1,5-Substituted nipecotic amides: Selective PDE8 inhibitors displaying diastereomer-dependent microsomal stability

Article abstract of DOI:10.1016/j.bmcl.2011.03.022

The first highly potent and selective PDE8 inhibitors are disclosed. The initial tetrahydroisoquinoline hit was transformed into a nipecotic amide series in order to address a reactive metabolite issue. Reduction of lipophilicity to address metabolic liab

Full text of DOI:10.1016/j.bmcl.2011.03.022

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3095–3098
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
1,5-Substituted nipecotic amides: Selective PDE8 inhibitors displaying
diastereomer-dependent microsomal stability
,
⇑
Michael P. DeNinno , Stephen W. Wright, Michael S. Visser à, John B. Etienne, Dianna E. Moore,
Thanh V. Olson, Benjamin N. Rocke, Melissa P. Andrews, Cynthia Zarbo, Michele L. Millham,
Brian P. Boscoe, David D. Boyer, Shawn D. Doran, Karen L. Houseknecht §
Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The first highly potent and selective PDE8 inhibitors are disclosed. The initial tetrahydroisoquinoline hit
was transformed into a nipecotic amide series in order to address a reactive metabolite issue. Reduction
of lipophilicity to address metabolic liabilities uncovered an interesting diastereomer-dependent trend in
turnover by human microsomes.
Received 14 January 2011
Revised 4 March 2011
Accepted 7 March 2011
Available online 11 March 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Phosphodiesterase 8
Pancreas
Diabetes
Microsomal stability
Phosphodiesterase 8B is a cAMP-specific isoform of the broader
class of phosphodiesterases (PDEs).1–3 Although mRNA for the clo-
sely related PDE8A has been found at low levels in most tissues,
PDE8B is more highly expressed in the brain, thyroid, pancreas
and adrenal cortex.4,5 Expression of the enzyme in pancreatic
b-cells4 suggested that inhibition of PDE8B might have therapeutic
utility in the treatment of diabetes. Such an agent could be used as
a mono-therapy, or in combination with drugs which signal
through cAMP such as those that act on the GLP-1 receptor.6 Other
PDE isoforms expressed in the pancreas include PDE1C, PDE3B,
PDE4 and PDE10.7 As no selective PDE8 inhibitors had been re-
ported, a high throughput screen (HTS) was conducted with the
goal of identifying selective tools for exploring the in vitro and
in vivo phenotype of PDE8B inhibition.
Figure 1. Structures of cAMP and IBMX.
The majority of hits from the HTS contained classical PDE phar-
macophores, such as the non-selective cAMP mimic inhibitor IBMX
(Fig. 1). One hit (1) based on a tetrahydroisoquinoline (THIQ) scaf-
fold looked particularly attractive due to its structural novelty. It
Figure 2. Structure of THIQ screening hit 1.
was hoped that this structural novelty would result in greater
specificity for derivatives of 1 against the PDE8B isozyme (Fig. 2).8
A standard SAR campaign around 1 gave 2 which had high po-
tency against PDE8B (IC50 = 5 nM) and excellent selectivity.9 How-
ever, it was determined that multiple analogs were positive in a
glutathione adduct-based reactive metabolite assay, presumably
at least partially through compound 3 which was detected as one
of the major products upon microsomal incubation (Fig. 3).10
⇑
Corresponding author. Tel.: +1 858 404 6664; fax: +1 858 404 6719.
E-mail address: mike_deninno@sd.vrtx.com (M.P. DeNinno).
Present address: Vertex Pharmaceuticals Inc., 11010 Torreyana Rd., San Diego, CA
92121, USA
à
Present address: Novartis Institutes for BioMedical Research, Inc., 250
Massachusetts Ave., Cambridge, MA 02139, USA
§
Present address: University of New England, 716 Stevens Ave., Portland, ME
04103, USA
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.03.022

