Identification of a novel class of orally active pyrimido[5,4-3][1,2,4] triazine-5,7-diamine-based hypoglycemic agents with protein tyrosine phosphatase inhibitory activity

Article abstract of DOI:10.1016/S0960-894X(03)00623-1

A novel series of orally active pyrimido[5,4-3][1,2,4]triazine-5,7-diamine- based hypoglycemic agents have been identified. These compounds show non-selective inhibitory properties against a panel of protein tyrosine phosphatases including PTP1B. Compounds 12 and 13 display oral glucose lowering effects in ob/ob mice.

Full text of DOI:10.1016/S0960-894X(03)00623-1

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2895–2898
Identification of a Novel Class of Orally Active
Pyrimido[5,4-3][1,2,4]triazine-5,7-diamine-Based Hypoglycemic
Agents with Protein Tyrosine Phosphatase Inhibitory Activity
Kevin R. Guertin,* Lina Setti, Lida Qi, Rachel M. Dunsdon, Brian W. Dymock,
PhilipS. Jones, Hilary Overton, Mathew Taylor, Glyn Williams, Joseph A. Sergi,
Karen Wang, Ying Peng, Marcia Renzetti, Rogely Boyce, Fiorenza Falcioni,
Ralph Garippa and Andree R. Olivier
Roche Research Center, Hoffmann-LaRoche Inc., 340 Kingsland St., Nutley, NJ07110, USA
Received 21 January 2003; accepted 14 May 2003
Abstract—A novel series of orally active pyrimido[5,4-3][1,2,4]triazine-5,7-diamine-based hypoglycemic agents have been identified.
These compounds show non-selective inhibitory properties against a panel of protein tyrosine phosphatases including PTP1B.
Compounds 12 and 13 display oral glucose lowering effects in ob/ob mice.
# 2003 Elsevier Ltd. All rights reserved.
Insulin resistance in the liver and peripheral tissues
coupled with elevated fasting plasma glucose and
impaired glucose tolerance are the hallmarks of Type 2
diabetes.1 The inability of endogenous insulin to sup-
press postprandial hepatic glucose output and stimulate
glucose uptake and disposal by the peripheral tissues
results in elevated plasma glucose levels. A sustained
elevation in plasma glucose eventually culminates in an
number of debilitating diabetic complications such as
atherosclerosis, coronary artery disease, nephropathy,
neuropathy and retinopathy.2 Thus, tight control of
blood glucose levels is essential in the early stages of the
disease.
domain) of the IR, resulting in attenuation of the intri-
sic tyrosine kinase activity of the receptor. The intra-
cellular PTPase PTP1B has been demonstrated to play a
significant role in the dephosphorylation of the IR.5
Furthermore, PTP1B gene knockout mice have shown
improved insulin sensitivity and resistance to weight
gain.6 Thus, small molecule PTP1B inhibitors have
considerable therapeutic potential for the treatment of
Type 2 diabetes and obesity.
Non-selective vanadium based PTPase inhibitors have
been found to increase IR tyrosine phosphorylation
levels in vitro and sensitize and mimic the actions of
insulin in vivo, effectively lowering plasma glucose levels
in diabetic animal models.7 The clinical utility of vana-
dium based antidiabetic agents is limited by their
cumulative toxicity and poor oral absorption and thus a
small molecule equivalent may have considerable ther-
apeutic advantage. Recently, Just and co-workers8
reported a series of periodinate based oxidants with in
vitro PTP1B inhibitory activity. The substituted pyr-
imido[5,4-3][1,2,4]triazine-5,7-diamine 1 was identified
through high throughput screening against LAR. This
compound was subsequently found to have residual
PTP1B activity as well, and served as a starting point
for the development of novel PTPase inhibitors. Herein,
we describe the initial SAR surrounding a novel series
Protein tyrosine phosphatases (PTPases) are enzymes
that regulate signal transduction pathways via dephos-
phorylation of biologically important proteins.3 In the
case of the insulin receptor (IR), PTPases believed to
play a role in its negative regulation include PTP1B,
leukocyte antigen receptor phosphatase (LAR), and
PTPa.4 These enzymes are believed to either directly or
indirectly (via downstream signaling) dephosphorylate
the active form (triphosphorylated in the regulatory
*Corresponding author. Tel.: +1-973-235-3635; fax: +1-973-235-
7122; e-mail: kevin.guertin@roche.com
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00623-1

2896
K. R. Guertin et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2895–2898
of orally active, non-selective ‘vanadate like’ pyri-
mido[5,4-3][1,2,4]triazine-5,7-diamine based-hypoglycemic
agents with PTPase inhibitory activity.
