Identification of pyridazino[4,5-b]indolizines as selective PDE4B inhibitors

Article abstract of DOI:10.1016/j.bmcl.2010.02.044

Substituted pyridazino[4,5-b]indolizines were identified as potent and selective PDE4B inhibitors. We describe the structure-activity relationships generated around an HTS hit that led to a series of compounds with low nanomolar affinity for PDE4B and high selectivity over the PDE4D subtype.

Full text of DOI:10.1016/j.bmcl.2010.02.044

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2163–2167
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of pyridazino[4,5-b]indolizines as selective PDE4B inhibitors
a,
a
a
b
Andrew F. Donnell a, , Paul J. Dollings , John A. Butera , Arlene J. Dietrich , Kerri K. Lipinski ,
*
*
Afshin Ghavami b, Warren D. Hirst b
a Worldwide Medicinal Chemistry, Pfizer Global Research and Development, CN 8000, Princeton, NJ 08543, USA
b Neuroscience Research Unit, Pfizer Global Research and Development, CN 8000, Princeton, NJ 08543, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Substituted pyridazino[4,5-b]indolizines were identified as potent and selective PDE4B inhibitors. We
describe the structure–activity relationships generated around an HTS hit that led to a series of com-
pounds with low nanomolar affinity for PDE4B and high selectivity over the PDE4D subtype.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 27 January 2010
Revised 8 February 2010
Accepted 8 February 2010
Available online 13 February 2010
Keywords:
Selective phosphodiesterase inhibitors
Pyridazino[4,5-b]indolizines
Hit-to-lead
Phosphodiesterase 4 (PDE4) is one of eleven PDE enzyme fami-
lies responsible for the regulation of intracellular levels of cyclic
adenosine monophosphate (cAMP) and cyclic guanosine mono-
phosphate (cGMP).1 PDE4 inhibition has been evaluated for the
treatment of depressive disorders.2 More recently, PDE4 inhibitors
have been explored for treating a wider range of central nervous
system (CNS) diseases including Parkinson’s disease,3 schizophre-
nia,4 and Alzheimer’s disease.5 However, clinical studies of the
PDE4 inhibitor rolipram were limited by side effects including nau-
sea and emesis that were thought to arise from inhibition of the
PDE4D subtype.6 Similarly, side effects have restricted the thera-
peutic index of the second-generation PDE4 inhibitors cilomilast
and roflumilast.7 Selective inhibition of the PDE4B subtype may
provide a means to achieve efficacy while potentially mitigating
these adverse events.8 Thus, we undertook a program directed to-
ward the discovery of PDE4B-selective compounds. Recent reports
highlighting chemotypes with varying levels of PDE4B selectivity
illustrate the interest in such an approach.9 Here, we report the
discovery of pyridazino[4,5-b]indolizines as a novel class of potent
and selective PDE4B inhibitors.
IC50 >100 l
M).11 Compound 1 also displayed reasonable biochem-
ical and physicochemical lead-like properties. These factors
prompted our use of 1 as a starting point in a hit-to-lead campaign
aimed at improving the affinity for PDE4B while maintaining or
elevating the level of selectivity over PDE4D. We targeted three
of the major regions of 1 for synthetic explorations: (1) the pen-
dant phenyl ring, (2) the pyrimidine tail-piece, and (3) the tricyclic
pyridazino[4,5-b]indolizine core.
