PTP1B inhibition and antihyperglycemic activity in the ob/ob mouse model of novel 11-arylbenzo[b]naphtho[2,3-d]furans and 11-arylbenzo[b]naphtho[2,3- d]thiophenes

Article abstract of DOI:10.1021/jm990260v

We have reported on a series of novel, potent, and selective PTPase inhibitors that function as oral antidiabetic agents. Compounds 6 and 7 are particularly notable in this regard. They were both low nanomolar inhibitors and lowered glucose in the diabeti

Full text of DOI:10.1021/jm990260v

J . Med. Chem. 1999, 42, 3199-3202
3199
insulin, and lower plasma glucose in diabetic animal
models.9 These studies all suggest that a selective, orally
active PTP1B inhibitor could function as an agent that
normalizes plasma glucose without inducing hypogly-
cemia; therefore, such an agent could be a major
advance in the treatment of type II diabetes. We disclose
here some of our efforts leading to potent and selective
PTP1B inhibitors that are orally active as antidiabetic
agents.
Ch em istr y. According to Scheme 1, 2-benzylben-
zothiophene10 (9) and 2-benzylbenzofuran10 (10) were
acylated with p-anisoyl chloride under standard Friedel-
Crafts conditions to provide 3-acyl derivatives 11 and
12, respectively (method a). Compound 11 was demeth-
ylated with pyridinium hydrochloride and dibrominated
to provide the benzbromarone analogue 2 (methods b
and c). Treatment of 11 or 12 with boron tribromide
etherate resulted in demethylation followed by concomi-
tant intramolecular cyclization and aromatization to
produce the 11-p-phenoxybenzo[b]naphtho[2,3-d]thio-
phene (13) and 11-p-phenoxybenzo[b]naphtho[2,3-d]-
furan (14) derivatives, respectively (method d). Bromi-
nation of 13 and 14 with 3 equiv of molecular bromine
resulted not only in dibromination ortho to the phenol
moiety but also in bromination at the 6 position to afford
tribromo derivatives 3 and 15 (method e).
Alkylation of phenol 3 with methyl bromoacetate
followed by methyl ester hydrolysis with aqueous potas-
sium hydroxide produced acetic acid congener 4 (method
f). Phenol 3 was also alkylated with diethyl trifluoro-
methanesulfonoxymethylphosphonate,11 and this (bis)-
ethyl ester derivative was further converted to the
phosphonic acid compound 5 using iodotrimethylsilane
(method g). Treatment of either phenol 3 or 15 with (S)-
2-hydroxy-3-phenylpropionic acid, methyl ester under
Mitsunobu conditions followed by methyl ester hydroly-
sis with aqueous potassium hydroxide gave (R)-2-
benzyloxyacetic acid derivatives 6 and 7, respectively
(method h). Likewise, treatment of phenol 3 with (S)-
2-hydroxy-1,3-dioxo-2-isoindolinebutyric acid, methyl
ester under Mitsunobu conditions followed by methyl
ester removal using iodotrimethylsilane gave (R)-2-
alkylated oxyacetic acid analogue 8 (method i).
