B. D. Stevens et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6588–6592
6589
plays a key role in the liver where it closely regulates hepatic glu-
cose uptake, glycogen synthesis, and gluconeogenesis.
O
S
4
O
NH
2
R1
3
R2
N
a, b
Given the overwhelming body of supporting human genetic
data and the repeated demonstration of glucose lowering efficacy
with small molecule allosteric activators, it is notable that GK re-
mains as of yet an undrugged target.27 Induction of clinical hypo-
glycemia remains a key hurdle that has led to attrition of
numerous GK activators.28 As such, several groups have proposed
novel strategies for enhancing the therapeutic index of these com-
pounds. One such strategy has involved the preferential activation
of GK within the liver, which can be accomplished by chemically
targeting compounds for uptake via OATP transporters that are ex-
pressed constitutively in human hepatocytes.29,30 Simultaneously,
the passive permeability of these compounds must be minimized
in order to limit exposures within peripheral tissues, thereby effec-
tively eliminating the potentiation of GSIS mediated via activating
GK in the pancreas.
For an orally (PO) dosed therapeutic relying on a liver targeting
strategy, a key balance between oral bioavailability versus periph-
eral exposure must be achieved. We recently demonstrated this
approach with the discovery of PF-04991532 (1, Fig. 1) as an orally
administered liver selective glucokinase activator, which advanced
to Phase 2 clinical studies for the treatment of T2DM.29,30 In Wistar
rats, PF-04991532 was shown to have enhanced liver to plasma
(8.8:1) and impaired pancreas to plasma (0.12:1) distribution
affording a substantial differential between liver and pancreas
exposure (75:1). This candidate was efficacious in various preclin-
ical models of diabetes and was recently shown to be efficacious in
T2DM patients with low hypoglycemia risk.30
H
O
CO2Me
TfO
CO2Me
c, d
R1
R1
N
H
O
S
6-11
CO2H
O
e, f
N
N
O
O
N
CO2H
S
R2
O
R2
N
5
N
Scheme 1. General synthesis of sulfonyl imidazole derivatives. Reagents and
conditions: (a) R1-MgBr, Li2CuCl2, Et2O, THF, À50 ? À78 °C; (b) Tf2O, 2,6-lutidine,
heptane, CH2Cl2, À10 °C; (c) K2CO3, EtOAc, 23 °C; (d) aq 6 N HCl, 100 °C; (e) benzyl
i
6-aminonicotinate, T3P, 2,6-lutidine, EtOAc, 10 ? 23 °C; (f) H2, Pd/C, PrOH.
for glucose was defined as
ing the enzyme’s Vmax was defined as b. Values of
1 with lower values of representing more substantial reductions
a
and the maximum fold effect on alter-
a
ranged from 0–
a
Km and increases in the enzyme’s glucose affinity for glucose. Val-
ues of b >1 represented activator induced increases in the enzyme’s
Vmax. An activator’s potency or EC50 was defined as the concentra-
tion affording a half maximal reduction in Km.
From a structure activity perspective, previous work in the
hepatoselective cycloalkyl propionate GKAs revealed a tolerance
of various R groups in the 4-imidazolyl position. Although lipo-
philic groups generally led to optimal potency, the pancreatic per-
meability of these compounds could be reduced by two to
threefold by introduction of a polar sulfonyl group (Table 1). Tar-
gets 6 and 7 containing a cyclopentyl group at R1 exhibited compa-
rable potency to the corresponding cyclohexyl derivatives 8 and 9,
but led to reduced maximal enzyme activation (b). In addition, the
aqueous kinetic solubility of cyclohexyl propionates proved infe-
In this study, sought to evaluate whether this liver targeting ap-
proach could be further extended to the targeting of intravenously
administered therapeutics. In particular, because intravenously
administered therapeutics do not rely on oral absorption, we
sought to explore whether even greater hepatic targeting might
be realized by further restricting passive permeability to an extent
not compatible with oral delivery. Herein we describe the identifi-
cation and optimization of an IV deliverable hepatoselective gluco-
kinase activator for the potential treatment of in-patient
hyperglycemia.
rior (e.g., 7 = 428 lM vs 9 = 198 lM at pH 6.5), thus rendering them
less attractive for IV use. Interestingly, attempts to increase polar
surface area and further reduce passive permeability through
incorporation of a pyranyl moiety (i.e., 10 and 11) resulted in sub-
stantial loss of activation potency. Both cyclopropyl and cyclobutyl
substituents were effective at R2, with cyclopropyl possessing a
small advantage in potency, a, and b values.
The synthesis of sulfonyl imidazole analogs of 1 generally fol-
lowed previously reported routes (Scheme 1).29 Various alkyl Grig-
nard reagents were reacted with commercially available methyl
(2R)-glycidate 2 to yield the corresponding free alcohols which
were readily converted into the intermediate triflates 3. Nucleo-
philic inversion of the chiral triflates using sulfonyl imidazoles 4
followed by acidic hydrolysis of the methyl ester afforded
carboxylic acid 5. Subsequent amidation of the free acids with
O-benzyl-6-aminonicotinate and hydrogenolysis of the protecting
group followed by chiral purification gave the final products
6–11 (Table 1).aBiochemical assay values reported as the geometric
mean of n >2 independent determinations.
Of the profiled compounds, 6 and 7 were selected for further
study based on their balance of physical properties and enzyme
activation profile. Previous work has highlighted the ratio of func-
tional effect in INS-1 cell lines vs EC50 in the biochemical glucoki-
nase enzyme assay as an effective gauge of pancreatic impairment;
hence, the dose response effect of activators 6 and 7 on glucose-
stimulated insulin secretion in INS-1 cells cultured in 5 mM glu-
cose.29 In this system, activator 6 exhibited an EC50 = 26
compound 7 had EC50 = 174 M highlighting the functional impair-
ment of these activators in a pancreatic cell line. By comparison
reference compound 1 had EC50 = 6.9 M in the INS-1 assay. Given
lM and
l
l
the enhanced functional impairment for compound 7, relative to 6,
in the INS-1 assay, this particular compound was selected for fur-
ther profiling. We next sought to evaluate the effect of 7 in primary
rat hepatocyctes (expressing organic anion transporters) in order
to compare its activity in these cells relative to the INS-1 cells. As
previously described, glucokinase activity in hepatocytes is regu-
lated though an interaction with glucokinase regulatory protein
(GKRP) which, during conditions of low glucose, binds the inactive
conformation of glucokinase and sequesters the enzyme to the
nucleus.29 As glucose concentrations increase, glucokinase dissoci-
ates from GKRP and enters the cytoplasm. Glucokinase activators
have been previously shown to disrupt this glucokinase-GKRP
interaction; thus to characterize the effects of compound 7 on
the glucokinase-GKRP interaction, a dose response evaluation
was conducted in freshly isolated Wistar rat hepatocytes at
The activity of activators prepared in these studies was evalu-
ated in a biochemical assay using human recombinant glucokinase
as previously described.31 In this assay, determination of an
activator’s maximum fold effect on reducing the glucokinase Km
H
N
N
N
F3C
O
N
CO2H
1
Figure 1. Structure of Hepatoselective GKA PF-04991532 (1).
Stevens, Benjamin D.
