Lead generation and optimization of novel GPR119 agonists with a spirocyclic cyclohexane structure

Article abstract of DOI:10.1016/j.bmcl.2018.12.041

We describe here the generation of a lead compound and its optimization studies that led to the identification of a novel GPR119 agonist. Based on a spirocyclic cyclohexane structure reported in our previous work, we identified compound 8 as a lead compound, being guided by ligand-lipophilicity efficiency (LLE), which linked potency and lipophilicity. Subsequent optimization studies of 8 for improvement of solubility afforded representative 21. Compound 21 had no inhibitory activity against six CYP isoforms and showed favorable pharmacokinetic properties and hypoglycemic activity in rats.

Full text of DOI:10.1016/j.bmcl.2018.12.041

Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Lead generation and optimization of novel GPR119 agonists with a
spirocyclic cyclohexane structure
Kazuhito Harada⁎, Jun Mizukami, Takashi Watanabe, Genki Mori, Minoru Ubukata,
Katsunori Suwa, Sumiaki Fukuda, Tamotsu Negoro, Motohide Sato, Takashi Inaba
Central Pharmaceutical Research Institute, Japan Tobacco Inc., 1-1, Murasaki-cho, Takatsuki, Osaka 569-1125, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords:
We describe here the generation of a lead compound and its optimization studies that led to the identification of
a novel GPR119 agonist. Based on a spirocyclic cyclohexane structure reported in our previous work, we
identified compound 8 as a lead compound, being guided by ligand-lipophilicity efficiency (LLE), which linked
potency and lipophilicity. Subsequent optimization studies of 8 for improvement of solubility afforded re-
presentative 21. Compound 21 had no inhibitory activity against six CYP isoforms and showed favorable
pharmacokinetic properties and hypoglycemic activity in rats.
GPR119 agonist
Spirocyclic structure
GPR119 is a G protein-coupled receptor (GPCR) expressed pre-
dominantly in pancreatic beta cells and enteroendocrine cells.
Activation of GPR119 promotes glucose-dependent insulin release in
the pancreas and secretion of incretins such as glucagon-like peptide-1
(GLP-1) in the gut. GPR119 has the potential to be a suitable target for
the development of orally active, small molecule agonists and has re-
ceived attention from many pharmaceutical companies in recent years.1
A number of synthetic GPR119 agonists have been reported to date,
several of which have advanced into clinical trials.2 Reported GPR119
agonists are often composed of three parts as depicted in Fig. 1: a pi-
peridine or a piperazine ring N-substituted with a carbamate or a het-
eroaryl group at the right side, a phenyl group substituted with a me-
thylsulfonyl or a heteroaryl group (X) at the left side, and a linker part
connecting the two parts.2,3 Of the three parts, the N-substituted pi-
peridine or piperazine ring as the right side moiety is a characteristic
structure.
for the ocular toxicity.
We therefore undertook the goal to discover a novel GPR119 agonist
with reduced lipophilicity, as compared to 5. We set out to generate a
new lead compound having a good balance of agonistic activity and
lipophilicity by using ligand-lipophilicity efficiency (LLE)6 as an index.
In seeking a new lead compound, we expected that the spirocyclic cy-
clohexane structure of 5 would be a versatile right side moiety in
GPR119 agonists. Thus, this structure was fixed as the right side moiety
and combined with linkers and left side moieties found in known
GPR119 agonists. Among them, we were interested in 1 (MBX-2982), 2
(PSN119-2), and 3 (Arena’s compound) (Fig. 1) with linkers and left
side moieties which could reduce lipophilicity.2,3 Each of the com-
pounds having the spirocyclic cyclohexane structure as the right side
moiety were synthesized and evaluated (Table 1).
All of the spirocyclic cyclohexane intermediates 26, 28, 29 and 32
were synthesized from commercially available compounds 22 and 30 as
shown in Scheme 1. Lithiation of ester 23 followed by reaction with
isobutylene oxide gave the desired lactone 24 with high diastereos-
electivity (98:2 d.r.).4 Reduction of 24 with LiAlH4 and subsequent
treatment with p-toluenesulfonic acid monohydrate furnished spiro
ether 25. Oxidation of 25 provided spirocyclic carboxylic acid 26.
