Pyrazoloquinolines as PDE10A inhibitors: Discovery of a tool compound

Article abstract of DOI:10.1016/j.bmcl.2011.12.080

A series of pyrazoloquinolines, possessing (hetero)arylhydroxymethyl substituents at the quinoline C-4 position were evaluated as PDE10A inhibitors. Among these, methylpyrimidyl analogue 15 was identified as having good rodent and monkey exposure, and a MED of 10 mg/kg in an in vivo model.

Full text of DOI:10.1016/j.bmcl.2011.12.080

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1335–1339
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Pyrazoloquinolines as PDE10A inhibitors: Discovery of a tool compound
William T. McElroy a, , Zheng Tan a, Kallol Basu a, Shu-Wei Yang a, Jennifer Smotryski a, Ginny D. Ho a,
⇑
Deen Tulshian a, William J. Greenlee a, Deborra Mullins b, Mario Guzzi b, Xiaoping Zhang b,
Carina Bleickardt b, Robert Hodgson b
a Department of Chemical Research—CV & CNS, Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
b Neurobiology Department of Biological Research, Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of pyrazoloquinolines, possessing (hetero)arylhydroxymethyl substituents at the quinoline C-4
position were evaluated as PDE10A inhibitors. Among these, methylpyrimidyl analogue 15 was identified
as having good rodent and monkey exposure, and a MED of 10 mg/kg in an in vivo model.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 7 November 2011
Revised 13 December 2011
Accepted 15 December 2011
Available online 21 December 2011
Keywords:
Phosphodiesterase inhibitor
PDE10A
Schizophrenia
Schizophrenia is a complex mental disorder that is estimated to
affect 0.5% of the world’s population.1 The predominant course of
treatment is administration of antipsychotic agents, with these
drugs generally categorized as being either ‘typical’ or ‘atypical’
antipsychotics. Both classes however largely rely on antagonism
of the dopamine D2 receptor to achieve the desired response.2
Although effective, these medications frequently produce undesir-
able side effects, including but not limited to weight gain, sedation,
and extrapyramidal syndrome (EPS). Phosphodiesterase 10A
(PDE10A) is a dual cAMP/cGMP hydrolyzing enzyme that has re-
cently emerged as a promising target for the treatment of schizo-
phrenia. Since it is believed that inhibition of PDE10A might
produce antipsychotic efficacy without interacting directly on
dopamine receptors, a drug relying on PDE10A inhibition may exhi-
bit a more desirable side effect profile than currently available
treatments.3–5
In order to advance our PDE10A program, a tool compound was
needed for complete assessment of both efficacy and side effect lia-
bilities. Such a compound needs to be a potent enzyme inhibitor
with adequate oral exposures in rodent and monkey. We recently
described the discovery of 1 (Fig. 1) as a potent inhibitor of
PDE10A.6–8 Compound 1 inhibited MK-801 induced hyperactivity9
with a MED of 3 mg/kg in the rat following oral dosing. Unfortu-
nately, 1 had no oral exposure in monkey and hence its EPS profile
could not be measured, as we intended to perform the assessment
in the monkey.10 To improve exposure in monkey with this class of
compounds, we continued SAR development using compound 1 as a
starting point. Since it was known that oxidation of the homo-
morpholine ring in 1 was a major metabolic pathway, we sought
to replace this group with aromatic and heteroaromatic substitu-
ents under the hypothesis that this might improve oral exposure.
Monkey hepatocyte clearance rate was used as an estimate of
exposure in plasma, as we were mindful that a loose relationship
existed between the two parameters.11 This proved to be a useful
tool to prioritize testing in vivo, and we set a goal of clearance being
<15 lL/min/million cells.
The compounds described in this report were prepared using the
general synthesis depicted in Scheme 1. For the present study, we
were interested only in analogues with C-8 methyl and C-6 meth-
oxy/chloro substituents on the quinoline core. Thiosemicarbazide
intermediates 4 and 5 were thus prepared upon treatment of the
corresponding aniline with carbon disulfide and hydrazine. Con-
densation with ethyl 2-chloroacetoaceate gave, following extrusion
of sulfur, pyrazole esters 6 and 7. Saponification and cyclization
with POCl3 then yielded pyrazoloquinolines 10 and 11. Substitution
of the chloride for a bromide was accomplished upon hydrolysis
(HCl, MeOH) followed by bromination (POBr3, DMF). This three step
sequence gave a higher overall yield than the direct conversion of 8
to 12 using POBr3. In order to obtain the desired targets, the bromo-
quinolines were allowed to react with n-BuLi, followed by a range of
heteroaromatic ketones and aldehydes. The yields for this step were
generally modest.
The SAR of representative analogues in the C-6-methoxy
quinoline series is presented in Table 1. All compounds prepared
were racemates unless otherwise noted. The PDE inhibitory activity
⇑
Corresponding author.
E-mail address: william.mcelroy@merck.com (W.T. McElroy).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.12.080

1336
W. T. McElroy et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1335–1339
Figure 1. Select data for compound 1.
Scheme 1. Synthesis of pyrazoloquinolines.