HN
N
N
N
N
N
N
N
1
PDE4B
PDE4D
K
K
i = 2.60
i = 59.0
µ
µ
M
M
Selectivity (D/B) = 23
A high-throughput screen (HTS) of our in-house compound li-
brary identified the novel substituted pyridazino[4,5-b]indolizine
1. This compound exhibited high affinity in our PDE4B binding as-
We designed a series of substituted-phenyl analogs to generate
SAR around the phenyl ring. The most practical route for the syn-
thesis of these compounds incorporated the diversity at an early
stage, as shown in Scheme 1, by treatment of methyl indolizine-
2-carboxylate (2) with the appropriate substituted benzoyl chlo-
rides. Warming this mixture at 50 °C in the presence of base
cleanly effected acylation of the 1-position of the indolizine. The
resulting ketoesters 3a–f were cyclized by condensation with
hydrazine to provide, after chlorination with phosphorus oxychlo-
ride, the chloropyridazino[4,5-b]indolizines 4a–f. The side chain
say (Ki = 2.60
lM), and it displayed 23-fold selectivity over
PDE4D.10 Comparable results were obtained from a cell-based
adenylyl cyclase functional assay (PDE4B IC50 = 13.2 lM, PDE4D
* Corresponding authors.
E-mail addresses: donnela@wyeth.com (A.F. Donnell), dollinp@wyeth.com (P.J.
Dollings).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.02.044

2164
A. F. Donnell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2163–2167
Cl
We then examined the role of the pyrimidine tail-piece. Using
CO2Me
O
an analogous synthetic pathway, we prepared several analogs with
substitution on the pyrimidine. We also replaced the pyrimidine
with other groups. All such changes significantly reduced the
PDE4B activity (data not shown). Subsequently, we prepared an ar-
ray of amide analogs as shown in Scheme 2. We elected to focus on
the scaffold with the 3-fluorophenyl group because of its higher
selectivity in the first round of SAR. Thus, introduction of an N-
Boc-piperazine side chain onto the chloropyridazino[4,5-b]indoli-
zine 4d, followed by deprotection and acylation with a variety of
acid chlorides, provided the amides 8a–g.
CO2Me
N
N
a
b, c
N
N
N
R1
R1
2
3a—f
4a—f
HN
H2N
N
d, e
HN
NC
N
N
N
N
Br
N
N
The PDE4 binding results of this library are shown in Table 2.
The acetamide 8a retained the affinity of the lead, although this
was attenuated with larger alkyl groups (8b and 8c). The benzam-
ide 8d was tolerated, and the pyrrole amide 8e was less active. Pyr-
idyl groups (8f, 8g) imparted sub-micromolar affinity and much
greater selectivity.
5
6
N
N
N
N
N
f
N
Concurrently, we explored modifications of the core moiety,
and developed the synthesis of a series of bicyclic pyrrolo[3,2-
d]pyridazine analogs, in which the ‘A-ring’ of the lead tricycle
has been truncated (Scheme 3). This synthesis commenced with
the N-methylation of the known trisubstituted pyrrole 9, which
was prepared by the HCl-promoted cyclization of a ketonitrile.12
Hydrogenolysis of the chloro group gave 10. Regioselective lithia-
tion of this species allowed for the introduction of substituted ben-
zoyl groups at the 2-position. We included the moieties that had
shown the most favorable activity in the tricyclic series. Cycliza-
tion of the resultant ketoesters 11a–d to the pyrrolo[3,2-d]pyrida-
zines 12a–d was followed by installation of the side chain. This
chemistry was performed analogously to that for the tricyclic ana-
logs, providing the bicyclic analogs 13a–d.
4
6
N
R1
7a—f
Scheme 1. Reagents and conditions: (a) R1C6H4COCl, NaOAc, 50 °C; (b) hydrazine
hydrate, EtOH, reflux; (c) POCl3, reflux; (d) DMF, 60 °C; (e) H2, Raney-Ni, 2 M NH3–
EtOH, rt; (f) NH4Cl, NMP, 100 °C.
was constructed by alkylation of the piperazine 5 with 3-bromo-
propanenitrile, followed by reduction of the nitrile. The resulting
amine 6 was introduced onto 4a–f by nucleophilic aromatic substi-
tution, giving the targeted compounds 7a–f. Using similar chemis-
try, we also replaced the phenyl ring with heteroaromatic groups
and various alkyl groups.