Curtius rearrangement of 26 in the presence of potassium tert-butoxide
gave the tert-butoxy carbonyl (Boc)-protected amine 27 while main-
taining high diastereomeric ratio (98:2 d.r.). Amine 28 was obtained as
the corresponding HCl salt by deprotection of the Boc group of 27.
Reduction of 27 with LiAlH4 under reflux and subsequent treatment
with HCl in AcOEt gave N-methylated amine 29 as the corresponding
In our previous work, a novel spirocyclic cyclohexane structure was
identified as an alternative to the right side moiety of compound 4
(Fig. 2).4 The representative
5 having high three-dimensionality
showed potent GPR119 agonistic activity with no cytochrome P450
(CYP) inhibitory activity. In the rat intraperitoneal glucose tolerance
test (ipGTT), 5 also displayed hypoglycemic activity with insulin se-
cretion dependent on glucose concentration.
However, corneal toxicity was observed in the rat toxicity test of 5.
In general, it is known that compounds with high lipophilicity may
show nonspecific off-target action and have high risk of toxicity.5 We
speculated that high lipophilicity of 5 (Clog P = 5.1) was one of reasons
⁎ Corresponding author.
E-mail address: kazuhito.harada@jt.com (K. Harada).
https://doi.org/10.1016/j.bmcl.2018.12.041
Received 18 September 2018; Received in revised form 1 December 2018; Accepted 17 December 2018
0960-894X/©2018ElsevierLtd.Allrightsreserved.
Pleasecitethisarticleas:Harada,K.,Bioorganic&MedicinalChemistryLetters,https://doi.org/10.1016/j.bmcl.2018.12.041

K. Harada et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Fig. 1. Representatives and typical structure of GPR119 agonists.
Fig. 2. Our previous work.
Table 1
Lead generation guided by LLE.
Compound
Left side-Linker
GPR119 EC50 (nM)
4
IA (%)
104
CLogPa
5.1
LLEb
3.3
5
6
7
85
89
79
3.7
3.4
3.3
3.6
109
8
13
91
4.0
4.0
a
b
The Clog P value was calculated using a software from ChemAxon.
LLE = pEC50 − Clog P.
2

K. Harada et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Scheme 1. Synthesis of spirocyclic cyclohexane intermediates 26, 28, 29 and 32. Reagents and conditions: (a) MeI, K2CO3, DMF, rt; (b) DHP, (−)-CSA, toluene, rt;
(c) isobutylene oxide, LiHMDS, THF, rt, 62% (3 steps); (d) LiAlH4, THF, 0 °C to rt, 79%; (e) p-TsOH·H2O, MeOH, 85 °C, 87%; (f) Dess-Martin reagent, CHCl3, 0 °C; (g)
NaClO2, NaH2PO4, 2-methyl-2-butene, t-BuOH, H2O, rt, 54% (2 steps); (h) DPPA, NEt3, KO-t-Bu, dioxane, 90 °C, 97%; (i) HCl, AcOEt, rt, 66%; (j) LiAlH4, THF, reflux;
(k) HCl, AcOEt, rt, 47% (2 steps); (l) DHP, (−)-CSA, toluene, rt; (m) isobutylene oxide, LiHMDS, THF, rt, 69% (2 steps); (n) LiAlH4, THF, 0 °C to rt, 89%; (o) p-
TsOH·H2O, MeOH, rt, 91%.
HCl salt. Spirocyclic alcohol 32 was synthesized with high diastereos-
electivity (97:3 d.r.) by the same procedure for 25.
21, respectively. Oxidative cleavage of 52, reductive amination with
aniline 40 and treatment under acidic conditions provided the final
product 19.
Compounds 6 and 7 were obtained from 26 and 32, respectively
(Scheme 2). Amide 33 was converted from 26. Reaction of 33 with
Lawesson’s reagent followed by reaction of ethyl bromopyruvate gave
thiazole ester 34. Reduction of 34 with LiAlH4 and subsequent coupling
with the corresponding phenol provided 6. Reaction of alcohol 32 with
2-bromopropanoic acid followed by reaction with hydroxyl amidine 36
and oxidation of the sulfide group afforded 7.