W. T. McElroy et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1335–1339
1337
Table 1
Table 2
PDE10A Inhibitory activity of pyrazoloquinolines possessing various C-4 benzyl
groups
PDE10A Inhibitory activity of pyrazoloquinolines possessing various linkers at the C-4
position
Entry
R
PDE10A
Ki (nM)
CYP3A4
Inhib.
Monkey
hepatocyte
clearance (lL/min/
million cells)
Entry Ar
PDE10A
Ki (nM)
CYP3A4
Inhib.
Monkey hepatocyte
clearance ( L/min/
million cells)
l
(IC50
M)
,
(IC50
,
lM)
l
ꢀ20
3.0 (co);
2.9 (pre)
1
2
12
2
(co); 0.8
(pre)
64
1
2
0.06
10
66
12.9
(co); 0.3
(pre)
2.7 (co);
<0.2 (pre)
67
NA
10.6 (co);
4.6 (pre)
3
4
54
NA
66
14.2
(co); 7.1
(pre)
3
4
4
1
66
66
4.2 (co);
0.6 (pre)
0.07
NA
7.8 (co);
4.4 (pre)
5
6
7
0.6
21
2
63
22
NA
4.9 (co);
1.2 (pre)
12.0
(co); 0.3
(pre)
5
0.3
62
0.7 (co);
<0.2 (pre)
8
9
50
2
NA
NA
60
icant CYP3A4 inhibition. We hypothesized that the CYP3A4 inhibi-
tion may be due to an interaction of the CYP enzyme with the
nonbonding electrons of the pyridine nitrogen.13 Thus, we prepared
the ortho-fluoro substituted pyridine (Table 1, entry 4). Unfortu-
nately, no improvement in CYP inhibition was observed, although
PDE10A potency was maintained. In order to further explore this,
a number of analogues in which the pyridine was replaced with
an azole were prepared. Selected examples are displayed in entries
5–7. In all cases this modification failed to give any significant
improvement in CYP inhibition. It was generally observed that ana-
logues in which the C-4 heterocycle was replaced with a phenyl
ring were less active than their heteroaromatic counterparts, and
exemplary data is presented in entries 7–9. A notable exception
was the 4-methoxy phenyl species (Table 1, entry 9), although this
compound was a potent CYP3A4 inhibitor. Thus it appeared that in
this subseries lowering the basicity of, or replacing the heteroaryl
ring does not reduce CYP3A4 inhibition.
We next investigated the effect of replacing the benzylic hydro-
xyl group of the analogues described above with other functional
groups. Although this would not necessarily be expected to reduce
the level of CYP inhibition, it was hypothesized that such structural
modifications would improve hepatocyte clearance values.14 Thus,
the alcohol underwent oxidation to the corresponding ketone upon
treatment with IBX. The hydroxyl group was also converted to a 2°
alkyl fluoride upon reaction with DAST. In order to prepare ana-
logues with a benzylic cyano group, 4-chloroquinolines were al-
lowed to react with anions derived from benzyl nitriles
(Scheme 1). The resulting SAR studies are listed in Table 2.
Two compounds possessing benzyl nitrile subunits (Table 2, en-
tries 1 and 2) were active in the PDE10A inhibition assay, although
1.2 (co);
<0.2 (pre)
10
83
NA
63
of target compounds was determined using a scintillation proxim-
ity assay with [3H]cAMP as substrate measuring the hydrolysis of
cAMP to AMP with recombinant human PDEs 1–11. Inhibitors hav-
ing PDE10A Ki <50 nM were selected for further evaluation. During
the course of this investigation, we became aware that many of
these analogues exhibited significant CYP3A4 inhibition. Therefore,
compounds were routinely screened for both reversible (preincu-
bation of the analogue with the CYP3A4 enzyme) and time-depen-
dent (coincubation of the analogue with CYP3A4) inhibition.12 As
an initial criteria, we were interested in analogues that exhibited
both time-dependent and reversible CYP3A4 IC50 >10 lM.
Pyrazoloquinolines possessing hydroxymethylpyridyl substitu-
ents were uniformly active. When the nitrogen atom was moved
from the 4- to 3- or 2- position, a loss in PDE10A inhibitory activity
was observed (Table 1, entries 1–3). This may be explained by the
ability of the 4-pyridyl isomer to better interact, either directly or
through a water bridge, with a nearby zinc atom in the PDE10 cat-
alytic site.6 The optically active 4-pyridyl analogue was obtained by
resolving the racemic mixture and is one of the most potent com-
pounds identified. Unfortunately, both the 4- and 3-pyridyl
analogues exhibited high monkey hepatocyte clearance with signif-

