The design and synthesis of human branched-chain amino acid aminotransferase inhibitors for treatment of neurodegenerative diseases

Article abstract of DOI:10.1016/j.bmcl.2005.07.058

The inhibition of the cytosolic isoenzyme BCAT that is expressed specifically in neuronal tissue is likely to be useful for the treatment of neurodegenerative and other neurological disorders where glutamatergic mechanisms are implicated. Compound 2 exhibited an IC₅₀ of 0.8 μM in the hBCATc assays; it is an active and selective inhibitor. Inhibitor 2 also blocked calcium influx into neuronal cells following inhibition of glutamate uptake, and demonstrated neuroprotective efficacy in vivo. SAR, pharmacology, and the crystal structure of hBCATc with inhibitor 2 are described.

Full text of DOI:10.1016/j.bmcl.2005.07.058

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2337–2340
The design and synthesis of human branched-chain amino
acid aminotransferase inhibitors for treatment
of neurodegenerative diseases
Lain-Yen Hu,* Peter A. Boxer, Suzanne R. Kesten, Huangshu J. Lei, David J. Wustrow,
David W. Moreland, Liming Zhang, Kay Ahn, Todd R. Ryder, Xiaohong Liu,
John R. Rubin, Kelly Fahnoe, Richard T. Carroll, Satavisha Dutta, Douglass C. Fahnoe,
Albert W. Probert, Robin L. Roof, Michael F. Rafferty, Catherine R. Kostlan,
Jeffrey D. Scholten, Molly Hood, Xiao-Dan Ren, Gerald P. Schielke, Ti-Zhi Su,
Charles P. Taylor, Anil Mistry, Patrick McConnell, Charles Hasemann and Jeffrey Ohren
Pfizer Global Research and Development, Ann Arbor, MI, USA
Received 13 May 2005; revised 18 July 2005; accepted 20 July 2005
Available online 6 September 2005
Dedicated to the memory of David W. Robertson, mentor, colleague and friend.
Abstract—The inhibition of the cytosolic isoenzyme BCAT that is expressed specifically in neuronal tissue is likely to be useful for
the treatment of neurodegenerative and other neurological disorders where glutamatergic mechanisms are implicated. Compound 2
exhibited an IC50 of 0.8 lM in the hBCATc assays; it is an active and selective inhibitor. Inhibitor 2 also blocked calcium influx into
neuronal cells following inhibition of glutamate uptake, and demonstrated neuroprotective efficacy in vivo. SAR, pharmacology,
and the crystal structure of hBCATc with inhibitor 2 are described.
Ó 2005 Elsevier Ltd. All rights reserved.
The cytosolic form of branched-chain aminotransferase
(BCATc), expressed in neuronal tissue, catalyzes the
transfer of an amino group from branched chain amino
acids (Leu, Ile, and Val) to a-ketoglutarate forming glu-
tamate. This appears to be a major neuronal biosynthet-
ic pathway for the de novo synthesis of glutamate.1 It
has been reported that leucine contributes ꢀ25% of
the nitrogen used for glutamate synthesis in the CNS.2
BCATc is highly expressed in the CNS and its expres-
sion is regulated.1 Therefore, inhibition of BCATc offers
an opportunity to slow glutamate synthesis and reduce
the amount of glutamate released during excitation in
neuronal tissues.2,3 The inhibition of BCATc is likely
to be useful for the treatment of a wide variety of neuro-
degenerative and behavioral disorders where glutama-
tergic mechanisms have been implicated.2,3 The
mitochondrial isoenzyme, BCATm, is more ubiquitous-
ly expressed and is found in muscle, heart, and other
peripheral tissues. BCATm is also expressed in brain
but only probably in glia.3 In brain its activity is only
about 1/4 that of BCATc. We have developed a series
of sulfonyl hydrazides, which were derived from a
high-throughput screen (HTS) hit (1). In this report,
the synthesis and SAR of these analogs and their ability
to block BCATc are described (see Fig. 1).
O
H
A
N
S
N
B
H
O
O
O
1
CF3
S
O
H
N
N
H
O
O
Cl
O
2
*
Figure 1. Inhibitors 1 and 2.
