Novel inhibitors of the high-affinity l-proline transporter as potential therapeutic agents for the treatment of cognitive disorders

Article abstract of DOI:10.1016/j.bmcl.2014.06.049

The incidence of cognitive disorders such as Alzheimer's disease continues to increase unabated. While cures for such diseases have eluded investigators, progress is being made on alleviating certain symptoms of these diseases. Mouse knockouts of the proline transporter (PROT), a high affinity Na ⁺/Cl^--dependent transporter, indicated its potential as a novel therapeutic target for cognition improvement. Herein we report our investigation into a novel class of PROT inhibitors.

Full text of DOI:10.1016/j.bmcl.2014.06.049

Bioorganic & Medicinal Chemistry Letters 24 (2014) 3886–3890
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel inhibitors of the high-affinity L-proline transporter as potential
therapeutic agents for the treatment of cognitive disorders
a
a
a
a
G. Greg Zipp a, , Joseph Barbosa , Michael A. Green , Kristen M. Terranova , Cynthia Fink ,
Xuan-Chuan Yu b, Amr Nouraldeen b, Alan Wilson b, Katerina Savelieva c, Thomas H. Lanthorn c,
S. David Kimball a
⇑
a Department of Medicinal Chemistry, Lexicon Pharmaceuticals, 350 Carter Rd., Princeton, NJ 08540, United States
b Department of Pharmaceutical Discovery, Lexicon Pharmaceuticals, 8800 Technology Forest Place, The Woodlands, TX 77381, United States
c Department of Neurology, Lexicon Pharmaceuticals, 8800 Technology Forest Place, The Woodlands, TX 77381, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The incidence of cognitive disorders such as Alzheimer’s disease continues to increase unabated. While
cures for such diseases have eluded investigators, progress is being made on alleviating certain symptoms
of these diseases. Mouse knockouts of the proline transporter (PROT), a high affinity Na+/Clꢀ-dependent
transporter, indicated its potential as a novel therapeutic target for cognition improvement. Herein we
report our investigation into a novel class of PROT inhibitors.
Received 4 June 2014
Revised 17 June 2014
Accepted 18 June 2014
Available online 1 July 2014
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Alzheimer’s disease
Dimentia
Proline transporter
Inhibitor
Acyl piperazine
Cognitive disorders affect the ability to think, to process and
store information, and to solve problems. Alzheimer’s disease
(AD), a progressive and fatal brain disease, is the most common
form of dementia. It currently directly affects as many as 35
million people worldwide and is projected to double in incidence
every 20 years, affecting 1 in 85 people globally by the year
2050.1 While early symptoms are often mistaken for aging or
stress, rapid advancement of the disease often leaves the patient
completely dependent on caregivers.2 Monetary costs of the
disease have been estimated at $160 billion annually, but the
incalculable cost of the burdens on main caregivers includes social,
psychological, and physical aspects as well.
While AD currently has no known cure, efforts to alter the
course of the disease and improve the quality of life for patients
can have a dramatic affect not only for those with the disease,
but those caring for them as well. Current treatments of the cogni-
tive impairment associated with AD include acetylcholinesterase
inhibitors3 and an NMDA receptor antagonist. A known feature of
AD is a reduction in the activity of the cholinergic neurons, which
many researchers presume underlies much of the short-term
memory loss associated with the disease. Acetylcholinesterase
inhibitors ameliorate the effects of the loss of cholinergic neurons
by reducing the degradation rate of acetylcholine and increasing its
effective concentration in the brain. Memantine, a noncompetitive
NMDA receptor antagonist approved for the treatment of moder-
ate-to-severe AD, inhibits the overstimulation of NMDA receptors
by glutamate, reducing the effects of this excitotoxicity pathway
endemic to AD.