Resu lts a n d Discu ssion . The inhibitory activity of
our compounds against recombinant PTP1B12 was as-
sessed using, as substrate, the phosphotyrosyl dode-
capeptide TRDI(P)YETD(P)Y(P)YRK, corresponding to
the 1142-1153 insulin receptor kinase regulatory do-
main, phosphorylated on the 1146, 1150, and 1151
tyrosine residues. Enzyme reaction progression was
monitored via the release of inorganic phosphate as
detected by the malachite green-ammonium molybdate
method.13
P TP 1B In h ibition a n d
An tih yp er glycem ic Activity in th e
ob/ob Mou se Mod el of Novel
11-Ar ylben zo[b]n a p h th o[2,3-d ]fu r a n s a n d
11-Ar ylben zo[b]n a p h th o[2,3-d ]th iop h en es
J ay Wrobel,* J anet Sredy,† Christopher Moxham,
Arlene Dietrich, Zenan Li, Diane R. Sawicki,†
Laura Seestaller, Li Wu,‡ Alan Katz,
Donald Sullivan, Cesario Tio, and Zhong-Yin Zhang‡
Wyeth-Ayerst Research, Inc., CN 8000,
Princeton, New J ersey 08543-8000, and Department of
Molecular Pharmacology, Albert Einstein College of
Medicine, Bronx, New York 10461
Received May 25, 1999
In tr od u ction . Non-insulin-dependent diabetes mel-
litus (type II) represents 80-90% of the human popula-
tion with diabetes, and worldwide estimates approach
215 million sufferers by 2010.1 In the patients with
diabetes, maintaining tight glycemic control appears to
be essential to prevent diabetic complications.2 However,
this is difficult since the current therapies for type II
diabetes have inherent problems including compliance,
ineffectiveness, and hypoglycemic episodes with insulin
and the sulfonylureas. Recently, troglitazone (marketed
as Rezulin) has been introduced as an agent that
enhances insulin action and lowers plasma glucose.3
However, troglitazone is not effective in all type II
patients; therefore, there still remains a great need for
more effective, orally administered agents,4 particularly
ones that normalize both glucose and insulin levels.
Excess hepatic glucose production and hepatic, skel-
etal muscle, and adipose tissue insulin resistance are
prime factors which contribute to type II diabetes.5
Insulin resistance is associated with a deficit in protein-
tyrosine phosphorylation in the insulin signal trans-
duction cascade. This deficit in tyrosine phosphorylation
in insulin-responsive tissues, muscle, liver, and adipose,
causes a reduction in the metabolic actions of insulin
and hyperglycemia. The decrease in protein-tyrosine
phosphorylation in cells does not appear to be due to
an inherent problem in the insulin receptor (IR) tyrosine
kinase but instead is caused by an elevation in protein-
tyrosine phosphatase activity.6 These enzymes (possibly
both transmembrane and intracellular PTPases) de-
phosphorylate the active form (triphosphorylated in the
regulatory domain) of the IR and attenuate its tyrosine
kinase activity. One PTPase, in particular, the intra-
cellular, nonreceptor PTPase PTP1B, appears to play a
major role in the dephosphorylation of the IR on the
basis of many biochemical and cellular studies7 and
according to a pivotal study with PTP1B knockout
mice.8 Furthermore, vanadium-containing inhibitors of
PTPases have been shown to increase IR tyrosine
phosphorylation, mimic cellular and in vivo actions of
Early in our lead screening program, we discovered
that the potent uricosuric agent benzbromarone14 (1)
was a weak rat PTP1B inhibitor (IC50 ) 26 µM). Since
this compound was a highly efficacious oral agent with
excellent absorption and pharmacokinetic parameters,15
we thought it a good starting point for a PTP1B
inhibitor analogue program. Standard SAR transforma-
tions on 1 led only to modest improvements in potency.
* Please address all correspondence to this author at Wyeth-Ayerst
Research, Inc., 145 King of Prussia Rd, Radnor, PA 19087.
‡ Albert Einstein College of Medicine.
† Present address: Institute for Diabetes Discovery, 23 Business
Park Dr, Branford, CT 06405.