Corresponding author. E-mail: Lain-Yen.Hu@pfizer.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.07.058

2338
L.-Y. Hu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2337–2340
Table 2. SAR of the B ring of compound 1
Compound 1 exhibited a moderate ability to inhibit
hBCATc (IC50 = 2.35 lM)4 and moderate selectivity
for hBCATc versus hBCATm. We, therefore, designed
a series of analogs, aimed at improving selectivity and
solubility for in vivo proof-of-concept studies. An
SAR study of ring A of inhibitor 1 was conducted veri-
fying the effect of simple substitution. Among the Me-
substituted derivatives, ortho- (3) and meta- (4) substitu-
tion is active while para-substitution (5) is not. The Cl-
substituted analogs are less potent than their methyl
counterparts, with the 2-Cl derivative (6) being more po-
tent than the 3-Cl (7) or 4-Cl (8) compound (see Tables 1
and 2).
O
H
S
N
R
N
H
O
O
Compound A ring
HBCATc IC50, lM (n = 2–4)
1
9
4-Dibenzofuran
2.35 0.01
18.3 5.0
O
N
N
10
11
12
13
14
15
16
17
36 2.2
33.3 8.3
13.3 2.5
2.5 0.07
4.2 0.45
12.8 2.8
15 2.7
Further modification of compound 1 focused on replac-
ing the dibenzofuran ring with smaller ring systems to
improve potency and solubility for in vivo studies. We
truncated the 4-dibenzofuran moiety by eliminating
one of the benzene rings, which yielded the 2-benzofuran
(9) with decreased hBCATc activity (IC50 = 18 lM). To
improve the solubility of 1, we modified the 2-benzofu-
ran with the nitrogen-containing derivatives (2-indole
(10), 3-quinoline (11), and 7-quinoline (12)). These
nitrogen-containing molecules showed lower hBCATc
activity. Modeling a series of analogs in the active site
with SYBYL10a suggested that smaller substituents at
the 2-benzofuran 5-position could fill a depression in
the active-site surface. Therefore, the 5-bromo-2-benzo-
furanyl (13), 5-chloro-2-benzofuranyl (14), and 5-meth-
oxy-2-benzofuranyl (15) analogs were compared. The
5-bromo and 5-chloro derivatives were the preferred
inhibitors with similar potency to the 4-dibenzofuran
analog (1). In the 2-indole series, the 5-bromo-2-indole
(16) was more potent than the 5-methoxy-2-indole (17)
(see Table 3).
N
Br
O
O
O
N
N
Cl
OMe
Br
OMe
56.8 0.7
Table 3. BACTc inhibitors
O
H
O
N
S
N
H
X
O
O
Cl
With the above findings, we considered that both the
solubility and activity of our BCATc inhibitors were
critical to the in vivo proof-of-concept study. We select-
ed 5-chloro-2-benzofuran as the preferred group for the
B-ring. We incorporated 2-Me-, 3-Me-, 2-Cl-, and 2-
CF3-phenyls as the A ring and constructed inhibitors
18–20 and 2.9 Among the four of them, the CF3 analog
(2) with an IC50 0.81 lM, was the most potent inhibitor
in this series.
Compound
X
hBCATc IC50
lM (n = 2–4)
,
RLMT1/2 (min)
18
19
20
2
2-Me
3-Me
2-Cl
1.16 0.23
1.2 0.2
5.7 1.3
5
6.7
10
14
2-CF3
0.8 0.05
An X-ray crystal structure was determined for the com-
plex of hBCATc with inhibitor 2 (Fig. 2). The ligand sits
immediately adjacent to the PLP with the linker forming
several hydrogen bonds: the sulfonyl and carbonyl oxy-
gens with Ala 332 and both hydrozide NHÕs with Thr
258. Substituents at position 4 of ring A encounter steric
interference from Tyr 88. Position 3 tolerates small sub-
stituents, probably with some rotation of the A-ring.
Substituents at position 2 are not sterically restricted.