The amino acid L-proline has been reported to play a role in reg-
ulating synaptic transmission in the mammalian brain.4 In general,
neurotransmitter systems typically employ various mechanisms
that inactivate signaling to regulate synaptic transmission and
protect against excitotoxicity. Many of these rely on the action of
a Na+-dependent transporter to facilitate rapid re-uptake of neuro-
transmitter ligands.5 The mammalian proline transporter (PROT) is
a high affinity Na+/Clꢀ-dependent transporter expressed in specific
regions of the brain, including the cerebral cortex and hippocam-
pus.6,7 It displays sequence homology to other neurotransmitter
transporters such as the glycine, dopamine, serotonin, and norepi-
nephrine transporters.7,9 PROT has been found to be expressed
selectively within synaptic terminals of a subset of glutamatergic
neurons and may play an important role in the regulation of
excitatory neurotransmission. Knockout mice lacking functional
PROT have demonstrated improved performance in learning and
memory tests and do not differ from wild-type mice in locomotor
⇑
Corresponding author. Tel.: +1 4128490808.
E-mail address: ggzipp@yahoo.com (G.G. Zipp).
http://dx.doi.org/10.1016/j.bmcl.2014.06.049
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

G. G. Zipp et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3886–3890
3887
basic pyrimidine (8) sacrificed a small amount of potency for
improved stability (S9 fraction, 100% remaining after 30 min.).
Much of the subsequent SAR investigation retained the pyrimidine
ring, and focused on changes to the central and western portions of
the molecule.
Assessment of the favored orientation of the biphenyl moiety
led to the synthesis of compounds 11 and 12 (Fig. 3). Increased
radial deviation from the original 4-biphenyl substituted com-
pound led to a decrease in potency. We therefore chose to continue
investigation of this hit by retaining the para substitution pattern
and focusing on changes to the distal phenyl ring.
Figure 1. Initial HTS hit.
activity and motor coordination.9 Three opioid peptide-related
peptides were found to inhibit L-proline uptake by rat brain synap-
Our general synthetic route for exploring the SAR of the biphe-
nyl region is depicted in Scheme 1. Thermal reaction of 2-chloro-
pyrimidine and piperidine gave amine 13. Subsequent EDC
coupling of 13 with 4-bromobenzoic acid afforded bromide 14,
our common intermediate for exploration of the terminal phenyl
ring through Suzuki coupling reactions with a diverse set of boro-
nic acids. The synthetic scheme was efficient, modular and
straightforward, and allowed for investigation of each of the lead
series’ component pieces.
tosomes and were subsequently confirmed to inhibit proline
uptake using recombinant PROT.10,11 However, only a single small
molecule inhibitor has been described for this transporter.12
Herein we report the discovery of a novel class of PROT inhibi-
tors. These compounds are potent in vitro, metabolically stable,
and exhibit good pharmacokinetics when dosed orally. The synthe-
sis, representative structure–activity relationships (SAR), and
in vivo activity are reported.
SAR and in vitro activity: In order to access in vitro potency of
Investigation into the SAR of the biphenyl moiety began with
the distal ring. We quickly found that the 20 position did not
tolerate groups larger than H or F (Table 1); for example, adding
a methyl or trifluoromethyl group at the 20 position reduced the
potency by 2- and 4-fold, respectively, as compared to the unsub-
stituted compound. Therefore, our main focus was directed at the
30 and 40 positions. Substitution at the 40 position was generally
well tolerated, and improvements in potency were realized with
small electron-withdrawing or neutral substituents such as methyl
PROT inhibitors, cell-based assays were established. The inhibition
3
levels were determined by measuring PROT mediated
H-proline
uptake in the presence of increasing compound concentration.
3
Nonspecific uptake was determined by measuring
uptake in the presence of 2 mM unlabeled proline.
