 Chronic Obstructive Pul-
monary Disease (COPD) affects 10–24 million adults in the USA.2
This disease is characterised by chronic cough, mucus hypersecre-
tion, breathlessness and a gradual decline in lung function. Neutro-
phils are implicated in the initiation, maintenance and
symptomatology of COPD,3 and are believed to play an important
pathophysiological role.4 Bronchial neutrophilia is the most signif-
icant cellular change in the disease, which correlates with airflow
obstruction. Furthermore the most important risk factor for the
development of COPD is cigarette smoking. However, since sys-
temically available A2A agonists were known to induce cardiovas-
Ph
HN
Ph
N
N
N
H
N
H
N
O
N
H
N
N
O
O
HO
2
OH
A2a IC50 6nM
Figure 2. Structure of urea containing lead.
O
OH
NH2
NH2
N
OH
N
N
N
Et
N
O
N
N
N
H
N
N
H
N
N
N
N
H
O
O
N
1
1a
GW 328267
CGS 21680
OH
OH
OH
OH
Figure 1. The structure of CGS 21680 and GW 328267.
* Corresponding author. Tel.: +44 01304 641375.
E-mail address: simon.mantell@pfizer.com (S.J. Mantell).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.05.027

4472
S. J. Mantell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4471–4475
Ph
Ph
HN
N
HN
N
Ph
O
N
N
Ph
N
N
N
OH
O
N
N
H
N
O
H
N
H
N
N
H
N
O
O
HO
OH
HO
A2A IC50 17nM
OH
A2A IC50 65nM
3
4
Ph
HN
Ph
N
Y = CONHEt
2A IC50 5nM
N
N
O
5
A
H
N
N
Y
N
H
N
H
N
O
Y = CH2OH
A2A IC50 45nM
6
O
HO
OH
Figure 3. Combining purine linked amide and purine linked methylene urea A2A agonist series.
cular effects, in particular hypotension, in laboratory animals,5 in-
haled administration of the agent was targeted from the outset.
Intratracheal administration6 of the adenosine A2A agonist, CGS
21680 1 inhibited lung inflammation in the rat but was associated
with falls in blood pressure.7 This has been supported by recent
clinical trial data with the inhaled adenosine A2A agonist GW
328267 1a which caused hypotension and tachycardia at doses
that limited clinical utility.8
By optimisation of the pharmacokinetic and physiochemical
properties of the compounds in order to minimise systemic
Table 1
SAR of urea containing A2A agonists
Ph
HN
Ph
H
N
N
N
N
O
N
H
N
X
O
O
HO
OH
12
Compound
X=
A2a IC50
O
7
7
N
N
H
N
H
O
8
9
5
N
N
N
Figure 4. Duration of action of 15.
H
H
O
Comparison of the effects of the adenosine A2a agonists (15
10
N
N
H
N
and GW328267 1a ) on diastolic blood pressure of
H
anaesthetised guinea pigs.
O
N
10
11
12
13
305
10
6
125
100
75
N
N
H
Me
O
N
H
N
H
50
O
O
300 µg GW328267 1a
10µg 15
micelle vehicle
N
H
N
H
25
0
0
100
200
300
400
4
N
H
N
H
Time (mins) after test compound
All data are mean + SEM n=4
N
Figure 5. Side effect liability of 15.

S. J. Mantell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4471–4475
4473
tendency of a number of structures related to 2 to absorb carbon
dioxide was evaluated. This study consisted of trying to form the
complex by washing a solution of the uncomplexed compound in
an organic solvent with saturated aqueous sodium hydrogen car-
bonate solution, drying the organic phase with anhydrous magne-
sium sulphate, filtering and solvent removal. The 1H NMR was
inspected to determine whether it was different to a spectrum
taken immediately prior to washing with the sodium hydrogen
carbonate solution. These fairly crude experiments indicated that
the presence of the urea linked through to a basic nitrogen by a
two carbon chain led to carbon dioxide complexation. As described
previously,9 the SAR of the series from which 2 is derived is fairly
tight, altering the urea portion or removing the basicity of the
nitrogen loses activity. Therefore, it became necessary to look at
changing 2 more substantially in order to move away from the car-
bon dioxide absorption properties. Previous SAR9 had also shown
that potent compounds could be obtained if a basic substituent
was linked through to the purine via an amide 2. Therefore, this
modification was incorporated into the series containing a urea
in the side chain (Figs. 1–5).