Table 1. SAR of the A ring of compound 1
O
H
O
S
A ring
N
N
H
O
O
Compound
A ring
Phenyl
hBCATc IC50, lM (n = 2–4)
˚
Fluorines of the 2-CF3 substituent lie 3.1 A from the
1
3
4
5
6
7
8
2.35 0.01
3.2 0.02
2.8 0.15
51.0 10.8
12.9 0.9
>35
phenoxyl oxygen of Tyr 159, which accept a hydrogen
bond from Arg 161 (not shown). A hydrogen bond to
the CF3 fluorine may stabilize the complex. The B ring
lies against a hydrophobic side of the pocket and mainly
interacts with Met 259. The furanyl oxygen may accept a
hydrogen bond from the thiol of Cys 333. The 5-chloro
2-Me-Phenyl
3-Me-Phenyl
4-Me-Phenyl
2-Cl-Phenyl
3-Cl-Phenyl
4-Cl-Phenyl
37.2 3.9

L.-Y. Hu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2337–2340
2339
Compound 2 decreased calcium influx in neuronal cul-
tures with an IC50 = 4.8 1.2 lM (n = 4).
Following a 30 mg/kg subcutaneous injection to Lewis
rats, inhibitor 2 reached a peak plasma concentration
(Cmax) of 8.28 lg/ml at 0.5 h (tmax). The mean plasma
exposure (AUC) value was 19.9 lg h/ml, and the mean
terminal half-life ranged from 12 to 15 h, indicating
favorable PK parameters of inhibitor 2. Since BCATc
is involved in the synthesis of glutamate, a model of slow
neurodegeneration was selected for the in vivo evalua-
tion of inhibitor 2. Daily administration of the mito-
chondrial neurotoxin, 3-nitroproprionic acid7 (3-NP),
has been shown to produce striatal lesions, which leads
to motor deficits. Figures 3A and B illustrate the deficits
in rotorod and beam walking performance and demon-
strate that inhibitor 2 (administered for 9 days) was able
to almost completely reverse the effects of 3-NP8. In
addition, histological examination of the brains from
these animals demonstrated significant reduction in the
number of sections with degenerating neurons using a
silver degeneration stain11 (Fig. 4).
Figure 2. X-ray crystal structure of compound 2 bound to human
cystosolic BCAT.10b
In summary, a series of potent and selective inhibitors of
BCATc has been developed and the X-ray crystal struc-
ture with inhibitor 2 bound to BCATc was obtained.
Using inhibitor 2 as a prototype, neuroprotection has
substituent contacts an amide NH from Gln 242 (not
shown).
25
20
15
Inhibitor 2 showed a favorable in vitro profile (potency,
selectivity). To insure that the compoundÕs selectivity for
the cystosolic enzyme in rat, inhibitor 2 was tested in a
recombinant rat BCATc assay and a crude rat BCATm5
assay. The IC50 at rat BCATc was 0.2 lM 0.02 (n = 2)
vs 3.0 lM 0.5 (n = 5) at rat BCATm. This selectivity
of approximately 15-fold justified further evaluation in
in vitro and in vivo assays. Inhibitor 2 was tested in a
neuronal cell culture model, which measures 45[Ca] in-
flux into neurons following treatment with the gluta-
mate uptake inhibitor, 1-pyrrolidine dicarboxylic acid.
This assay was specifically developed to evaluate
BCATc inhibitors assuming that reducing the source
of glutamate would reduce the downstream effects of
glutamate, such as calcium influx and cell death6.
*
10
5
0
3-NP +
Vehicle +
Cmpd 2
3-NP +
Vehicle
Cmpd 2
Figure 4. Histology following 3-NP administration and inhibitor 2.
Asterisks indicate p < 0.05 versus 3-NP plus vehicle treatment.
A
50
40
30
20
10
0
B
*
15
12
9
*
6
3
0
3-NP +
Vehicle
3-NP +
Cmpd 2
Vehicle +
Cmpd 2
3-NP +
Vehicle
3-NP +
Cmpd 2
Vehicle +
Cmpd 2
Figure 3. Effect of Inhibitor 2 on (A) rotorod performance and (B) beam walking in animals treated for 9 days with the 3-NP. Means SEM.
Asterisks indicate p < 0.05 versus 3-NP plus vehicle treatment.