H-proline
To begin our investigation of inhibitors of PROT, we analyzed
internal libraries of small molecules using high-throughput screen-
ing and a COS1 cell line expressing human PROT. A large number of
compounds were found to have inhibitory activity; however, many
of these compounds were ultimately judged to have poor ADME
properties, and poor brain penetration in particular (Fig. 1). Com-
pound 1 showed dose-dependent inhibition of PROT with an IC50
(21, PROT IC50 = 0.034
Electron-rich substituents such as methoxy led to a measurable
decrease in potency (26, PROT IC50 = 0.23 M), while some steric
lM) and chloro (23, PROT IC50 = 0.052 lM).
l
of 0.160 lM, however this compound showed poor mouse liver
S9 stability.13
Replacement of the pyridine ring included substituted phenyl
rings as well as other heteroaromatic rings (Fig. 2). The phenyl
substituted compound 2 did improve the stability of the series
(68% remaining after 30 min), but sacrificed significant potency.
Similarly, substituted phenyl rings led to less active compounds,
suggesting that substitution at the para position was not tolerated.
Only two variations retained desirable levels of potency, 2-thiazol-
yl (5) and 2-pyrimidyl (8). While the thiazole substituted com-
pound had similar potency to 1, it also had poor metabolic
stability. The modification of changing the pyridyl moiety to a less
Figure 3. Optimal phenyl orientation.
Scheme 1. Reagents and conditions: (a) triethylamine, EtOH, 100 °C; (b) 4-
bromobenzoic acid, EDC, HOBt, DIPEA, DMF; (c) 2-methylphenylboronic acid,
Pd(dppf)Cl2, K3PO4, DME, H2O, 80 °C.
Figure 2. Replacement of the pyridyl moiety.

3888
G. G. Zipp et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3886–3890
Table 1
Distal phenyl ring SAR
substituted compound 17 proved to be the most potent compound
(PROT IC50 = 0.036 M) and had good metabolic stability (S9 frac-
tion, 90% remaining at 30 min) and good selectivity over the dopa-
mine (DAT) and glycine (GlyT) transporters (DAT IC50 = 3.4 M and
l
O
2
3
l
N
R
GlyT IC50 > 10 lM, respectively). These compounds had, in general,
N
N
poor solubility and high plasma protein binding (>98%).14 To
improve these characteristics, changes to the biphenyl ring system
were pursued.
N
2'
4'
3'
Compd
R
PROT IC50 (lM)
Replacing the distal ring with saturated or heteroaromatic rings
generally led to substantial decreases in potency (Table 2). The
cyclohexyl analog and basic nitrogen-containing heterocycles were
found to be micromolar inhibitors of PROT. The exception to these
results came from substituted thiophene derivatives. Improved
potencies relative to the initial HTS compounds were observed
when thiophene rings with small substituents replaced the distal
phenyl ring (32, 33, and 34). The potencies of these compounds
were similar to the previously synthesized phenyl analogs, but
the compounds did not afford any significant improvement in
solubility or in the protein binding of the series.
Attempts to investigate the effects of the linker groups between
the two phenyl rings to possibly address the low solubility led to
the synthesis of a number of linker-spaced biphenyl analogs. As
shown in Scheme 2, the synthesis of these analogs began with
the acylation of piperidine 13 with either 4-amino- or 4-hydrox-
ybenozic acid by treatment with HOBt and EDC affording interme-
diates 35 and 36, respectively, in good yields. Subsequent
functionalization of the phenol 35 by either alkylation under
Finklestein-type reaction conditions or by Mitsunobu reaction with
appropriate alcohols afforded ethers of the general form 39. Mono-
alkylated derivatives of the aniline intermediate were prepared by
either treatment of 36 with alkyl bromides and potassium carbon-
ate at 80 °C in DMF or by reductive amination with aldehydes
employing sodium cyanoborohydride as the reducing reagent. In
the case of the reductive aminations, acetic acid and molecular
sieves were necessary to achieve good yields. Alternatively, the
anilines could be derivatized as carbamates by treatment with car-
bonyl diimidazole (CDI) and then with an alcohol to afford the
desired compounds.