Table 1 (continued)
12
Compound
X=
A2a IC50
O
14
15
5
N
H
N
H
N
O
O
N
N
O
4
N
H
N
H
N
H
N
H
16
0.4
OH
exposure after delivery by inhalation, we hoped to identify an
efficacious compound with an acceptable therapeutic index. In a
previous publication in this journal we described our strategy for
achieving lung selectivity and the SAR of a series of A2A agonists,
leading up to the discovery of 2.9
Adding together the purine sidechains in 4 and 3 gave two
potent compounds 5, where Y = CONHEt, and 6, where Y = OH.
Compound 5 was significantly more active than 6 so the series
Y = CONHEt was pursued. 5 and all the others in Table 1 were pre-
pared using the synthetic route in Scheme 1.11
Compound 2 showed high potency and projected selectivity for
lung over cardiovascular effects. However, when a solution of the
compound in an organic solvent was washed with a saturated
aqueous solution of sodium hydrogen carbonate 2 appeared to
form a carbonate adduct that was stable to the buffer used in our
human neutrophil assay and to dosing to our in-vivo model.10 This
complex had a lower affinity for the A2A receptor than the parent.
Since the compound will be exposed to carbon dioxide in the lung
the observation of adduct formation meant that 2 could not be pro-
gressed as a clinical candidate as it would be very difficult to esti-
mate doses required for efficacy. Thus, our A2A SAR exploration
continued with the objective of finding a compound with a similar
pharmacological and pharmacokinetic profile to 2 but without the
problem of carbon dioxide complexation. The first step in achiev-
ing these objectives was to determine which of the structural fea-
tures was responsible for carbon dioxide absorption. Therefore the
A series of compounds where the urea containing sidechain of 5
was modified were prepared.
Incorporation of the urea containing sidechain that gave good
activity in 2 to give 7 gave similar potency. The comparable activity
of 7 to 9 demonstrates that there is a substantial amount of flexi-
bility in what substituents on the basic nitrogen will give activity.
However, methylation of the urea portion of 5 to give 10 causes a
drop off in activity. With the knowledge that there is flexibility in
the SAR of the basic nitrogen substituents but that this may be
more limited in the urea portion, a series of compounds that did
not contain the basic centre in 5 were synthesised. The benzyl 12
and phenethyl 13 derivatives were synthesized. The benzylic
derivative 12, was found to be slightly more potent. The close in
SAR around the benzylic substituent of 12 was then examined.
Replacement of the phenyl in 12 with 2-pyridyl to give 13
Ph
O
HN
Ph
N
N
Ph
HN
N
H
O
N
Ph
O
N
N
N
N
H
O
O
O
O
OAc
OBz
N
H
N
O
i
OBz
OH
OH
Ph
ii
HN
N
Ph
Ph
HN
N
N
N
O
N
Ph
R1
N
O
N
N
R1
N
O
N
N
N
X
NR3R4
H
N
H
O
N
X
O
O
NRR2
O
O
NR2
iii
HO
OH
HO
OH
NR3R4
Where R1=R2 then R=H, if R1=R2 then R=protecting group.
Scheme 1. Reagents and conditions: (i) N,O-Bis(trimethylsilylacetamide, then purine and TMSOTf, toluene, reflux, 2.5 h; (ii) sodium carbonate, methanol, rt, 4 h; (iii) HNR1–
X–NRR2, 105 °C; (iv) toluene/isopropyl alcohol 4:1, reflux.

4474
S. J. Mantell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4471–4475
Table 2
Potency, molecular properties and side affect liability for analogues
Rat Club (ml/min/kg)
TPSAc
Log D pH 7.4
MW
Free Cmax (nM)
a
d
Compound
A2A IC50 (nM)
HLM Clint
(ll/min/mg)
2
8
15
6
5
4
135
81
>255
3500
750
5500
179
208
221
2.0
0.9
3.4
713
755
778
1.8
12.3
0.2
a
Human Liver Microsome (HLM) stability given as intrinsic clearance (Clint).
Rat unbound clearance from IV bolus PK at 1 mg/kg.