initial screening hit 1 and compound 13 the in vitro
potency is significantly improved (ca. 33-fold) under the
low DTT conditions. The precise role of the diamino-
pyrimidine moiety was not investigated in this study.
The chloromethylpyrimido[5,4-3][1,2,4]triazine-5,7-di-
amine compound 5 served as a convenient synthetic
intermediate for the preparation of PTPase inhibitors 6–
15 and was prepared as outlined in Scheme 1. S-Methyl-
ation of commercially available arylthiol 2 followed by
nitrosylation afforded the nitroso-thiomethyl derivative
3. Nucleophilic displacement of the thiomethyl moiety
with hydrazine provided cyclization precursor 4 in good
yield. Reaction of 4 with chloroacetaldehyde diethyl
acetal under acidic conditions gave, upon dilution of the
reaction mixture with water, compound 5 as a stable
light brown solid. Upon heating in ethanol, inter-
mediate 5 readily underwent reaction with a variety of
secondary amines affording PTPase inhibitors 6–11 as
outlined in Scheme 2. In particular, the piperazine deri-
vative 9 could be further functionalized via alkylation
with a variety of benzylic halides to provide PTPase
inhibitors 12–15 as described in Scheme 3.9
While the enzyme is active at both DTT concentrations,
the in vitro inhibitory activity of the PTPase inhibitors
appears to be sensitive to the concentration of DTT
used in the assay. The sensitivity of the PTP1B inhibi-
tory potency of inhibitors 6–15 to DTT concentration
would suggest a possible redox reaction between the
inhibitors and DTT. It is also possible that the com-
pounds undergo redox chemistry with the PTP1B active
site cysteine residue thus inactivating the enzyme. Sha-
ver and co-workers have speculated that the mode of
PTPase inhibition by peroxovanadium compounds may
involve irreversible oxidation of the catalytic cysteine
Table 1. PTP1B inhibitory activity of compounds 1 and 6–15
Initially, PTPase inhibitors 6–15 were evaluated for
their PTP1B inhibitory activity and the results are out-
lined in Table 1. Compounds were assayed against
PTP1B using a fluorescence based assay10 and the IC50’s
were measured in the presence of 300 nM and 2 mM
dithiothreitol (DTT). Clearly, hydrophobic benzylic
sidechains (e.g., 12–15) are favored on the piperazine
ring with the 4-biphenylmethyl 13 showing the best
PTP1B inhibitory potency (Table 1). In comparing the
Compd
X
PTP1BIC (mM)a
97 (>100)
1
6
29 (>100)
24 (>100)
24 (>100)
29 (91)
7
8
9
10
11
12
13
14
21 (87)
Scheme 1. Reagents and conditions: (a) KOH, MeI, H2O, 71%; (b)
NaNO2, 2 N AcOH, 50 ꢀC, 86%; (c) hydrazine hydrate, DMF, 50 ꢀC,
86%; (d) chloroacetaldehyde diethyl acetal, concd HCl, DMF, rt to
80 ꢀC then NH4OH, 38%.
>100 (>100)
6.2 (98)
2.9 (18)
Scheme 2. Reagents and conditions: (a) secondary amine, EtOH,
80–100 ꢀC.
3.5 (33)
15
4.5 (40)
aIC50 determined in the presence of 300 nM dithiothreitol (DTT). IC50
as measured in the presence of 2 mM DTT in parentheses. All mea-
surements were carried out using a common MES buffer system.