8
15
16
17
18
19
20
21
22
23
24
25
26
27
H
0.28
0.653
1.26
20-Me
20-CF3
30-Cl
30-CF3
30-OCF3
30-CN
0.036
0.071
0.073
0.19
0.034
0.077
0.052
0.076
0.19
40-Methyl
40-Ethyl
40-Cl
40-CF3
40-CN
40-MeO
40-F
0.23
0.26
Table 2
Biphenyl replacements
Compd
R
PROT IC50 (lM)
8
28
29
30
31
32
33
34
Phenyl
0.28
7.01
7.08
2.23
>10
0.15
0.034
0.061
Cyclohexyl
2-Pyridyl
1-Pyrolo
1-Imidazolo
5-Chloro-2-thiophenyl
4-Methyl-2-thiophenyl
5-Methyl-2-thiophenyl
The synthesized compounds were tested for inhibition of PROT,
and Table 3 summarizes the results from these efforts. Insertion of
amide or carbamate linkers between the two phenyl rings ren-
dered the compounds inactive (42 and 43). More subtle changes
such as insertion of a carbonyl (37), oxygen (38) or amine (41)
afforded compounds that retained potency. However, while some
of these compounds did show some improved protein binding
tolerance was noted for the 4-ethyl (22, PROT IC50 = 0.077
l
M) and
4-trifluoromethyl analogs (24, PROT IC50 = 0.076
lM). Substitution
at the 30 position of the phenyl ring led to slight improvements in
potency relative to the 40 position. In particular, the 3-chloro
Scheme 2. Reagents and conditions: (a) EDC, HOBt, DMF, rt; (b) BnBr, NaI, K2CO3, acetone, 50 °C; (c) BnOH, DEAD, PPh3, THF, 0 °C to rt; (d) BnBr, K2CO3, DMF, 80 °C; (e)
PhCHO, NaCNBH3, HOAc, 4 Å molecular sieves, EtOH; (f) carbonyldiimidazole, BnOH, THF, rt.

G. G. Zipp et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3886–3890
3889
Table 3
Table 4
Linker investigation
Piperazine modifications
O
2
5
1
3
N
N
4
N
6
N
Y
Compd
X
n
PROT IC50
(
lM)
Diamine
X
8
37
38
39
40
41
42
43
—
C@O
O
O
O
NH
C(O)NH
OC(O)NH
0
0
0
1
2
1
1
1
0.28
2.70
1.95
0.27
1.02
0.87
>10
Compd
Diamine
X
Y
PROT IC50
0.70
(lM)
1
44
Cl
H
H
N
N
4
>10
N
N
45
Cl
0.053
(41, 95% bound), most did not improve either the solubility or PPB,
and metabolic stability was also problematic.15 Unable to improve
the ADME properties of the series relative to compound 37, we
chose to investigate the eastern portion of the compounds holding
the biphenyl portion of the series constant.
46
47
N
Cl
H
H
H
0.024
0.477
N
N
N
OH
Having established the SAR of the biphenyl moiety, we turned
our attention to the piperazine ring. Substitution on the piperazine
was generally well tolerated (Table 4). Substitution at the 2 and 6
positions by small alkyl groups was found to improve the potency
of the compounds. Gem dimethylation (45), 2,6-dimethyl substitu-
tion (46), and the beta epimer of the mono-methylated compounds
(51) had the most advantageous effect on the potency of the series.
Substitution at the 3 position was disfavored (47), as were larger
alkyl groups such as tert-butyl (44). Polar groups such as alcohols
(48) and carboxylic acids (49) were also generally tolerated,
though some potency was lost in certain cases (50).
48
49
Cl
Cl
H
H
0.076
0.128
N
O
N
N
OH
N
N
O
N
50
Cl
H
0.392
N
N
N
As shown in Figure 4, fragmentation of the piperazine ring to
the open chain compound 54 led to a dramatic reduction of
potency. Homologation to larger ring systems such as the 7-mem-
bered homopiperazine (53) improved the potency by twofold.