10.1021/jm990260v CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/10/1999

3200 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17
Communications to the Editor
Sch em e 1a
a
Reagents: (a) p-anisoyl chloride, SnCl4, CS2, rt, 9 to 11 (85%), 10 to 12 (84%); (b) pyridinium hydrochloride, 228 °C (90%); (c) Br2,
HOAc, H2O, rt (95%); (d) BBr3, CH2Cl2, -78 °C to rt, 11 to 13 (56%), 12 to 14 (42%); (e) Br2, HOAc, rt, 13 to 3 (KOAc added, 95%), 14 to
15 (83%); (f) 3 to 4, BrCH2CO2Et, K2CO3, DMF, rt (98%)/KOH, H2O, THF, MeOH (91%); (g) 3 to 5, TfOCH2PO3Et2, NaH, THF, 0 °C to
rt (88%)/TMSI, CH2Cl2, 0 °C to rt (44%); (h) (S)-HOCH(CH2Ph)CO2Et, diethyl azodicarboxylate, PPh3, benzene, reflux (3 to 6, 90%; 3 to
7, 82%)/KOH, H2O, THF, MeOH (for 6, 67%; for 7, 98%); (i) 3 to 8, (S)-R-hydroxy-1,3-dioxo-2-isoindolinebutyric acid, methyl ester, diethyl
azodicarboxylate, PPh3, benzene, reflux (81%)/TMSI, CH2Cl2, -10 °C to rt (52%).
Ta ble 1
For example, extension of the 2-ethyl moiety of 1 to a
4-day ob/ob mouseg
(% decrease)
benzyl group to provide benzothiophene 2 (IC50 ) 4 µM,
rPTP1B) improved inhibition values to the low micro-
molar range. We decided to build rigid analogues of 1
and 2 in an effort to study the effect of the conformation
of the dibromophenolic residue with respect to the
benzothiophene nucleus. This led to the tribrominated
(benzo[b]naphtho[2,3-d]thiophene-11-yl)phenol 3 (IC50
) 0.23 µM, rPTP1B), the first of our analogues to break
the micromolar barrier.
hPTP1B
IC50 (nM)a (mg/kg/day)
doseb
compd
glucosec
insulinc
3
4
5
6
384
386
145
61
100
100
90
75
50
25
10
25
10
5
d
23
d
d
61
d
50
35
26
26
42
29
d
85
71
72
d
86
69
56
e
7
8
83
11
1 (ip)
38
d
100
100
200
d
ciglitazonef
43
34
39
d
troglitazonef
a
Substrate is the triphosphorylated peptide TRDI(P)YETD-
(P)Y(P)YRK. The reported IC50 values are from direct regression
b
curve analyses that are significant (p < 0.05). All doses are po
unless otherwise indicated. c All values from drug-treated mice,
other than d, are significant vs vehicle-treated mice: p < 0.05.
d
Less than 15% decrease at dose tested. e Not determined. f Ref-
g
erence standard. Plasma glucose and insulin values of 200-350
mg/dL and 300-775 µIU/mL, respectively, in untreated mice.
However, the highly lipophilic and poorly water-
soluble 3 did not demonstrate efficacy at high doses in
an oral study done in our primary in vivo model, the
ob/ob mouse (vide infra). An acetic acid moiety was
introduced onto the phenol of 3, to afford the oxyacetic
acid 4, in an effort to improve water solubility and,
hence, intestinal absorption. Compound 4 retained
PTP1B potency (IC50 ) 0.15 µM, rPTP1B) and was
active orally in the ob/ob mouse model. Around this time
we received the human clone of the PTP1B protein, and
compounds 3 and 4 demonstrated comparable potency
against this human enzyme (IC50’s of 384 and 386 nM,
respectively, Table 1).
We surmised that compound 4 was binding to the
active site of PTP1B and, therefore, hypothesized that
the acetic acid residue of 4 might bind to the side chain

Communications to the Editor
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17 3201
Ta ble 2
IC50 (µM)
PTPasea
hPTP1B
LAR
PTP1R (LRP)
CD45
PTP1C (SH-PTP1)
hPTPâ
PTP-PEST
HePTP360
VHR
4
7
0.39 ( 0.1
1.5 ( 0.2
17 ( 3
9.1 ( 3.2
5.0 ( 3.3
0.60 ( 0.3
2.3 ( 0.7
2.7 ( 0.6
7.9 ( 1.6
0.08 ( 0.01
0.42 ( 0.08
13 ( 3
6.3 ( 2.4
9.0 ( 5
0.5 ( 0.2
2.2 ( 0.9
2.1 ( 0.6
11 ( 3
a
All PTPases are human. Substrate for all PTPases is the
triphosphorylated peptide TRDI(P)YETR(P)Y(P)YRK at 50 µM
concentration. Three independent measurements were performed
for IC50 determinations. Similar results were obtained in multiple
measurements. The reported values are the average of all experi-
ments, and the errors are standard deviations.