Scheme 3. Reagents and conditions: (a) RI or RBr, K2CO3, DMF, rt.

K. R. Guertin et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2895–2898
2897
residue.11a Under high DTT conditions, one may expect
to observe less inhibition of PTP1B with inhibitors 6–15
as competing redox chemistry with DTT would become
prominent. Indeed, Ramachandran and co-workers11b
have demonstrated that increasing concentrations of
DTT (as well as b-mercaptoethanol and reduced glu-
tathione) effectively decrease the PTP1B inhibitory
activity of pervanadate in vitro and similar inhibition
kinetics were observed for pervanadate with PTP CD45.
The redox hypothesis described herein for PTPase inhi-
bitors 1 and 6–15 would suggest that these compounds
should be rather non-selective in their redox actions, as
all PTPases (cytosolic and transmembrane bound
receptor like enzymes) contain the conserved 11 amino
acid sequence motif (I/V)HcxAGxxR(S/T)G which spe-
cifies the active site of the phosphatase. The conserved
cysteine and arginine residues within this motif are
essential for catalytic activity.15 The PTPase inhibitors
6, 12, and 13 were further evaluated against other
PTPases PTPa, LAR and SHP-2 and the results are
outlined in Table 2. Indeed, the inhibitors show little
selectivity in their inhibitory activity across this panel of
PTPases and are reminiscent of the vanadate based
inhibitors previously reported in the literature.7
It is likely that the triazine ring of inhibitors 6–15 is
responsible for the redox activity of these PTPase inhi-
1
bitors as we have observed using H NMR and MS,
that in the presence of DTT and absence of oxygen, the
structurally related pyrimidotriazine 16 rapidly and
quantitatively undergoes reduction to the corresponding
dihydro derivative 17 which is reversible in the presence
of atmospheric oxygen (Scheme 4).12 A simple starch/
iodide test indicates that ca. 1% of the reduction pro-
duct of oxygen is either hydrogen peroxide or super-
oxide. It is possible that PTPase inhibitors 6–15 catalyse
the reversible oxidation of the active site cysteine of
PTPases to the sulfenic acid via a similar mechanism
involving the generation of peroxide or superoxide. In
an in vivo setting, DTT would be replaced by high levels
of reduced glutathione. Recently, Goldstein et al. have
demonstrated in vivo that insulin stimulation generates
a burst of intracellular hydrogen peroxide that rever-
sibly inhibits PTP1B, thereby enhancing the early insu-
lin cascade.13 Kinetic and reversibility experiments
performed using GST-PTP1B (1–321) construct indi-
cated that compounds of the pyrimido[5,4-3][1,2,4]tri-
azine-5,7-diamine class are reversible and competitive
inhibitors of PTP1B in the presence of low (300 nM)
DTT concentrations.14 Furthermore, the PTP1B inhib-
itory potency of these compounds is completely abolished
when the assay is conducted in the presence of catalase
to sequester any generated hydrogen peroxide. This
result clearly indicates that hydrogen peroxide plays a
central role in the inhibition of PTP1B under the assay
conditions and a redox cycle similar to that outlined
in Scheme 4 is likely to be operative for compounds 1
and 6–15.14 Although generated hydrogen peroxide is
apparently responsible for the PTP1B inhibitory
potency of 1 and 6–15, the substituted piperazine side
chains of 12–15 appear to improve potency somewhat,
presumably by altering the standard redox potential of
the triazine moiety.