GPR119 agonistic activity was evaluated for the human GPR119
receptor in a cell-based cAMP assay.7 The results expressed the potency
as EC50 values and the inherent activity (IA) as percentages which were
compared to the test compound with oleoylethanolamide (OEA), an
endogenous ligand of GPR119 (defined as 100% activation).
As expected, compounds 6–8 all showed lower values of lipophili-
city than 5. In addition, they exhibited acceptable GPR119 agonist
activity, which indicated that the spirocyclic cyclohexane structure of 5
was a versatile right side moiety in GPR119 agonists. Among 6–8,
Compounds 8–11, 16 and 17 were synthesized from 32 (Scheme 3).
Reaction of 32 with the corresponding dichloropyrimidine and sub-
sequent reaction with the corresponding aniline or phenol afforded the
final products.
pyrimidine derivative
8 expressed the highest LLE value (4.0).
Compounds 12–15 and 20 were synthesized by starting from the
corresponding anilines (Scheme 4). The intermediates 45, 46, 48 and
49 were coupled with 32 to afford the final products 12–15. The re-
action of 29 with 49 gave the final product 20.
Therefore, we selected 8 as a new lead compound having a better
balance of agonistic activity and lipophilicity.
Next, we moved to an optimization process, where we focused on
improvement for values of solubility of 8 (Table 2). Compound 8
showed considerably lower values of solubility for Japanese Pharma-
copoeia 1st fluid for a dissolution test adjusted to pH 1.2 (JP1) and Fed-
state simulated intestinal fluid (FeSSIF). In general, solubility is one
factor that governs oral absorption, and compounds with low solubility
Compounds 18, 19 and 21 were converted from 28 (Scheme 5).
Reaction of 28 with the corresponding dichloropyrimidine and sub-
sequent methylation of the NH group gave intermediates 50–52. Re-
action of 50 and 51 with aniline 40 afforded the final products 18 and
Scheme 2. Reagents and conditions: (a) NH4Cl, NEt3,
EDC, HOBt, CH3CN, rt, 96%; (b) Lawesson’s reagent,
THF, reflux; (c) ethyl bromopyruvate, EtOH, rt, 32%
(2 steps); (d) LiAlH4, THF, 0 °C to rt, 94%; (e) 4-(1H-
tetrazol-1-yl)phenol, PPh3, DMAD, toluene, rt, 76%;
(f) 2-bromopropanoic acid, NaH, dioxane, rt, 62%;
(g) EDC, HOBt, i-Pr2NEt, DMF, rt; (h) toluene, reflux;
(i) m-CPBA, CHCl3, 0 °C, 55% (3 steps).
3

K. Harada et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Scheme 3. Reagents and conditions: (a) the corresponding 4,6-dichloropyrimidine, NaH, THF, rt, 83–93%; (b) 40, Pd(OAc)2, ( )-BINAP, NaO-t-Bu, dioxane,
100 °C, 12%; (c) the corresponding aniline 40–42, Pd(OAc)2, 1,1′-bis(di-t-butylphosphino)ferrocene, NaO-t-Bu, dioxane, 100 °C, 26–57%; (d) 43, K2CO3, TBAI,
DMSO, 130 °C, 67%.
Scheme 5. Reagents and conditions: (a) the corresponding 4,6-di-
chlorolpyrimidine, K2CO3, DMSO, rt; (b) MeI, NaH, DMF, rt, 25–70% (2 steps);
(c) 40, PdCl2(dppf), ( )-BINAP, NaO-t-Bu, DMF, 80 °C, 22–38%; (d) 4,6-di-
chloro-5-allylpyrimidine, i-Pr2NEt, K2CO3, DMF, 80 °C; (e) MeI, NaH, DMF,
60 °C, 26% (2 steps); (f) NaIO4, KOsO4, acetone, H2O, rt, 10%; (g) 40, NaBH
(OAc)3, TFA, CHCl3, rt; (h) conc. HCl, n-PrOH, 100 °C, 66% (2 steps).