F igu r e 1. Compound 6 docked into the active site of PTP1B
with the R-benzyl group of 6 shown in magenta.
fold more potently than PTPR, which has also been
implicated as a negative regulator of insulin signaling.
Compound 7, however, was generally more than 25
times selective for PTP1B over other PTPases. The
exception was hPTPR, where compound 7 was 7-fold
selective.
The compounds were also evaluated as antidiabetic
agents in the ob/ob mouse model.20 Insulin resistance
in this model has been associated with a reduction in
insulin-induced protein-tyrosine phosphorylation in tis-
sues such as liver, and markedly elevated PTPase
activities in these tissues have been observed leading
to the conclusion that PTPases may cause or contribute
to the reduced phosphorylation of the IR.21 Increased
abundance of PTP1B in the livers of ob/ob mice has also
been noted.22
As stated previously, compound 4 showed oral glu-
cose- and insulin-lowering activity at a dose of 100 mg/
kg/day in the ob/ob mouse model. This result served as
a proof-of-concept that justified further testing of inhibi-
tors in this model. In fact, much more robust oral activ-
ity was seen with (R)-benzyl congeners 6 and 7. Both
of these compounds showed a dose-response relation-
ship in reductions of both glucose and insulin. The mini-
mum effective dose (MED) for plasma glucose lowering
for 7 was between 5 and 10 mg/kg/day, although the
MED for insulin lowering with this compound was not
reached in this dose-response study. Compound 7 was
more potent when administered ip, showing a strong
glucose-lowering response at 1 mg/kg/day. The potent
inhibitor 8 was not active at the highest dose tested (100
mg/kg). Although compound 8 has a calculated log P
value similar to that of 7, the precise nature for the
reduced potency in vivo (i.e. solubility, absorption,
metabolism, etc.) was not determined. Compound 7 was
further found to exhibit high mean plasma concentra-
tions (up to 120 µg/mL at the 4-h time point) after a
single 10 mg/kg oral dose in ob/ob mice over a 24-h
period, and these levels were not significantly different
when compared to ip administration. This demonstrated
that compound 7 was well-absorbed and did not undergo
extensive first-pass metabolism.
of active site Arg 221 via a charge-charge interaction
and to the main chain amides of the phosphate binding
loop via hydrogen bond interactions similar to the
tyrosine phosphate residue of a peptide substrate un-
dergoing phosphate hydrolysis.16 We were encouraged
further with this hypothesis when we found that phos-
phonate analogue 5 showed a slight potency increase
over 4 (IC50 of 145 versus 386 nM, Table 1). The X-ray
crystal structure of PTP1B17 was used as a starting
point for our modeling efforts, and compound 4 was
docked into the active site and minimized with the acetic
acid moiety pointing toward the catalytic sequence
residues V213HCSAGIGR221SG of PTP1B. This docking
experiment revealed the existence of a large pocket of
which the inhibitor 4 was not taking advantage. We
believed that a lipophilic residue emanating from the
acetic acid R-carbon could fill this pocket and enhance
the binding interactions of the inhibitor with the
enzyme. In fact, placement of a (R)-benzyl moiety in the
R position resulted in compound 6 (Figure 1) and a 6-fold
potency increase to an IC50 of 61 nM. The furan
derivative of 6, compound 7, was also synthesized and
showed similar potency to 6 (IC50 of 83 nM, Table 1).