3096
M. P. DeNinno et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3095–3098
Figure 3. Structures of analog 2 and microsomal metabolite 3.
Figure 4. Structure of heterocycle 4 and fluorinated THIQ analog 5.
Scheme 1. Preparation of 5-substituted nipecotic amide analogs. Reagents and
conditions: (a) R1ZnX (1.5 equiv), Pd(dppf)Cl2 (0.025 equiv), THF, 50 °C, 3 h; (b) H2
(50 psi), PtO2, H2SO4, EtOH, 20 °C, 24 h; (c) Boc2O (1.5 equiv), iPr2NEt (3 equiv),
dioxane, 20 °C, 18 h;. (d) NaOH (2 equiv), MeOH, reflux, 18 h; (e) R2NH2 (1 equiv),
propanephosphonic anhydride (1.2 equiv), Et3N (3 equiv), DMF, 20 °C, 18 h; (f) HCl
(5 equiv), Et2O, MeOH, 20 °C, 2 h; (g) ArCH2Cl (1.5 equiv), iPr2NEt (5 equiv), MeCN,
20 °C, 18 h.
We then directed our efforts upon solving the reactive metabo-
lite problem by any of three means: (i) replacement of the fused
aryl ring with more polar fused rings, (ii) blocking potential sites
of metabolism on the isoquinoline, and (iii) elimination of the
fused aryl ring.
The first strategy failed to afford compounds with PDE8B
inhibitory activity (e.g., analog 4). Blocking of the isoquinoline 4-
position with a fluorine atom to afford 5 resulted in a 75-fold loss
of potency (IC50 = 370 nM) without significantly improving meta-
bolic stability (Fig. 4).
Table 2
PDE8B activity of trans-5-substituted nipecotate analogs
Elimination of the fused ring to reduce overall lipophilicity and
reduce the driving force for metabolism to the isoquinolonium spe-
cies 3, could be accomplished with a reduction in potency. How-
ever, PDE8B inhibition potency was strongly influenced by even
small structural changes (Table 1).
Substitution of the piperidine ring was explored to improve the
potency of the nipecotate template 6. These compounds were pre-
pared as shown in Scheme 1. This route was designed to afford pri-
marily the cis isomers, which preliminary modeling suggested
should better mimic the THIQ template. Yields were generally
good; following hydrogenation and amide bond formation, the
separable cis/trans isomers were isolated in ratios of 3:1 to 10:1
depending on the substituent R1.
R1
N
N
Ph
O
N
N
O
H
Compound
R1
IC50 (nM)
c log D
6
H
240
40
6
2.9
3.7
4.2
4.5
5.0
4.5
2.9
3.0
3.8
3.7
11
12
13
14
15
16
17
18
19
Methyl
Ethyl
n-Propyl
n-Butyl
Phenyl
CH2OMe
OH
6
15
0.9
3.2
2700
7
We were surprised to find that the cis isomers were uniformly
inactive (IC50 >2000 nM, data not shown), whereas the trans iso-
mers were the active diastereomers (Table 2).11
OAc
OCONHEt
120
Table 1
PDE8B activity of analogs lacking a fused ring
Addition of small alkyl groups at the C-5 position led to an in-
crease in potency with ethyl analog 12 providing the optimal bal-
ance of potency and lipophilicity. More potent derivatives could be
made (15) at the expense of lipophilicity and molecular weight.
Compound 12 showed similar potency and selectivity to 2, how-
ever, unlike 2 it did not produce reactive metabolites. The primary
liability of 12 was its poor microsomal stability, contributing to
high in vivo clearance and low oral bioavailability.12 This result
was not unexpected given the compound’s high log D of 4.2.13
Metabolite identification studies showed that multiple sites under-
went microsomal oxidation. Given the modular nature of the lead,
our goal was to identify more polar replacements for each of its
three segments as a means to reduce microsomal metabolism. In
order to make more significant changes to log D, a matrix set of
22 compounds was designed by combining two or three of the
new building blocks. Data for selected compounds is shown in
Table 3.14
N
R
N
N
Ph
N
Ph
O
A
O
N
O
NH
N
H
O
N
I
II
Compound
Structure
A
IC50 (nM)
c log D
6
7
I
I
CH2
O
240
>10,000
2400
>10,000
1700
2.9
2.9
2.1
2.7
3.4
8
9
10
I
II
II
SO2
H
CF3