2340
L.-Y. Hu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2337–2340
CO2 incubator at 37 °C and then spiked with 10 ll of
either dH2O or 26 mM L-trans-pyrrolidine-2,4-dicarboxy-
late (LPDC) [500 lM final concentration]. Plates were
then store in a humidified CO2 incubator at 37 °C. 45Ca
uptake was determined thirty minutes following the
addition of LPDC. Cultures were washed trice with
500 ml HBSS and then dissolved with lysate buffer.
Individual lysates were combined with 10 ml scintillation
cocktail and counted for emissions using a Tricarb
2900TR Packard Liquid Scintillation Analyzer; (b) Mar-
coux, F. W.; Probert, A. W. Stroke 1988, 19, 148.
been demonstrated both in vitro and in vivo. This sug-
gests that inhibitors of BCATc may represent a novel
therapeutic strategy for the treatment of neurological
disorders, where glutamate toxicity has been implicated.
References and notes
1. (a) Hutson, S.; Berkich, D.; Drown, P.; Xu, B.; Aschner,
M.; LaNoue, K. J. Neurochem. 1998, 71, 863; (b) Hall, T.;
Wallin, R.; Reinhart, G.; Hutson, S. J. Biol. Chem. 1993,
268, 3092; (c) Davoodi, J.; Drown, P.; Bledsoe, R.; Wallin,
R.; Reinhart, G.; Hutson, S. J. Biol. Chem. 1998, 273,
4982.
2. (a) Kholodilov, N.; Neystat, M.; Oo, T.; Hutson, S.;
Burke, R. Mol. Brain Res. 2000, 75, 281; (b) Lieth, E.;
LaNoue, K.; Berkich, D.; Xu, B.; Ratz, M.; Taylor, C.;
Hutson, S. J. Neurochem. 2001, 76, 1712.
7. Schulz, J. B.; Matthews, R. T.; Klockgether, T.; Dichgans,
J.; Beal, M. F. Mol. Cell. Biochem. 1997, 174, 193.
8. The study began with 3 days of dosing with compound 2
at 30 mg/kg, SC, BID (60 mg/kg/day) and continued daily
for 14 days. The mitochondrial inhibitor 3-nitropropionic
acid (3-NP, 46 mg/kg, IP) was administered every other
day up to day 9 (day 1 defined by the first day of 3-NP
dosing). Behavioral testing was done at baseline, during 3-
NP administration (day 6), at day 14 (3 days after the end
of compound 2 dosing), and again at day 20. Data in
Figure 3 represent the maximum effect seen on two of the
motor function endpoints. As evaluated by ANOVA,
there was an overall statistical difference for each of the
behavioral endpoints, and Bonferroni corrected post hoc
t-tests were used for specific comparisons. After dosing
and behavioral tests, the rats were cardiac-perfused,
washed with cold saline, and brains were removed. The
right hemispheres were immediately immersed into 4%
paraformaldehyde and kept under 4 °C for at least a week.
The brain samples were then paraffin embedded and
continuous brain sections (40 lm in thickness) were
obtained from olfactory bulb to the end of occipital
cortex. Every 8th section was processed for cupric silver
staining that demonstrates degenerative areas.11 For
evaluation of brain damage, sections were examined with
a light microscope for detecting positive staining black
areas that implied areas of neuronal death. Total
number of sections with the positive findings from each
group was finally compared with other groups (Newman–
Keuls).
3. Jouvet, P.; Rustin, P.; Taylor, D. L.; Pocock, J. M.;
Felderhoff-Mueser, U.; Mazarakis, N. D.; Sarraf, C.;
Joashi, U.; Kozma, M.; Greenwood, K.; Edwards, A. D.;
Mehmet, H. Mol. Biol. Cell 2000, 11, 1919.