Further optimization at the R-carbon of 6, via replace-
ment of the (R)-benzyl group by a (R)-2-phthalimido-
ethyl side chain, gave compound 8. With an IC50 of 11
nM (Table 1), this compound was 6 times more potent
than the benzyl derivative 6. Furthermore a kinetic
analysis, using the monophosophotyrosine-containing
insulin receptor peptide TRDIYETD(P)YYRK, as sub-
strate, revealed that 8 bound competitively with a
potent Ki of 3 nM.18
The PTPase domains of receptor and nonreceptor
PTPases are highly conserved with ∼35% mean se-
quence identity between known phosphatases.19 There-
fore it is critical that inhibitors of PTPases used for
therapeutic purposes show requisite selectivity, espe-
cially in vivo. To address this issue at the in vitro level,
compounds 4 and 7 were evaluated as inhibitors against
a battery of PTPases presented in Table 2 using the
triphosphorylated IR peptide as substrate. Both com-
pounds showed only modest selectivity (3-6-fold, re-
spectively) over LAR, a PTPase also implicated as a
negative regulator of the insulin receptor.6 On the other
hand, compounds 4 and 7 inhibited PTP1B 40- and 160-
Su m m a r y. We have reported on a series of novel,
potent, and selective PTPase inhibitors that function as
oral antidiabetic agents. Compounds 6 and 7 are
particularly notable in this regard. They were both low
nanomolar inhibitors and lowered glucose in the diabetic

3202 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 17
Communications to the Editor
L.; Kennedy, B. P. Increased Insulin Sensitivity and Obesity
Resistance in Mice Lacking the Protein Tyrosine Phosphatase-
1B Gene. Science 1999, 283, 1544-1548.
ob/ob mouse at doses at or below 10 mg/kg/day, po.
Compound 7 generally showed selectivities against
other PTPases of greater than 25-fold. These compounds
and other analogues are currently being further evalu-
ated.
(9) Evans, J . L.; J allal, B. Protein Tyrosine Phosphatases: Their
Role in Insulin Action and Potential as Drug Targets. Exp. Opin.
Invest. Drugs 1999, 8, 139-160.
(10) Nutaitis, C. F.; Charles, F.; Patragnoni, R.; Goodkin, G.; Neigh-
bour, B.; Obaza-Nutaitis, J . Reduction of Heterocyclic Alcohols
with Sodium Borohydride-Trifluoroacetic acid. Preparation of
Bis-Heterocyclic Methanes. Org. Prep. Proced. Int. 1991, 23,
403-411.
(11) Phillion, D. P.; Andrew, S. S. Synthesis and Reactivity of Diethyl
Phosphonomethyltriflate. Tetrahedron Lett. 1986, 1477-1480.
(12) Barford, D.; Keller, J . C.; Flint, A. J .; Tonks, N. K. Purification
and Crystallization of the Catalytic Domain of Human Protein
Tyrosine Phosphatase 1B Expressed in Escherichia coli. J . Mol.
Biol. 1994, 239, 726-730.
(13) Lanzetta, P. A.; Alvarez, L. J .; Reinach, P. S.; Candia, O. A. An
Improved Assay for Nanomole Amounts of Inorganic Phosphate.
Ann. Biochem. 1979, 100, 95-97.
Ack n ow led gm en t. We are grateful to the Discovery
Analytical Department of Wyeth-Ayerst for elemental
analyses, 400-MHz 1H NMR, mass spectroscopy, and
HPLC data and to Lauryl Van Neste for plasma level
analysis of compound 7.
Su p p or tin g In for m a tion Ava ila ble: Details of the syn-
thesis and analysis of compounds 2-8 and procedures for the
biological evaluations of these compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
(14) Yu, T.-F. Pharmacokinetic and Clinical Studies of a New
Uricosuric Agent-Benzbromarone. J . Rheumatol. 1976, 3, 305-
312.