Scheme 4. Reagents and conditions: (a) 4,6-dichloro-5-methylpyrimidine, Pd
(OAc)2, NaO-t-Bu, 1,1′-bis(di-t-butylphosphino)ferrocene, dioxane, 120 °C,
16%; (b) 32, NaH, THF, reflux, 22–69%; (c) 2-(4,6-dichloropyrimidin-5-yl)
acetaldehyde, NaBH(OAc)3, THF, rt, 85%; (d) the corresponding 4,6-di-
chlorolpyrimidine, n-PrOH, reflux, 39%; (e) 29, Pd(OAc)2, DavePhos, K2CO3,
dioxane, 105 °C, 5%.
transformed by the dimethylcarbamoyl group showed higher values of
solubility, particularly against JP1. However, agonistic activity of 9 was
decreased about three times compared with 8.
Next, we focused on molecular planarity, which is known to influ-
ence crystal packing. Disruption of molecular planarity would be ex-
pected to decrease the efficiency of crystal packing, leading to im-
proved solubility.9 In the case of 8, molecular planarity of the two
aromatic rings was considered as the cause of low solubility.
Therefore, we investigated how solubility and agonistic activity
are related to low oral absorption.
In our previous work4 and in other reports,8 conversion of the me-
thylsulfonyl group at the left side to the dimethylcarbamoyl compound
exhibited considerable improvement in solubility. Compound
9
4

K. Harada et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 2
SAR and physicochemical properties of pyrimidine derivatives.
Compound
Left side-Linker
GPR119 EC50 (nM)
13
IA (%)
91
Clog Pa
4.0
LLEb
4.0
JP1c (μM)
5.3
FeSSIF (μM)
15.8
8
9
33
59
101
62
4.5
4.1
3.7
3.2
128.5
< 0.5
65.2
33.4
10
11
12
139
92
17
3.9
4.2
3.0
–
4.4
8.6
> 1000
13.0
42.5
13
5
90
3.7
4.6
18.9
15.2
14
15
16
> 1000
15
38
87
82
4.1
3.6
4.4
–
2.8
7.0
4.6
25.5
19.8
67.5
4.2
3.0
42
17
18
98
45
61
4.7
4.3
2.3
3.1
0.7
4.3
128
> 475
49.8
a
b
c
The Clog P value was calculated using a software from ChemAxon.
LLE = pEC50 − Clog P.
Japanese Pharmacopoeia 1st fluid for a dissolution test adjusted to pH 1.2.
5

K. Harada et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 3
Fine tuning of pyrimidine derivatives.
Compound
Left side-Linker
GPR119 EC50 (nM)
45
IA (%)
128
CLogPa
4.3
LLEb
3.1
hMsc (%)
91
rMsd (%)
93
JP1e/FeSSIF (μM)
> 475 / 49.8
18
19
39
118
4.0
3.4
99
94
> 475 / 75.3
20
21
16
19
108
120
3.9
3.7
3.9
4.0
89
80
96
418.2 / 24.3
473.3 / 31.1
103
a
b
c
d
e
The Clog P value was calculated using a software from ChemAxon.
LLE = pEC50 − Clog P.
Percent of compound (5 μM) remaining after 1 h incubation with human liver microsomes (0.2 μg/mL).
Percent of compound (5 μM) remaining after 1 h incubation with rat liver microsomes (0.2 μg/mL).
Japanese Pharmacopoeia 1st fluid for a dissolution test adjusted to pH 1.2.
were affected by disrupting molecular planarity. Compound 10 with the
O atom instead of the NH group connecting the two aromatic rings in 8
showed little improvement in solubility and its agonistic activity de-
creased four times. In addition, the F atom on the benzene ring had
little effect in solubility, and the agonistic activity of compound 11
without the F atom decreased about ten times compared with 8. On the
other hand, N-methylation of 8 slightly improved solubility and the
agonistic activity of N-methylated 12 was greatly weakened. These
results suggested that the intramolecular hydrogen bond between the F
atom and the H atom of the NH group in 8 might be important to
achieve highly agonistic activity.10 This outcome was also indicated
that steric repulsion between the N-methyl group and the methyl group
at the 5-position on the pyrimidine ring in 12 significantly changed the
molecular arrangement and led to decreased agonistic activity.