The analogs with substitution at the 2- and 4-positions of the
phenyl ring had markedly lower PDE4B affinity. Likewise, replace-
ment of the phenyl group with heteroaromatic groups and alkyl
groups was not tolerated (data not shown). Substitution at the
3-position, particularly with electron-withdrawing groups, was
tolerated. Selected compounds are displayed in Table 1. The
methyl-substituted analog 7a was somewhat less active than the
lead 1. Analogs containing a halogen (7b–d) were essentially equi-
potent with 1, although PDE4B selectivity was eroded with the bro-
mo and chloro species. The fluoro analog 7d was the only
compound that displayed modestly higher selectivity than 1. Nota-
bly, the nitro-substituted analog 7f showed a 10-fold increase in
PDE4B binding affinity, but with a twofold decrease in selectivity.
Pyrrolo[3,2-d]pyridazine 13a, the direct bicyclic analog of the
lead tricycle 1 exhibited 10-fold less affinity for PDE4B (Table 3).
The loss of affinity was not as severe for the halogen- and nitro-
substituted analogs 13b–d, although they were slightly less active
than their tricyclic comparators. In contrast to the tricyclic analogs,
the bicycles, in particular the nitro species 13d, displayed high lev-
els of PDE4B-selectivity.
In a further iteration, we incorporated the pyridyl amide tail-
pieces onto this scaffold and more exhaustively explored this motif
with a variety of heteroaromatic amides. We retained the 3-nitro-
phenyl group in these studies because of the high PDE4B affinity
and selectivity shown by 13d. The synthesis of these compounds
was carried out analogously to that described in Scheme 2 for the tri-
cyclic species. We examined a range of amides, some of which are
displayed in Table 4. Among pyridines, the 4-pyridyl isomer 14a
had reduced affinity, while the 3- and 2-pyridyl amides 14b and
14c displayed both high affinity and selectivity. Five-membered ring
heterocycles (14d–f) showed varying levels of activity. Among dia-
zines, the pyrazine amide 14h showed the greatest affinity
(77 nM) and also exhibited 58-fold selectivity over PDE4D.
We then returned to the original tricyclic scaffold with this
optimized pyrazine amide tail-piece, and prepared the unsubsti-
Table 1
Substituted phenyl analogs
HN
N
N
N
N
N
N
N
R1
Cl
HN
N
R1
PDE4B Kia
(
lM)
PDE4D Kia
(lM)
Selectivityb (D/B)
N
Compd
N
N
N
N
R2
1
H
2.60
59.0
23
13
12
7
29
3
a, b, c
N
N
7a
7b
7c
7d
7e
7f
Me
Br
Cl
F
CN
NO2
4.30 2.77
1.44 0.65
2.01 1.10
2.70 1.14
1.03 0.38
54.4 6.2
17.5 9.2
14.2 2.9
83.0 2.8
3.35 1.20
2.40 0.14
F
F
4d
8a—g
0.210 0.099
11
a
Values that are the mean of two or more experiments are shown with their
Scheme 2. Reagents and conditions: (a) tert-butyl 4-(3-aminopropyl)piperazine-1-
carboxylate, NH4Cl, NMP, 100 °C; (b) HCl–MeOH, 60 °C; (c) R2COCl, i-Pr2NEt, THF,
0 °C to rt.
standard deviations.
b
Ratio of Ki (PDE4D)/Ki (PDE4B).

A. F. Donnell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2163–2167
2165
Table 2
Piperazine amide analogs
HN
N
N
N
N
R2
N
F
Compd
R2
PDE4B Kia
(l
M)
PDE4D Kia
(
lM)
Selectivityb (D/B)
8a
1.90
>100
>53
>20
O
O
8b
8c
4.90
14.0
>100
61.0
4
O
8d
8e
8f
2.70
5.20
0.413
22.2
65.0
19.0
8
O
O
12
46
N
N
N
O
8g
0.241
18.0
75
O
a
Values are the result of one experiment.