(15) Ferber, H.; Vergin, H.; Hitzenberger, G. Pharmacokinetics and
Biotransformation of Benzbromarone in Man. Eur. J . Clin.
Pharm. 1981, 19, 431-435.
(16) J ia, Z.; Barford, D.; Flint, A. J .; Tonks, N. J . Structural Basis
for Phosphotyrosine Peptide Recognition by Protein Tyrosine
Phosphatase 1B. Science 1995, 268, 1754-1762.
(17) Barford, D. J .; Flint, A. J .; Tonks, N. K. Crystal Structure of
Human Protein Tyrosine Phosphatase 1B. Science 1994, 263,
1397-1404.
(18) A detailed report on the kinetics of this compound and other
related PTPase inhibitors will be submitted for publication.
(19) Barford, D. J .; J ia, A.; Tonks, N. K. Protein Tyrosine Phos-
phatases Take Off. Nature Struct. Biol. 1995, 2, 1043-1053.
(20) Wrobel, J .; Li, Z.; Dietrich, A.; McCaleb, M.; Mihan, B.; Sredy,
J .; Sullivan, D. Novel 5-(3-Aryl-2-propynyl)-5-(Arylsulfonyl)-
thiazolidine-2,4-diones as Antihyperglycemic Agents. J . Med.
Chem. 1998, 41, 1084-1091.
Refer en ces
(1) Bennett, P. H. Primary Prevention of NIDDM: A Practical
Reality. Diabet. Metab. Rev. 1997, 13, 583-605.
(2) The Diabetes Control and Complications Trial Research Group.
The Effect of Intensive Treatment of Diabetes on the Develop-
ment and Progression of Longterm Complications in Insulin-
Dependent Diabetes Mellitus. N. Engl. J . Med. 1993, 329, 977-
986.
(3) Kumar, S.; Boulton, A. J . M.; Beck-Nielsen, H.; Berthezene, F.;
Muggeo, M.; Persson, B.; Spinas, G. A.; Donoghue, S.; Lettis,
S.; Stewart-Long, P.; Group, F. T. T. Troglitazone, an Insulin
Action Enhancer, Improves Metabolic Control in NIDDM Pa-
tients. Diabetologia 1996, 30, 701-709.
(4) Witcher, J . W. Implications of the DCCT. A Pharmaceutical
Industry Viewpoint and Commentary. Diabetes Rev. 1994, 2,
272-276.
(5) Kahn, C. R. Insulin Action, Diabetogenes, and the Cause of Type
II Diabetes. Diabetes 1994, 43, 1066-1084.
(6) Goldstein, B. J .; Ahmad, F.; Ding, W.; Li, P.-M.; Zhang, W.-R.
Regulation of the Insulin Signaling Pathway by Cellular Protein-
Tyrosine Phosphatases. Mol. Cell. Biochem. 1998, 182, 91-99.
(7) Byon, J . C. H.; Kusari, A. B.; Kusari, J . Protein-Tyrosine
Phosphatase-1B acts as a Negative Regulator of Insulin Signal
Transduction. Mol. Cell. Biochem. 1998, 182, 101-108.
(8) Elchebly, M.; Payette, P.; Michaliszyn, E.; Cromlish, W.; Collins,
S.; Loy, A. L.; Normandin, D.; Cheng, A.; Himms-Hagen, J .;
Chan, C.-C.; Ramachandran, C.; Gresser, M. J .; Tremblay, M.
(21) Sredy, J .; Sawicki, D. R.; Flam, B. R.; Sullivan, D. Insulin
Resistance is Associated with Abnormal Dephosphorylation of
a Synthetic Phosphopeptide Corresponding to the Major Auto-
phosphorylation of the Sites of the Insulin Receptor. Metabolism
1995, 44, 1074-1081.
(22) Sredy, J . Unpublished data.
J M990260V