In order to eliminate the steric repulsion, cyclization of the two
methyl groups in 12 was investigated.11 As anticipated, this cyclization
boosted the agonistic activity significantly, and the thus-obtained 13
showed an EC50 value of 4 nM and LLE value of 4.6. The similar im-
provement in agonistic activity was observed in the indoline deriva-
tive10,12 15 which maintained coplanarity compared with compound
14. These results suggested that maintaining of coplanarity in the two
aromatic rings was important to demonstrate potent agonistic activity.
Unfortunately, cyclized compounds 13 and 15 exhibited little im-
provement in solubility against only acidic medium (JP1) as compared
with 8.
for the two media, especially for acidic medium (JP1). The agonistic
activity of 18 was only about three times less than 8.
Optimization of 8 led to compounds 9 and 18 which exhibited good
solubility and acceptable agonistic activity. Metabolic stability of 9 was
moderate (percentage remaining at 1 h = 77% in human liver micro-
somes, 44% in rat liver microsomes), whereas 18 exhibited high me-
tabolic stability as shown in Table 3.
Further optimization of 18 was carried out with the aim of enhan-
cing agonistic activity, while maintaining good solubility and metabolic
stability (Table 3). Cyclized products 19 and 20 with favorable profiles
in solubility and metabolic stability, especially the indoline derivative
20, showed about three times improvement in agonistic activity com-
pared with 18. This result suggested that steric repulsion between the
substituent at the 5-position of the pyrimidine and the N-methyl group
attenuated the agonistic activity of 18 and 19, when compared with the
ether forms 8 and 13, respectively. Therefore, compound 21 without
the methyl group at the 5-position of pyrimidine in 18 was synthesized
and evaluated. As
a result, improvement of agonistic activity
(EC50 = 19 nM) was observed while maintaining good solubility and
metabolic stability. The lipophilicity of 21 (Clog P = 3.6) was con-
siderably lower than that of 5 (Clog P = 5.1), resulting in a better LLE
value (4.0).
We finally succeeded in identifying a novel GPR119 agonist 21
which exhibited reduced lipophilicity compared with
5 and ad-
ditionally demonstrated good solubility and metabolic stability.
The pharmacokinetic profiles of compound 21 were evaluated in
Sprague-Dawley rats (Table 4). Compound 21 exhibited long-lasting
and good oral availability. In addition, 21 showed no inhibitory activity
(IC50 > 50 μM) against all six CYP isoforms (CYP 3A4, 2D6, 1A2, 2C9,
Another approach to decrease the efficiency of crystal packing while
maintaining the coplanar arrangement was investigated. The methyl
group at the 5-position of the pyrimidine ring was converted to either
an ethyl group or an isopropyl group. Although ethyl derivative 16
showed slight improvement in solubility against FeSSIF, the agonistic
activity diminished about three times compared with 8. In isopropyl
derivative 17, no improvement in solubility was observed and further
agonistic activity attenuation occurred.
Table 4
Phamacokinetic profiles of 21 in Sprague-Dawley rats.
Rat IV (1 mg/kg)
Rat PO (3 mg/kg)
Finally, we investigated increasing basicity as a way to improve
solubility. We expected that conversion of the ether linkage of 8 to an
N-methyl group could improve solubility. More than expected, com-
pound 18 demonstrated significant improvement of values of solubility
t1/2β (h)
3.7
CL (L/h/kg)
0.8
Vdss (L/kg)
4.0
MRT (h)
5.1
BA (%)
99
6

K. Harada et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
2C8, 2C19).
References
The effect of compound 21 on plasma glucose levels in the in-
traperitoneal glucose tolerance test (ipGTT) in Sprague-Dawley rats was
evaluated.13 Glucose solution was intraperitoneally administered at
16 h after oral administration of 21 (10 mg/kg) or vehicle. The plasma
glucose levels were measured at 0, 30 and 60 min after glucose loading.
Plasma glucose levels just before glucose loading were slightly different
between the vehicle and 21 groups. The plasma glucose levels at 30 and
60 min after glucose loading in 21 group were 14% and 11% lower than
that in the vehicle group, respectively. Compound 21 suppressed the
increase in peak plasma glucose levels after glucose loading.
In summary, lead generation studies guided by LLE led to the
identification of compound 8 having a spirocyclic cyclohexane struc-
ture, which was a versatile right side moiety in GPR119 agonists.