1338
W. T. McElroy et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1335–1339
Table 3
PDE10A Inhibitory activity of pyrazoloquinolines possessing C-6-chloro substitution
Entry
1
R
PDE10A Ki (nM)
CYP3A4 Inhib. (IC50
,
lM)
Monkey hepatocyte clearance (
lL/min/million cells)
0.2
0.5 (co); 0.3 (pre)
25
2
11
2.1 (co); 1.1 (pre)
46
3
4
39
1
12.2 (co); 1.8 (pre)
>20 (co); >20 (pre)
NA
37
5
6
0.4
21
1.0 (co); 7.4 (pre)
0.8 (co); 0.3 (pre)
NA
NA
7
3
>20 (co); 7.9 (pre)
8
Figure 2. Selected data from analogues 14 and 15.

W. T. McElroy et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1335–1339
1339
they were strong CYP3A4 inhibitors. In the case of the unsubstitut-
ed phenyl ring analogue (entry 1), no improvement in clearance
was observed. We also explored 4-pyridyl derivatives, since these
were the most potent compounds in the hydroxy series. Although
Ki values were in the single digit nanomolar range, the 3° alcohol
(Table 2, entry 3), ketone (Table 2, entry 4), and alkyl fluoride
(Table 2, entry 5) exhibited high clearance values in monkey hepa-
tocytes. Disappointingly, modification of the benzylic substituent
had no effect on clearance rate.
Finally, we examined the SAR in the C-6-chloro pyrazoloquino-
line series. Analogues in this series generally exhibited lower
monkey hepatocyte clearance rates relative to their C-6 methoxy
counterparts, and selected results are presented in Table 3. Unsub-
stituted pyridines remained active PDE10A inhibitors (Table 3, en-
tries 1–3), although CYP3A4 inhibition was a recurring problem.
The addition of a fluorine atom adjacent to the nitrogen atom
was an effective strategy for decreasing CYP inhibition (Table 3, en-
try 4). This may be due to decreasing the basicity of the pyridine
nitrogen,13 and suggested that analogues with less basic heterocy-
les might be less potent CYP3A4 inhibitors in this subseries, if not
in the C-8 methoxy collection described in Table 1. Unfortunately,
an oxazole (Table 3, entry 5) and imidazole (Table 3, entry 6) both
exhibit significant CYP inhibition. Finally, a methyl-substituted
pyrimidine (Table 3, entry 7, 15) had a PDE10A Ki = 3 nM with
acceptable clearance rate in monkey hepatocytes and reasonable
CYP3A4 profile.
In conclusion, we have described a series of pyrazoloquinolines,
possessing (hetero)arylhydroxymethyl substituents at the quino-
line C-4 position as PDE10A inhibitors. 4-Pyridyl analogues were
found to be extremely potent, but exhibited high clearance rates
in monkey hepatocytes. Replacing the benzylic hydroxyl group
with other functionalities failed to lower these values. Replacing
the quinoline C-6 methoxy group with a chloro substituent
generally led to reduction of the clearance rate in monkey
hepatocytes. Within this series, the methylpyrimidyl compound
15 emerged as a tool compound with a good overall profile.
Complete evaluation of this compound in a variety of rodent and
monkey models, including efficacy and side effect profiles, would
assist in developing PDE10A inhibitors for the treatment of
schizophrenia.
Acknowledgments
The authors wish to acknowledge Drs. Mark Liang, Jianshe
Kong, Jesse Wong, and Ms. Teresa Andreani and Mr. Tao Meng
for preparation of the intermediates, and Dr. Tze-Ming Chan and
his group for structure confirmation of some analogues. We
acknowledge Drs. Diane Rindgen and Xiaoming Cui and the DMPK
group for acquiring hPXR and pharmacokinetic data, and Drs. Steve
Sorota and Tony Priestley for providing hERG data.
References and notes
Based on this data, methylpyrimidine 15 was selected for fur-
ther evaluation (Fig. 2). This compound inhibited MK-801 induced
hyperactivity in rat with an MED of 10 mg/kg when given orally.9,15
Analogue 15 does not exhibit CYP450 induction in rat liver slice