Replacement of the piperazine ring with analogous cyclic diamine
moieties yielded positive results. The fused bicyclic [3.3.0] ring sys-
tem (56) retained the potency and demonstrated improved PK
properties, specifically increased brain levels after oral dosing.
Constraining the cyclic diamine portion with bridging groups led
to compounds with improved PK and superior potency. While
the bicyclo-[2.2.1] compound (55), which forces the six-member
ring into a boat conformation, was significantly less active at inhib-
51
N
N
H
H
Cl
Cl
0.021
0.057
52
iting PROT (IC50 = 1.34
conformation, bicyclo[3.2.1]octane compounds 57 and 58, exhibit
even greater potency (PROT IC50 = 0.027 and 0.018 M, respec-
lM), the compounds that exhibit a chair
l
tively). Although these compounds continued to exhibit higher
than desired protein binding, the pharmacokinetic profiles of 57
and 58 led us to investigate them further.
Compounds 57 and 58 were determined to be quite selective
within a class of associated small molecule transporters. The dopa-
mine transporter (DAT) was not inhibited by 57 up to a maximum
tested concentration of 10
the glycine transporter. Compound 58 was slightly less selective
for DAT (IC50 = 6.1 M), but more selective for GlyT (IC50 > 10 M).
lM, and it was 37-fold selective against
l
l
The metabolic stability of both compounds was acceptable, with
75% and 70% remaining when exposed to the mouse S9 fraction,
respectively. Compound 58 was found to inhibit HERG at 59% at
10 lM concentration, and dosing of mice (IV, 1 mpk) resulted in
lethargic animals and, as a result, in vivo investigation of this com-
pound was not progressed. Following a single oral administration
of Compound 57 to C57BL/6 albino mice at 100 mpk, the com-
pound was rapidly absorbed, well tolerated and exhibited good
exposure with a Tmax of 0.38 h and a Cmax of 4.5
lg/mL and a
Figure 4. Piperazine modifications.

3890
G. G. Zipp et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3886–3890
ments. While the pharmacological effects of these compounds
have not been conclusively demonstrated, continued SAR investi-
gation of this series and the identification of compounds which
demonstrate reproducible effects in in vivo models will be subse-
quently reported.
Acknowledgment
The authors would like to thank Ken Carson for helpful discus-
sion and assistance in manuscript preparation.
Figure 5. Trace conditioning results of compound 57.
References and notes
1. Brookmeyer, R.; Johnson, E.; Ziegler-Graham, K.; Arrighi, M. H. Alzheimers
Demen. 2007, 186.
AUC0–1 of 24 lg h/mL. Analysis of total brain concentration at 6 h
2. Waldemar, G.; Dubois, B.; Emre, M.; Georges, J.; McKeith, I. G.; Rossor, M.;
Scheltens, P.; Tarsika, P.; Winblad, B. Eur. J. Neurol. 2007, 14, e1.
3. (a) Camps, P.; Munoz-Torrero, D. Mini-Rev. Med. Chem. 2002, 2, 11; (b) Benzi, G.;
Moretti, A. Eur. J. Pharmacol. 1998, 346, 1; (c) Rösler, M.; Retz, W.; Retz-
Junginger, P.; Dennler, H. J. Behav. Neurol. 1998, 11, 211; (d) Scott, L. J.; Goa, K. L.
Drugs 2000, 60, 1095; (e) Stahl, S. M. J. Clin. Psychiatry 2000, 61, 813.
4. Renick, S. E.; Kleven, D. T.; Chan, J.; Stenius, K.; Milner, T. A.; Pickel, V. M.;
Fremeau, R. T., Jr. J. Neurosci. 1999, 19, 21.
after dosing revealed a level of brain concentration of 2.26
lg/mL,
resulting in a calculated brain-to-plasma ratio of 1.7.