Further efforts to improve solubility succeeded in discovery of a novel
GPR119 agonist 21. Compound 21 exhibited orally active, long-lasting
pharmacokinetic profiles and showed a hypoglycemic effect after 16 h
of administration in rats. Further identification of novel GPR119 ago-
nists with a spirocyclic cyclohexane structure will be reported in due
course.
1. (a) Chu Z-L, Jones RM, He H, et al. Endocrinology. 2007;148:2601;
(b) Chu Z-L, Carroll C, Alfonso J, et al. Endocrinology. 2008;149:2038.
2. Ritter K, Buning C, Halland N, Pöverlein C, Schwink L. J Med Chem. 2016;59:3579.
3. Semple G, Lehmann J, Wong A, et al. Bioorg Med Chem Lett. 2012;22:1750.
4. Harada K, Mizukami J, Kadowaki S, et al. Bioorg Med Chem Lett. 2018;28:1228.
5. Waring MJ. Expert Opin Drug Discov. 2010;5:235.
6. Edwards MP, Price DA. Annu Rep Med Chem. 2010;45:381.
7.. GPR119 agonistic activity was evaluated in HEK293 cells overexpressing human
GPR119 using a cAMP HiRange assay kit (Cisbio). The test compounds in stimulation
buffer (Hanks’ Balanced Salt Solution (Invitrogen), 20 mM HEPES, 0.1% bovine serum
albumin, 0.1 mM 3-isobutyl-1-methylxanthine, and 0.1 mM 4-(3-butoxy-4-methox-
ybenzyl)-2-imidazolidinone) were plated in 96-well half-area plates. The cells were plated
onto the plates at 12500 cells/well in buffer mix (stimulation buffer with cAMP-d2, ac-
cording to the manufacturer’s protocol) then incubated for 30 min at room temperature.
After that, anti-cAMP cryptate conjugate in lysis buffer was added into the wells and
incubated for 3 h at room temperature in the dark. Cellular cAMP levels were measured
using a micro plate reader. The EC50 value of each compound was determined as the
concentration of the test compound required to achieve 50% of the maximal oleoy-
lethanolamide (OEA) stimulated response.
8. (a) Matsuda D, Kobashi Y, Mikami A, et al. Bioorg Med Chem Lett. 2016;26:3441;
(b) Matsuda D, Kobashi Y, Mikami A, et al. Bioorg Med Chem. 2017;25:4339.
9. Ishikawa M, Hashimoto Y. J Med Chem. 2011;54:1539.
10. Koshizawa T, Morimoto T, Watanabe G, et al. Bioorg Med Chem Lett. 2017;27:3249.
11. Katamreddy SR, Carpenter AJ, Ammala CE, et al. J Med Chem. 2012;55:10972.
12. Several indoline derivatives in GPR119 agonists have been reported. (a) Kaneko T,
Shida T, Baba T, Matsumoto K, Aoki K. WO 2009051119; 2009; (b) Darout E,
Deninno MP, Futatsugi K, et al. WO 2010128414; 2010; (c) Rikimaru K, Imoto H,
Kamaura M. WO 2010095663; 2010; (d) Tran JA, Chen C, WO 2010149685; 2010.
(e) Sato K, Sugimoto H, Rikimaru K, et al. Bioorg Med Chem 2014, 22, 1649.
13.. We conducted ipGTT in place of oral glucose tolerance test (oGTT) to evaluate the
inherent glucose lowering effect of the compound without nonspecific effects, such as
gastric emptying inhibitory action. The ipGTT was performed in the rats (7 weeks of age)
after overnight fasting. Compounds were administrated by oral gavage. After 16 h, glu-
cose (1 g/kg of body weight) was administrated intraperitoneally to rats and blood
samples were collected from the tail vein prior to and at 10, 30 and 60 min after glucose
loading. Plasma glucose levels were measured using an automatic analyzer
(Hitachi7170S, Tokyo, Japan). Plasma insulin levels were measured with a rat-insulin
ELISA kit (Morinaga Institute of Biological Science, Yokohama, Japan).
Acknowledgments
We thank Dr. Hisashi Shinkai for helpful discussions. We also ap-
preciate Dr. Hiromasa Hashimoto for suggestions on writing this
manuscript.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2018.12.041.
7