Ratio of Ki (PDE4D)/Ki (PDE4B).
a
Table 3
Bicyclic analogs
CO2Et
O
CO2Et
CO2Et
Cl
a, b
c
d
N
HN
N
N
N
N
N
H
N
N
R1
N
N
9
10
11a—d
O
HN
N
R1
N
N
NH
N
N
N
Compd
R1
PDE4B Kia
(l
M)
PDE4D Kia
(l
M)
Selectivityb (D/B)
e, f
N
N
N
13a
13b
13c
13d
H
Cl
F
NO2
26.2
>100
80.9 33.0
>100
>4
15
>24
>100
5.50 2.97
4.40 0.42
1.00
R1
R1
13a—d
>100
12a—d
a
Values that are the mean of two or more experiments are shown with their
standard deviations.
Scheme 3. Reagents and conditions: (a) KOH, EtOH, rt, then Me2SO4, acetone; (b)
Pd-C, HCO2NH4, MeOH, rt; (c) n-BuLi, THF, À78 °C, then, when R1 = H, Cl, or F:
R1C6H4COCl, or, when R1 = NO2: 3-NO2–C6H4C(O)N(OMe)Me; (d) hydrazine hydrate,
EtOH, rt; (e) POCl3, reflux; (f) 6, NH4Cl, NMP, 100 °C.
b
Ratio of Ki (PDE4D) / Ki (PDE4B).
parallel the trends observed in the binding assay. The variations
appeared to correlate with cell permeability, as measured by PAM-
PA. Despite this variability, the cyclase assay corroborated com-
pound 16 as a potent and selective PDE4B inhibitor (PDE4B
tuted-phenyl analog 15. This compound exhibited 41 nM PDE4B
binding affinity and an impressive 205-fold selectivity over PDE4D
(Table 5). By adding the 3-nitro substituent to the phenyl ring (16),
we further increased the affinity to 5.6 nM, while sacrificing some
of the selectivity gain. Selected compounds were examined in the
functional, cell-based cAMP assay, but the potency did not always
IC50 = 0.389 lM, PDE4D IC50 = 4.90 l
M).13 We further profiled 16
for selectivity against other PDEs and did not observe any signifi-
cant activity against these targets (PDEs 1–3, 5, and 6:
IC50 >10 lM). Likewise, compound 16 was selective for PDE4B

2166
A. F. Donnell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2163–2167
Table 4
Bicyclic analogs with piperazine amides
HN
N
N
N
N
R2
N
NO2
M)
Compd
R2
PDE4B Kia
(l
PDE4D Kia
(lM)
Selectivityb (D/B)
>2
N
N
14a
52.2
>100
O
O
14b
0.432
15.3
35
14c
14d
14e
0.238
5.08
11.5
>100
48.6
48
N
O
O
O
N
>20
50
N
N
S
0.977
O
O
O
N
14f
0.222
4.44
12.7
51.2
57
12
S
N
N
14g
N
N
14h
0.077 0.013
4.44 0.06
58
O
a
Values that are the mean of two or more experiments are shown with their standard deviations.
Ratio of Ki (PDE4D)/Ki (PDE4B).
b
Table 5
Pyrazine amide analogs
16 had poor oral bioavailability in PK studies. Nevertheless, these
compounds are potentially valuable tools for understanding the
structural features responsible for PDE4B selectivity.
N
N
In summary, we have identified a series of novel, potent,
PDE4B-selective pyridazino[4,5-b]indolizines. By utilizing the re-
sults of SAR studies around HTS hit 1, we prepared compound
15, which exhibited >200-fold selectivity for PDE4B over PDE4D,
and, in compound 16, we produced a species with single-digit
nanomolar affinity for PDE4B.
HN
N
N
N
N
O
N
R1
M)
Acknowledgments
Compd
R1
PDE4B Kia
(l
PDE4D Kia
(
lM)
Selectivityb (D/B)
15
16
H
NO2
0.041 0.009
0.0056 0.0001
8.37 2.33
0.147 0.004
205
26
We thank the management of the Neuroscience and Medicinal
Chemistry departments for their support of this program.
a
Values that are the mean of two or more experiments are shown with their
standard deviations.
b
References and notes
Ratio of Ki (PDE4D)/Ki (PDE4B).