studies and is approximately within our desired criteria of
CYP450 inhibition. Additionally, this compound was inactive
1. Goldner, E. M.; Hsu, L.; Waraich, P.; Somers, J. M. Can. J. Psychiatry 2002, 47, 833.
2. Kapur, S.; Mamo, D. Prog. Neuor-Phychoph. 2003, 27, 1081.
3. Kehler, J.; Ritzén, A.; Greve, D. R. Expert Opin. Ther. Patents 2007, 17, 147.
4. For a review, see: Marino, M. J.; Knutsen, L. J. S.; Williams, M. J. Med. Chem.
2008, 51, 1077.
5. For other PDE10 programs, see: (a) Verhoest, P. R.; Chapin, D. S.; Corman, M.;
Fonseca, K.; Harms, J. F.; Hou, X.; Marr, E. S.; Menniti, F. S.; Nelson, F.; O’Connor,
R.; Pandit, J.; Proulx-LaFrance, C.; Schmidt, A. W.; Schmidt, C. S.; Suiciak, J. A.;
Liras, S. J. Med. Chem. 2009, 52, 5188; (b) Chappie, T. A.; Humphrey, J. M.; Allen,
M. P.; Estep, K. G.; Fox, C. B.; Lebel, L. A.; Liras, S.; Marr, E. S.; Menniti, F. S.;
Pandit, J.; Schmidt, C. J.; Tu, M.; Williams, R. D.; Yang, F. V. J. Med. Chem. 2007,
50, 182; For a recent review, see: (c) Kehler, J.; Ritzén, A.; Greve, D. R. Expert
Opin. Ther. Patents 2009, 19, 1715.
6. Ho, G. D.; Yang, S.-W.; Smotryski, J.; Bercovici, A.; Nechuta, T.; Smith, E. M.;
McElroy, W. T.; Tan, Z.; Tulshian, D.; Mckittrick, B.; Greenlee, W. J.; Hruza, A.;
Xiao, L.; Rindgen, D.; Guzzi, M.; Zhang, X.; Bleickardt, C.; Mullins, D.; Hodgson,
R. Bioorg. Med. Chem. Lett. in press, doi.org/10.1016/j.bmcl.2011.11.127.
7. Yang, S.-W.; Smotryski, J.; McElroy, W. T.; Ho, G. D.; Tulshian, D.; Greenlee, W.
J.; Mullins, D.; Hodgson, R.; Xiao, L.; Hruza, A. Bioorg. Med. Chem. Lett. 2012, 22,
235.
8. Ho, G. D.; Yang, S.-W.; Smith, E. M.; McElroy, W. T.; Basu, K.; Smotryski, J.; Tan,
Z.; McKittrick, B. A.; Tulshian, D. B. PCT Int. Appl. WO 2010062559, 2010.
9. Rung, J. P.; Carlsson, A.; Markinhuhta, K. R.; Carlsson, M. L. Prog. Neuor-
Phychoph. 2005, 29(5), 833.
against a panel of 23 kinases (IC50 >30 lM). On 3 mg/kg oral dos-
ing, 15 exhibited excellent exposure (AUC = 11,496 nM h) with a
half life of 7.1 h in monkey. It also displayed good oral exposure
in rodent with an AUC of 4286 nM h over 6 h at 10 mg/kg dose.
An ex vivo study (after sacrificing the rodent at the 6 h time point)
revealed that the amount of 15 present in the brain was below the
limit of detection. However, the concentration of 15 in the plasma
at the 6 h time point (220 nM) was also minimal, making a deter-
mination of brain to plasma ratio impossible. As a closely related
analogue, pyridine 14 was also profiled. Similar to 15, compound
14 inhibited MK-801 induced hyperactivity with
a MED of
10 mg/kg. This analogue exhibited excellent exposure in the rat
(AUC = 11,040 nM h) over 6 h. Notably, the plasma concentration
at the 6 h time point of the study was 810 nM, making a determi-
nation of the brain to plasma ratio feasible. At the 6 h time point,
the amount of 14 in the brain was 115 ng/g and the brain to plasma
ratio was 1.5. In fact, several analogues in this series had brain to
plasma ratios >0.5, helping support our hypothesis that the reduc-
tion in MK-801 hyperactivity is a result of PDE10A inhibition. The
overall profiles of 14 and especially 15 render these analogues
valuable tool compounds to investigate PDE10A inhibition as a
treatment for schizophrenia.
10. Varty, G. B.; Hodgson, R. A.; Pond, A. J.; Grzelak, M. E.; Parker, E. M.; Hunter, J. C.
Psychopharmacology 2008, 200(3), 393.
11. Lau, Y. Y.; Sapidou, E.; Cui, X.; White, R. E.; Cheng, K.-C. Drug Metab. Dispos.
2002, 30(12), 1446.
12. Yan, Z.; Caldwell, G. W. Curr. Top. Med. Chem. 2001, 1, 403.
13. Zlokarnik, G.; Grootenhuis, P. D. J.; Watson, J. B. Drug Discovery Today 2005,
10(21), 1443.
14. Nassar, A.-E. F.; Kamel, A. M.; Clarimont, C. Drug Discovery Today 2004, 9(23),
1020.
15. Data for this study was collected at the 1 h time point following dosing po.