4. (a) Determination of IC50 values for BCAT inhibitors. The
reaction mixtures (200 ll) contained 25 mM NaPi (pH
7.8), 0.75 mM L-leucine, 0.5 mM a-ketoglutarate, 10 lM
pyridoxal phosphate, 1 mM NAD+, 4 mM ADP, 2 mM
dithiothreitol, 12.5 mM EDTA, the indicated compound,
saturating level of glutamate dehydrogenase (GDH), and
50–100 nM hBCATc. The final concentration of DMSO
was kept at 1%. The reactions were carried out in 96-well
clear polystyrene plates. The reaction was incubated at
37 °C and the absorbance at 340 nm was collected over a
period of 20 min with readings using a SpectraMax
Microplate Spectrophoto-meterÒ (Molecular Devices,
Palo Alto, CA) equipped with Softmax ProÒ software
(Molecular Devices, Palo Alto, CA). IC50s were calculated
from 8 to 10 point curves. Human BCATc was expressed
in Escherichia coli and purified according to the method
described.1c Glutamate dehydrogenase from bovine liver
(Sigma) was purified further, as previously described.4b
The amount of GDH was determined for each preparation
to ensure the saturating level for the assay; (b) OÕCarra, P.;
Griffin, T.; OÕFlaherty, M.; Kelly, N.; Mulcahy, P.
Biochim. Biophys. Acta 1996, 1297, 235.
9. (a) All the compounds were synthesized by coupling the
aryl sulfonyl chloride and aryl carboxylic acid hydrazide
in pyridine, as described in compound 2. Compound 2:
A
solution of 2-trifluorophenyl sulfonyl chloride
(1 mmol) and 5-chloro-benzofuran-2-carboxy-lic acid
hydrazide (1 mmol) was stirred for 18 h at room
temperature in pyridine (10 ml). After the pyridine was
removed in vacuo, the residue was washed with water
and purified by column over SiO2 eluting with 1:2
EtOAc/hexanes (60% yield). MS: 419.9 (M + 1 for
5. Inhibitors were solubilized in DMSO at a primary
concentration of 10 mM and then diluted to give a final
concentration between 0.3 and 100 lM. The assay is done
in a 96-well plate. Total reaction volume of 100 ll
consisted of 58 ll reaction mixture with 2ll of specified
inhibitor and 38 ll water. The reaction was started with
2 ll BCAT enzyme. The 96-well reaction plate was
incubated for 30 min at 37 °C. The reaction was stopped
with 100ll of 2% acetic acid (1% Final) and analyzed for
glutamate with LC/MS/MS.
6. (a) Cultures were prepared, as described in Ref. 6b. Ten-
to fourteen-day-old cerebrocortical cultures (48-well
plates) were washed once with 500 ml of 37 °C HBSS
and then plated with 500 ll of 37 °C supplemented MEM
(see formula below). A 2 ll volume of test agents (in 100%
DMSO) was then spiked in at 260 times final desired
concentrations. Plates were stored in a humidified CO2
incubator at 37 °C for 2 h by which time 30 lCi/ml 45Ca in
MEM [0.60 lCi/ml final concentration] was added. Plates
were then stored for one additional hour in a humidified
C16H10N2O4SF3Cl);
mp:
192–193.4 °C.
Analysis
(C16H10N2O4SF3Cl) calcd. C: 45.89, H: 2.41, N: 6.69,
S: 7.66, F: 13.61. Found: C: 46.16, H: 2.41, N: 6.60, S:
7.38, F: 13.30. 1H NMR (DMSO-d6) d 7.44 (dd, 1H,
J = 9.04 and 2.20 Hz), 7.53 (s, 1H), 7.63 (d, 1H,
J = 8.79 Hz), 7.75–7.82 (m, 2H), 7.84 (d, 1H,
J = 1.95 Hz), 7.93 (dd, 1H, J = 2.2 and 9.03 Hz), 8.13
(dd, 1H, J = 2.44 and 6.59 Hz), 10.19 (s, 1H), and 11.06
(s, 1H); (b) Hu, L.-Y.; Kesten, S.R.; Lei, H.; Wustrow,
D.J.; Ryder, T.R.E.P. EP 1,314,723, 2003.
10. (a) Tripos Associates, St. Louis, MO; (b) Coordinates
have been deposited with the RCSB Protein Data Bank
with the ID cods 2ABJ.
11. Switzer, R. C. Toxicol. Pathol. 2000, 28, 70.