1. (a) Houslay, M. D.; Schafer, P.; Zhang, K. Y. J. Drug Discovery Today 2005, 10,
1503; (b) Bender, A. T.; Beavo, J. A. Pharmacol. Rev. 2006, 58, 488.
2. (a) Wachtel, H. Neuropharmacology 1983, 22, 267; (b) Przegalinski, E.; Bigajska,
K. Pol. J. Pharmacol. Pharm. 1983, 35, 233; (c) Zhang, H.-T. Curr. Pharm. Des.
2009, 15, 1688.
versus the PDE4A and PDE4C subtypes (>fivefold selectivity in the
cyclase assay). Subsequent characterization of these compounds
found that they inhibit hERG channels. Additionally, compound

A. F. Donnell et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2163–2167
2167
3. Yang, L.; Calingasan, N. Y.; Lorenzo, B. J.; Beal, M. F. Exp. Neurol. 2008, 211, 311.
4. (a) Millar, J. K.; Pickard, B. S.; Mackie, S.; James, R.; Christie, S.; Buchanan, S. R.;
Malloy, M. P.; Chubb, J. E.; Huston, E.; Baillie, G. S.; Thomson, P. A.; Hill, E. V.;
Brandon, N. J.; Rain, J.-C.; Camargo, L. M.; Whiting, P. J.; Houslay, M. D.;
Blackwood, D. H. R.; Muir, W. J.; Porteous, D. J. Science 2005, 310, 1187; (b)
Kanes, S. J.; Tokarczyk, J.; Siegel, S. J.; Bilker, W.; Abel, T.; Kelly, M. P.
Neuroscience 2007, 144, 239; (c) Siuciak, J. A.; Chapin, D. S.; McCarthy, S. A.;
Martin, A. N. Psychopharmacology 2007, 192, 415.
5. (a) Rose, G. M.; Hopper, A.; De Vivo, M.; Tehim, A. Curr. Pharm. Des. 2005, 11,
3329; (b) Ghavami, A.; Hirst, W. D.; Novak, T. J. Drugs R. D. 2006, 7, 63; (c)
Blokland, A.; Schreiber, R.; Prickaerts, J. Curr. Pharm. Des. 2006, 12, 2511; (d)
McLachlan, C. S.; Chen, M. L.; Lynex, C. N.; Goh, D. L. M.; Brenner, S.; Tay, S. K. H.
Arch. Neurol. 2007, 64, 456; (e) Reneerkens, O. A. H.; Rutten, K.; Steinbusch, H.
W. M.; Blokland, A.; Prickaerts, J. Psychopharmacology 2009, 202, 419.
6. Robichaud, A.; Stamatiou, P. B.; Jin, S.-L. C.; Lachance, N.; MacDonald, D.;
Laliberté, F.; Liu, S.; Huang, Z.; Conti, M.; Chan, C.-C. J. Clin. Invest. 2002, 110,
1045.
4 °C. The supernatant was aliquoted and frozen at À20 °C until use in assay.
Compounds were diluted to a final concentration ranging from 100 M to
0.3 nM. Each assay well contained 200 l of diluted lysate, 25 l of diluted
compound, and 25
of [methyl-3H]-rolipram (2.5 nM final). Plates were
l
l
l
ll
incubated at 30 °C for 1 h with gentle rotation. The reaction was terminated by
filtering with ice-cold assay buffer (10 mM Tris, pH 7.5 and 50 mM MgCl2) on a
GF/B filter mat using a Filtermate harvester (Packard). Plates were dried at
37 °C for 1 h and then counted on a TopCount liquid scintillation counter
(Packard). CPM values were expressed as percent specific binding and plotted
against compound concentration. A curve was fitted using a four-parameter
logistic fit and the IC50 value was determined. The Ki was calculated using the
Cheng–Prusoff equation, pKi = IC50/(1+L/Kd), where L is the concentration of
free radioligand used in the assay and Kd is the dissociation constant of the
radioligand for the receptor. The IC50 is the concentration of competing ligand
that displaces 50% of the specific binding of the radioligand.