Topological polar surface area.
b
c
d
Free Cmax was dose normalised to 1 mg/kg IT, all compounds were delivered as solutions.
in the series profiled 8, was found to have a higher Cmax than 2. This
was concerning but consistent with our previous observations9
that Cmax is at least partly driven by unbound clearance (clear-
ance/free fraction). The unbound clearance for 8 is 750 ml/min/
kg, approximately fivefold lower than for 2 and the Cmax of 8 is
approximately fivefold higher. Comparison of 2 with compound
15 also shows that increasing unbound clearance reduces Cmax
but there appears a much greater reduction in Cmax than would
be expected from the increase in unbound clearance, there is only
a 1.6-fold increase in unbound clearance but a ninefold drop in the
Cmax. This shows that it is unlikely that increasing clearance is the
sole contributor to reducing Cmax for compound 15 and that
increasing the lipophilicity and molecular weight of the molecules
is impacting on other factors that affect Cmax, most likely slower
transfer from lung to blood. Contribution to systemic levels of
the compounds in Table 2 by absorption from the gut is highly
unlikely given the very high Topological Polar Surface Area (TPSA)
of the compounds.17 The passive transcellular diffusion rate of
compound 15 was found to be negligible in a Parallel Artificial
Membrane Permeation assay, further supporting our hypothesis
that none of the compounds in Table 2 would be significantly
absorbed from the gut. 15 was taken forward into phase I studies
in man.
Table 3
Pharmacokinetics properties of compound 15 in healthy male humans
a
Dose (
l
g)
n
Cmax (nM)
100
200
200
400
800
7
3
6
9
10
7
0.45
0.30
0.30
0.19
0.19
0.12
1600
a
Free Cmax was dose normalised to 1 mg/kg.
increased potency slightly, as did addition of a basic substituent
onto the phenyl ring to give 14. Addition of a carboxylic acid to
the phenyl ring in 12 to give 16 increased potency 15 fold. Incorpo-
ration of a piperidylpyridine sidechain gave a potent compound,
15. The 2-aminopyridine in 15 is basic enough to allow salt
formation and hence aid solid form selection but because of
conformational restriction of the sidechain it was hypothesised
that it might not form a complex with carbon dioxide. 15 was
found not to complex carbon dioxide and was progressed to
in vivo studies. In order to compare the duration of action of a
compound in the lung with its activity in the systemic circulation
a model where A2A activity in the lung and the systemic circulation
could be determined simultaneously was required. Unfortunately,
due to differences in A2A receptor pharmacology between man and
common laboratory animals, the development of such an assay
with a neutrophil-dependant endpoint was not possible. An alter-
native model was thus adopted based on the ability of intrathecally
dosed A2A agonists to inhibit intravenously dosed capsaicin
induced bronchoconstriction in the anaesthetised guinea pig.13
Compound 14 showed a >5 h duration in the capsaicin induced
bronchoconstriction model at the same dose that showed no
effects on diastolic blood pressure after intrathecal administration
in the guinea pig model.
This is a significantly extended duration of action compared to
salmeterol and indicates that the compound could be expected to
be dosed twice daily in the clinic at worst and most likely once a
day. It is unclear why the duration of action is extended. Possible
factors include compound onset/offset kinetics,14 an exosite bind-
ing receptor as postulated for salmeterol,15 partitioning of lipo-
philic amines into the lipid bilayers of smooth muscle,16 poor
permeability from the lung into systemic circulation or slow disso-
lution of a compound that has crystallised after inhalation. The
overall systemic side-effect liability for an A2A inhaled agent is
thought to be dependent upon the free plasma Cmax achieved fol-
lowing inhalation. Side effect liability was assessed by comparing
the compounds normalised free Cmax (i.e., free Cmax/dose) subse-
quently referred to as free Cmax in this text. This assumes that
the required dose will be proportional to the potency of the com-
pound, the free plasma concentrations will also scale proportion-
ally to this dose and hence the effect of potency on side effect
liability cancels out. As can be seen in Table 3 the first compound
Healthy male subjects (age 18–45 years) in 2 cohorts received
single escalating doses of 15 administered by an aerosol device over
the dose range 100–1600
determined at doses above 100
l
g. Plasma drug concentrations were
g with a validated analytical meth-
l
od using HPLC with mass spectrometric detection. Pharmacody-
namic measurements of haemodynamic parameters were
recorded up to 4 h post dose along with standard safety evaluation
at times during the study. Mean half-life values for 15 were in the
range 3–6 h across the dose range studied. Exposure increased with
increasing dose although in a subproportional manner (Table 3).
Cmax values were similar to those observed in the preclinical model.