The well characterized ob/ob mouse model was chosen
for in vivo evaluation of selected inhibitors.16 The glu-
cose lowering properties of inhibitors 12 and 13 in male
ob/ob mice are summarized in Table 3. Compounds
were dosed po q.d. for 5 days at 50 mg/kg and blood
glucose was measured 2 h post dose on day 5. Inhibitors
12 and 13 display significant glucose lowering effect in
mice with an ED25 (parent drug) around 30 mg/kg po
for 12.17,18 The pharmacokinetic properties of 13 are
summarized in Table 4. The compound is rapidly and
extensively distributed in mice. The achieved systemic
exposure (Cmax) of 13 approximates the measured IC50
Table 2. Selectivity of inhibitors 6, 12, and 13 against relevant
PTPases
Compd
IC50 (mM)a
PTP1B
LAR
PTPa
SHP-2
6
12
13
12
8
8
28
12
16
28
24
15
14
9
5
aIC50’s were determined in a common glutarate buffer system in the
presence of 300 nM DTT.
Table 3. Glucose lowering properties of 12 and 13 in male ob/ob mice
Compd
% change in blood glucosea
Vehicle
12
13
À0.5
À23b
À19b
aCompounds were dosed at 50 mg/kg orally for a 5-day period and
blood glucose was measured 2 h post dose on day 5; values are percent
change relative to pretreatment group with 10 mice per group.
bp<0.05 when compared to vehicle treated mice.
Table 4. Pharmacokinetic (PK) properties of 13
Compd
PK Propertiesa
t1/2
h
CL
(mL/kg/min)
Vss
(L/kg)
Cmax
(mM)
F
(%)
13
1.0
104
3.1
4.0
97
Scheme 4. Catalytic cycle for the reduction of triazine derivatives with
DTT.
aiv and po pharmacokinetic parameters were determined post 10 mg/
kg bolus and 50 mg/kg doses, respectively, in male C57BL/6J mice.

2898
K. R. Guertin et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2895–2898
against PTP1B in vitro under the low DTT conditions
(Table 1). The compound has a favorable t1/2 and
excellent oral bioavailability. The large steady state
volume of distribution (Vss) of 13 is approximately 6
times the total body water volume of the animal sug-
gesting deeptissue and cell penetration. On the other
hand, the compound suffers from high systemic clear-
ance which exceeds hepatic blood flow of the mouse.
8. Leung, K. W. K.; Posner, B. I.; Just, G. Bioorg. Med.
Chem. Lett. 1999, 9, 353.
9. Satisfactory spectral data was obtained for all compounds.
Final products 6–15 were purified by C18 reversed-phase
HPLC, eluting with 0.1% TFA, H2O/CH3CN gradient, lyo-
philized and isolated as their trifluoroacetate salts.
10. Assay conditions: PTP1B (0.5–2 nM) was incubated with
compound for 15 min in buffer containing 37.5 mM MES pH
6.2, 140 mM NaCl, 2 mM EDTA, 0.05% BSA and either 300
nM or 2 mM DTT. The reaction was initiated by the addition
of 50 mM substrate [phosphorylated peptide 1146(TRDI-
(pY)E]. After 20 min at room temperature (22–25 ꢀC), the
reaction was stopped with KOH and the amount of free
phosphate measured using malachite green as previously
described (Harder, K. W.; Owen, P.; Wang, L. K.; Asbersold,
R.; Clark-Lewis, I.; Jirik, F. R. Biochem J. 1994, 298, 395). In
order to measure the PTPase activity across a panel of phos-
phatases, similar assay conditions were used as described
except 6,8-difluoro-4-methylumbelliferylphosphate (DiFMUP,
Molecular Probes) was used as substrate at the Km for each
enzyme and 37.5 mM diethylglutarate pH 6.2 was used in
place of MES. The reaction was stopped with KOH and the
fluorescence of the dephosphorylated substrate measured
(excitation: 360 nM/emission: 460 nM).
In conclusion,
a novel series of pyrimido[5,4-3]-
[1,2,4]triazine-5,7-diamine-based hypoglycemic agents
have been discovered. These compounds display non-
selective or ‘vanadate like’ PTPase inhibitory properties
in vitro and 12 and 13 were found to effectively lower
blood glucose levels in ob/ob mice upon oral adminis-
tration. From a toxicological standpoint, the lack of a
metallic component suggests that compounds of this
class may offer an advantage over the vanadium-based
anti-diabetic agents currently in clinical development.