11. Inducible CHO cells stably expressing b2 adrenergic receptor and either
PDE4B3 or PDE4D4 were plated at 2 Â 104 cells/well in 96-well plates in Ham’s
F12 medium, 10% FBS, 100 units penicillin streptomycin and induced with
50 nM mifepristone for 48 h before use in assay. Media was removed from cells
and replaced with serum-free Ham’s F12 medium and incubated at 37 °C for
7. Spina, D. Br. J. Pharmacol. 2008, 155, 308.
8. Srivani, P.; Usharani, D.; Jemmis, E. D.; Sastry, G. N. Curr. Pharm. Des. 2008, 14,
3854.
9. (a) Huang, Z.; Liu, S.; Zhang, L.; Salem, M.; Greig, G. M.; Chan, C. C.; Natsumeda,
Y.; Noguchi, K. Life Sci. 2006, 78, 2663; (b) Giovannoni, M. P.; Cesari, N.;
Graziano, A.; Vergelli, C.; Biancalani, C.; Biagini, P.; Piaz, V. D. J. Enzyme Inhib.
Med. Chem. 2007, 22, 309; (c) Kranz, M.; Wall, M.; Evans, B.; Miah, A.;
Ballantine, S.; Delves, C.; Dombroski, B.; Gross, J.; Schneck, J.; Villa, J. P.; Neu,
M.; Somers, D. O. Bioorg. Med. Chem. 2009, 17, 5336; (d) Naganuma, K.; Omura,
A.; Maekawara, N.; Saitoh, M.; Ohkawa, N.; Kubota, T.; Nagumo, H.; Kodama, T.;
Takemura, M.; Ohtsuka, Y.; Nakamura, J.; Tusjita, R.; Kawasaki, K.; Yokoi, H.;
Kawanishi, M. Bioorg. Med. Chem. Lett. 2009, 19, 3174.
10. Inducible CHO cells stably expressing b2 adrenergic receptor and either
PDE4B3 or PDE4D4 were seeded at 6 Â 107cells in a 10-cell factory (Corning) in
Ham’s F12 medium, 10% FBS, 100 units penicillin streptomycin and induced
with 50 nM mifepristone for 48 h before harvesting. Cell pellets were lysed in
50 ml Ripa buffer (50 mM Tris, pH 7.5, 20 mM NaCl, and 1% NP40) for 15 min
on ice. Particulate matter was removed by centrifugation at 27,000g for 1 h at
15 min. Compounds were diluted to a final concentration ranging from 100
to 0.3 nM. Isoproterenol (10 M final) was added to compound wells. Controls
included four wells each of Ham’s F12 medium alone, isoproterenol alone, and
10 M rolipram with isoproterenol. Cells were incubated at 37 °C for 15 min,
then 100 mM perchloric acid was added to stop the reaction and lyse cells for a
minimum of 15 min. In Optiplates (Packard), 50 l of cell lysate was added per
well, plus 50
l each of [125I]-cAMP, primary rabbit cAMP antibody, and SPA
lM
l
l
l
l
PVT fluomicrospheres with anti-rabbit secondary-antibody according to
manufacturer’s directions. Plates were incubated at room temperature
overnight and then read on a TopCount liquid scintillation counter (Packard).
12. Foley, L. H. Tetrahedron Lett. 1994, 35, 5989.
13. For compound 15, the PDE4B IC50 was 6.7
unable to determine an accurate value for selectivity against PDE4D due to
limitations of the assay at higher concentrations (IC50 >100 M).
lM in the cyclase assay. We were
l