There were no clinically significant effects on sitting blood pres-
sure measurements or peak expiratory flow rate. A dose dependent
increase in sitting pulse rate was observed between 200 and
1600 lg, these effects were generally not considered to be clini-
cally significant.
In conclusion, a series of potent A2A agonists have been identi-
fied. Compound 15 showed >8 h activity in a guinea pig lung based
duration of action model and low side effect liability in a rat car-
diovascular side effect liability model has been identified. Com-
pound 15 has been taken through to the clinic and at doses
predicted to have antineutrophil effects in the lung no clinically
significant side effects were observed.
Acknowledgements
The authors of this manuscript would like to thank Petra Chaffe,
Adrian Barnard, Gary Salmon, Rob Webster and Gill Allen for their
invaluable contributions to generating the data in this manuscript.
The senior author would like to highlight the identification of the
chemical series disclosed herein by Peter Stephenson and the

S. J. Mantell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4471–4475
4475
8. Trevethick, M. A.; Mantell, S. J.; Stuart, E. F.; Barnard, A.; Wright, K. N.; Yeadon,
M. Br. J. Pharmacol. 2008, 155, 463.
9. Mantell, S. J.; Stephenson, P. T.; Monaghan, S. M.; Maw, G. N.; Trevethick, M. A.;
Yeadon, M.; Keir, R. F.; Walker, D. K.; Jones, R. M.; Selby, M. D.; Batchelor, D. V.;
Rozze, S.; Chavaroche, H.; Hobson, T. J.; Dodd, P. G.; Lemaitre, A.; Wright, K. N.;
Stuart, E. F. Bioorg. Med. Chem. Lett. 2008, 18, 1284.
identification of compound 15 by Peter Stephenson and David
Batchelor as being critical contributions to this Letter.
References and notes
10. Ureas are known to form strong complexes with carboxylic acids in the solid
phase: Schlenk, H.; Holman, R. T. J. Am. Chem. Soc. 1950, 72, 5001; Smith, G.;
Baldry, K. E.; Byriel, K. A.; Kennard, C. H. L. Aust. J. Chem. 1997, 50, 727.
11. Synthetic procedures for all compounds are given in Mantell, S. J.; Stephenson,
P. T. WO2001094368.
1. Dalby, R.; Suman, J. Adv. Drug Delivery Rev. 2003, 55, 779.
2. Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2006. Global
Strategy for Diagnosis, Management and Prevention of COPD.
www.goldcopd.com.
3. Wise, R. A.; Tashkin, D. P. Am. J. Med. 2007, 120, S4.
4. Stockley, R. A. Quart. J. Med. 1995, 88, 141.
12. A2A Screening details in WO2002000676.
13. Morimoto, H.; Yamashita, M.; Imazumi, K.; Matsida, A.; Ochi, T.; Seki, N.;
Mizuhara, H.; Fujii, T.; Senoh, H. Eur. J. Pharmacol. 1993, 240, 121. All
experiments using animals were carried out in compliance with UK
Legislation and subject to local ethical review.
14. Disse, B.; Speck, G. A.; Rominger, K. L.; Witek, T. J.; Hammer, R. Life Sci. 1999, 64,
457.
5. Hutchison, A. J.; Webb, R. L.; Oei, H. H.; Ghai, G. R.; Zimmerman, M. B.;
Williams, M. J. Pharm. Exp. Ther. 1989, 251, 47.
6. The micelle vehicle used for dosing the compounds in the in-vivo model is
prepared by dissolving 1 mg of compound in 0.66 g of a 0.9% w/v aqueous
sodium chloride solution and 0.39 g of a solution that was prepared by mixing
11.55 g of glycerol formol, 0.56 g of benzyl alcohol and 8.97 g of 12-
hydroxystearate (BASF Aktiengesellschaft solutol HS15).
15. Nials, A. T.; Sumner, M. J.; Johnson, M.; Coleman, R. A. Br. J. Pharmacol. 1993,
108, 507.
7. Fozard, J. R.; Ellis, K. M.; Villela Dantas, M. F.; Tigani, B.; Mazzoni, L. Eur. J.
Pharmacol. 2002, 438, 183.
16. Anderson, G. P.; Linden, A.; Rabe, K. F. Eur. Respir. J. 1994, 7, 569.
17. Hou, T.; Wang, J.; Zhang, W.; Xu, X. J. Chem. Inf. Model. 2007, 47, 208.