M. P. DeNinno et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3095–3098
3097
Table 3
PDE8B activity of selected matrix compounds
R1
N
R3
N
O
A
R2
N
H
Compound
R1
R2
R3
A
O
IC50 (nM)
7.3
c log D
HLM Clint
O
20
Ethyl
Ph
Ph
O
3.6
320
240
O
21
22
CH2OMe
Ethyl
O
O
2.0
17
2.6
1.9
NH2
N
N
N
110
N
O
23
24
CH2OMe
CH2OMe
NH
O
13
65
0.64
0.91
63
41
N
O
O
O
N
Me
N
25
Ph
O
3.9
26
3.0
1.6
2.5
2.6
2.6
105
190
49
N
O
N
26
Ethyl
NH
O
N
NH2
N
27a
28a
29
H
H
Ph
58
O
Ph
Ph
O
166
66
84
N
N
Me
N
O
290
O
NH2
N
30
Ph
Ph
O
0.12
3.6
>300
a
Analogs 27 and 28 were tested as single 3R enantiomers.
In addition to maintaining potency, other positive attributes of
enantiomers of the core piperidine (Table 4). This compound was
chosen based on its properties and the fact that it had the lowest
Clint. Only one of the two trans isomers possessed PDE8B inhibitory
activity. There was no difference in microsomal clearance between
either of the pairs of cis or trans diastereomers. Experimental pKa’s
and log D’s were also determined since these parameters, if diver-
gent, could also help to explain the diastereomer-dependent
microsomal clearance data. Small but not significant differences
in pKa and log D were observed.
Insights into the binding conformation of these inhibitors
necessitated the use of homology models due to the inability to
crystallize the poorly expressing and difficult to purify PDE8B pro-
tein. After numerous docking experiments, a result that accounted
for the majority of the known structure activity relationships was
identified and is depicted in Figure 6, with 30 as the ligand
(Fig. 7). In addition to nicely filling lipophilic pockets, the following
interactions are possible. (1) A hydrogen bond from the protonated
piperidine NH to Tyr253 (could also be nitrogen lone pair to tyro-
sine OH). (2) Amide NH hydrogen bond to the backbone carbonyl
our leads 6 and 12, such as permeability, aqueous solubility and
selectivity over other phosphodiesterases (>1000-fold, except
equipotent against PDE8A) were also preserved. Again, the cis iso-
mers were inactive. However, microsomal stability was our great-
est concern. Figure 5 shows the correlation of c log D and HLM
intrinsic clearance for the set of compounds.
Despite an obvious relationship between c log D and micro-
somal clearance (Clint), reduction of c log D by four orders of mag-
nitude did not afford potent compounds with low clearance
(considered to be <10 lL/min/mg in this assay). Oddly, it was
found that the cis diastereomers (red markers) were significantly
more stable across a broad range of c log D. The unsubstituted
nipecotic amide analogs (R1 = H, blue markers) displayed clearance
values between the cis and trans isomers. Taken together, these
data suggest a conformational component to the metabolism. A
similar trend was also observed with RLM (data not shown).
One compound (24) from the series was separated into its four
component isomers in order to rule out differences between the