After demonstrating no effects distinguishable from vehicle in
an Open Field experiment,16 this compound was selected for test-
ing using a trace conditioning model to evaluate for potential cog-
nitive improvement.17 In the experiment, subjects are presented
with a neutral conditioned stimulus (CS) that is paired, after an
interval of time, with an unconditioned stimulus (US) that is averse
in nature. This is an associative learning task that by the nature of
the delay, or trace interval, will engage the hippocampus and pre-
frontal cortex. Mice learn to associate the two stimuli and will
exhibit behavioral responses such as freezing when the CS is pre-
sented alone. Mice will also associate the context, or environment,
in which the training of the CS–US occurs, and will exhibit similar
behavior even in the absence of the conditioned stimulus. Com-
pound 57 was dosed in a cohort of C57 mice (n = 13) at 100 mpk
PO along with a vehicle control group.18 As shown in Figure 5,
the results of the context freezing portion of the experiment indi-
cated a possible improvement in response to conditioned stimulus
(CS) in the drug treated subjects as compared to the vehicle control
group, however the data was not statistically significant (P = 0.16).
In summary, we have reported a novel series of highly potent
inhibitors for the mammalian proline transporter (PROT). Investi-
gation of the modular SAR of this series has yielded compounds
that have demonstrated stability in vivo, and have achieved good
plasma and brain levels in oral dosing pharmacokinetic experi-
5. Iversen, L. L. In Handbook of Psychopharmacology; Plenum: New York, 1975; Vol.
2, pp 381–442.
6. Shafgat, S.; Velaz-Faircloth, M.; Henzi, V. A.; Whitney, K. D.; Yang-Feng, T. L.;
Seldin, M. F.; Fremeau, R. T., Jr. Mol. Pharmacol. 1995, 48, 219.
7. Crump, F. T.; Fremeau, R. T., Jr.; Craig, A. M. Mol. Cell. Neurosci. 1999, 13, 25.
8. Velaz-Faircloth, M.; Guadano-Ferraz, A.; Henzi, V. A.; Fremeau, R. T., Jr. J. Biol.
Chem. 1995, 270, 15755.
9. Mouse and human versions of PROT have 97% homology at the protein level.
10. Rhoads, D. E.; Peterson, N. A.; Raghupathy, E. J. Biol. Chem. 1983, 258, 12233.
11. (a) Fremeau, R. T., Jr.; Velaz-Faircloth, M.; Miller, J. W.; Henzi, V. A.; Cohen, S.
M.; Nadler, J. V.; Shafqat, S.; Blakely, R. D.; Domin, B. Mol. Pharmacol. 1996, 49,
1033; (b) Galli, A.; Jayanthi, L. D.; Ramsey, I. S.; Miller, J. W.; Fremeau, R. T., Jr.;
DeFelice, L. J. J. Neurosci. 1999, 19, 6290.
12. Yu, X.-C.; Zhang, W.; Oldham, A.; Buxton, E.; Patel, S.; Nghi, N.; Tran, D.;
Lanthorn, T. H.; Bomont, C.; Shi, Z.-C.; Liu, Q. Neurosci. Lett. 2009, 451, 212.
13. After 30 min of incubation with the S9 fraction of mouse liver homogenate 53%
of the parent compound remained.
14. Compound 17, for instance, was determined to be 99.2% protein bound in an
equilibrium dialysis assay. Compound 23: 97.8%.
15. After 30 min of incubation with the S9 fraction of mouse liver homogenate,
45% of compound 41 and 63% of compound 39 remained.
16. Prut, L.; Belzung, C. Eur. J. Pharmacol. 2003, 463, 3.
17. (a) Bangasser, D. A.; Waxler, D. E.; Santollo, J.; Shors, T. J. J. Neurosci. 2006, 26,
8702; (b) Runyan, J. D.; Moore, A. N.; Dash, P. K. J. Neurosci. 2004, 24, 1288.
18. Burman, M. A.; Simmons, C. A.; Hughes, M.; Lei, L. J. Neurosci. Methods 2014,
222, 111.