Acknowledgements
11. (a) Shaver, A.; Ng, J. B.; Hall, D. A.; Soo Lum, B.; Posner,
B. I. Inorg. Chem. 1993, 32, 3109. (b) Huyer, G.; Liu, S.; Kelly,
J.; Moffat, J.; Payette, P.; Kennedy, B.; Tsaprailis, G.;
Gresser, M. J.; Ramachandran, C. J. Biol. Chem. 1997, 272,
843.
We thank Drs. Jeff Tilley and Nader Fotouhi for helpful
discussions and Dr. Hans Kuehnle for in vivo eval-
uation of compounds 12 and 13.
12. Upon reaction of 16 with DTT, the CH proton of the tri-
azine ring of product 17 is shifted upfield by approximately 3
ppm compared to 16, indicating a loss in aromaticity of this
ring. In DMSO-d6, two NH protons could be detected at 6.93
and 8.73 ppm. Mass spectrometry revealed a species whose
mass is 2 units higher than parent 16, corresponding to the
addition of hydrogen and consistent with the structure 17.
13. Mahadev, K.; Zilbering, A.; Zhu, L.; Goldstein, B. J . J .
Biol. Chem. 2001, 276, 21938.
14. Olivier, A. R. Unpublished results. A detailed account of
this work will be published separately.
15. (a) Fischer, E. H.; Charbonneau, H.; Tonks, N. K.
Science 1991, 253, 401. (b) Pot, D. A.; Dixon, J. E. Biochim.
Biophys. Acta 1992, 1136, 35.
References and Notes
1. Defronzo, R. A. Diabet. Rev. 1997, 5, 177.
2. Klein, R. Diabetes Care 1995, 18, 258.
3. Neel, B. G.; Tonks, N. K. Curr. Opin. Cell Biol. 1997, 9, 93.
4. (a) Goldstein, B. J. J. Cell. Biochem. 1992, 31, 33. (b)
Goldstein, B. J. Receptor 1993, 3, 1. (c) Ahmad, F.; Goldstein,
B. J. Biochim. Biophys. Acta 1995, 1248, 57.
5. Byon, J. C. H.; Kusari, A. B.; Kusari, J. Mol. Cell. Bio-
chem. 1998, 182, 101.
6. 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.; Kennedy, B. P. Science 1999, 283, 1544.
7. (a) Huyer, G.; Liu, S.; Kelly, J. J. Biol. Chem. 1997, 272,
843. (b) Tsiani, E.; Fantus, G. Trends Endocrinol. Metab.
1997, 8, 51. (c) Posner, B. I.; Faure, R.; Burgess, J. W.; Bevan,
A. P.; Lachance, D.; Zhang-Sun, G.; Fantus, I. G.; Ng, J. B.;
Hall, D. A.; Lum, B. S. J. Biol. Chem. 1994, 269, 4596. (d)
Yale, J. F.; Lachance, D.; Bevan, A. P.; Vigeant, C.; Shaver,
A.; Posner, B. I. Diabetes 1995, 44, 1274. (e) Halberstan, M.;
Cohen, N.; Shlimovich, P.; Rosetti, L.; Shamoon, H. Diabetes
1996, 45, 659. (f) Fantus, I. G.; Tsiani, E. Mol. Cell. Biochem.
1998, 182, 109.
16. Coleman, D. L. Diabetologia 1978, 14, 141.
17. As determined by C, H, N, O, F combustion analysis, the
trifluoroacetate salt of compounds 12 and 13 contains 3 mol of
TFA per mol of parent 12 and 13. Thus a 50-mg/kg dose of 12
corresponds to 28 mg/kg of parent drug.
18. Preliminary toxicological data was obtained following the
5 day dosing study in ob/ob mice (Table 3). No significant
adverse treatment related effects were observed on organ
weights, hematological parameters or serum chemistries. A
histopathological evaluation of selected tissues was per-
formed and no treatment related microscopic changes were
observed.