3098
M. P. DeNinno et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3095–3098
250
100
75
50
Figure 7. Structure of 30 used as the ligand in homology model docking
experiments.
25
This conformation, in which the amide group was in an axial
configuration, would be readily accessible in compound 2 given
the lack of any A1-3 strain. It would also explain the preference
for the trans isomer in the nipecotic analogs, since the cis isomers
would force both substituents in a di-axial arrangement. The abso-
lute configuration of the amide was also the same as the single
enantiomer nipecotic amide analogs 27 and 28 further supporting
this binding model.
10
5
Colored by diastereomer
cis trans NA
0.5
1
1.5
2
2.5
3
3.5
4
cLogD
In summary, the first highly potent and selective PDE8 inhibi-
tors have been identified. A scaffold-hopping exercise was success-
fully utilized to circumvent a reactive metabolite issue. Efforts to
reduce microsomal turnover through lowering lipophilicity led
to an interesting observation that the cis diastereomers were
significantly more stable in human microsomes than the trans.
Although the microsomal clearance of the series was responsive
to reductions in lipophilicity, it could not be moved into a range
deemed sufficiently low for clinical study. The disclosure and
optimization of additional PDE8B scaffolds will be the subject of
future publications.
Figure 5. Correlation of microsomal stability versus c log D. Pairs of diastereomers
are connected with dotted lines.
Table 4
Data for individual isomers of 24
N
O
N
O
O
O
N
O
N
H
Supplementary data
Isomer
PDE8B IC50
43 nM
HLM Clint
pKa
log D
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.03.022.
trans 1
trans 2
cis 1
48
49
<8
<8
7.2
1.0
1.0
0.8
0.8
>10
>10
>10
l
l
l
M
M
M
7.8
cis 2
References and notes
1. Hayashi, M.; Matsushima, K.; Ohashi, H.; Tsunoda, H.; Murase, S.; Kawarada, Y.;
Tanaka, T. Biochem. Biophys. Res. Commun. 1998, 250, 751.
2. Soderling, S. H.; Bayuga, S. J.; Beavo, J. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,
8991.
3. Soderling, S. H.; Bayuga, S. J.; Beavo, J. A. J. Biol. Chem. 1998, 273, 15553.
4. Dov, A.; Abramovitch, E.; Warwar, N.; Nesher, R. Endocrinology 2008, 149, 741.
5. Horvath, A.; Giatzakis, C.; Tsang, K.; Greene, E.; Osorio, P.; Boikos, S.; Libè, R.;
Patronas, Y.; Robinson-White, A.; Remmers, E.; Bertherat, J.; Nesterova, M.;
Stratakis, C. A. Eur. J. Genet. 2008, 16, 1245.
6. Pyne, N. J.; Furman, B. L. Diabetologia 2003, 46, 1179.
7. Pratley, R. E.; Gilbert, M. Rev. Diabet. Stud. 2008, 5, 73.
8. Such isoform selectivity has been a major issue in past PDE research, due to a
highly conserved catalytic domain among PDE family; see: Ke, H.; Wang, H.
Curr. Top. Med. Chem. 2007, 7, 391.
9. Selectivity was >1000-fold for all other PDE isoforms except PDE8A
(IC50 = 1.8 nM).
10. The isoquinoloinium species 3 was identified upon incubation with human
liver microsomes, and was shown to alkylate glutathione in this assay. See
Mutlib, A. E.; Shockcor, J. P. (2003) In Drug Metabolizing Enzymes-Cytochrome
P450 and Other Enzymes in Drug Discovery and Development Lee, J.S.; Obach, R.
S.; Fisher, M.B., Eds.; Fontis Media, S.A., Lausanne, Switzerland, p 33.
11. Unless otherwise noted and where applicable, compounds were tested as
diastereomerically pure racemic mixtures. Analogs containing an
enantiomerically pure amide substituent (e.g., 24) were mixtures of
diastereomers.
Figure 6. Homology model of 30 binding to PDE8B catalytic domain.
12. Rat PK for compound 12: Cl = 98 ml/min/kg; Vdss = 1.4 L/kg; F = 7%; rat PK for
compound 24: Cl = 146 ml/min/kg; Vdss = 2.1 L/kg.
13. van de Waterbeemd, H.; Smith, D. A.; Jones, B. C. J. Comput. Aided Mol. Des.
2001, 15, 273.
14. Data for the complete set of analogs can be found in the Supplementary data.
of Pro270. (3) A hydrogen bond from Tyr64 to the oxazole ring oxy-
gen atom, perhaps through a water molecule.