Phenotypic screening to identify small-molecule enhancers for glucose uptake: Target identification and rational optimization of their efficacy

Article abstract of DOI:10.1002/anie.201310618

Small-molecule glucose uptake enhancers targeted to myotubes and adipocytes were developed through a phenotypic screening linked with target identification and rational optimization. The target protein of glucose-uptake enhancers was identified as a nucle

Full text of DOI:10.1002/anie.201310618

.
Angewandte
Communications
DOI: 10.1002/anie.201310618
Target Identification
Phenotypic Screening to Identify Small-Molecule Enhancers for
Glucose Uptake: Target Identification and Rational Optimization of
Their Efficacy**
Minseob Koh, Jongmin Park, Ja Young Koo, Donghyun Lim, Mi Young Cha, Ala Jo,
Jang Hyun Choi, and Seung Bum Park*
Abstract: Small-molecule glucose uptake enhancers targeted
to myotubes and adipocytes were developed through a pheno-
typic screening linked with target identification and rational
optimization. The target protein of glucose-uptake enhancers
was identified as a nuclear receptor PPARg (peroxisome
proliferator-activated receptor gamma). Subsequent optimiza-
tion of initial hits generated lead compounds with high potency
for PPARg transactivation and cellular glucose uptake.
Finally, we confirmed that the chirality of optimized ligands
differentiates their PPARg transcriptional activity, binding
affinity, and inhibitory activity toward Cdk5 (cyclin-dependent
kinase 5)-mediated phosphorylation of PPARg at Ser273.
Using phenotype-based lead discovery along with early-stage
target identification, this study has identified a new small-
molecule enhancer of glucose uptake that targets PPARg.
promising candidates to treat type 2 diabetes and related
complications.
To facilitate the discovery of novel anti-diabetic agents,
we envisioned the conjunction of target identification with
phenotypic screening: 1) Image-based high-throughput
screening (HTS)[3] using a fluorescent glucose analogue,
GB2, to identify small-molecule modulators of glucose
uptake in myotubes;[4] 2) identification of the target protein
of initial hits from a phenotypic assay by FITGE (fluores-
cence difference in two-dimensional gel electrophoresis)
technology;[5] and 3) rational hit-to-lead optimization.
First, we performed the image-based HTS of a 3000-
membered drug-like compound library derived from a priv-
ileged substructure-based diversity-oriented synthesis strat-
egy[6] in a 96-well plate format. Fluorescence intensity of GB2
in the cytoplasm of the differentiated C2C12 myotubes was
quantified from captured images in an automated fashion,
which identified small-molecule enhancers of cellular glucose
uptake. Among the initial hit compounds, we were partic-
ularly interested in two isoxazole-containing compounds
(P29C06 and P29C07) because they selectively enhanced
cellular glucose uptake in C2C12 myotubes, but not in
undifferentiated C2C12 myoblasts. Their different glucose
uptake capabilities related to the differentiation status of
C2C12 cells were quantitatively investigated in myoblasts and
myotubes by direct comparison with ampkinone, a small-
molecule glucose uptake enhancer that functions by the
induction of AMP-activated protein kinase phosphorylation
(Figure 1a; Supporting Information, Figure S1a,b).[7] This
phenotypic observation suggested that the cell differentia-
tion-dependent enhancement of cellular glucose uptake may
be caused by the elevated quantity of target proteins in
myotubes compared to the pre-differentiated counterparts.
Given the fact that both P29C06 and P29C07 also enhanced
cellular glucose uptake in adipocytes, our initial hit com-
pounds were considered to affect energy homeostasis.[8]
After identifying the initial hits from the phenotypic
screening, we pursued the target identification of glucose
uptake enhancers using FITGE technology, a fluorescence-
guided target identification method involving a photoaffinity
probe.[5] Preliminary structure–activity relationship studies
led us to modify P29C06 to generate a probe (P29C06-Az) for
FITGE. As shown in Figure 1c, P29C06-Az contains aryl
azide as a photoaffinity group and an acetylene moiety as
a bioorthogonal handle for the visualization of cross-linked
proteomes using a fluorescence reporter.[9] Despite structural
changes, the probe retained its bioactivity in the myotubes
O
wing to the increasing morbidity associated with type 2
diabetes involving microvascular complications, the identifi-
cation of effective methods to control hyperglycemia is
considered a top priority in the biomedical community.[1]
Normal glucose levels can be maintained by regulating insulin
sensitivity in skeletal muscles, glycogenolysis in the liver, and
lipolysis in fat tissue.[2] Therefore, the development of novel
small molecules that regulate cellular glucose uptake or
insulin sensitivity in muscles or fat tissues would provide
[*] Dr. M. Koh, Dr. J. Park, J. Y. Koo, Dr. M. Y. Cha, A. Jo,
Prof. Dr. S. B. Park
Department of Chemistry, Seoul National University
Seoul 151-747 (Korea)
E-mail: sbpark@snu.ac.kr
D. Lim, Prof. Dr. S. B. Park
Department of Biophysics and Chemical Biology
Seoul National University
Seoul 151-747 (Korea)
Prof. Dr. J. H. Choi
School of Nano-Bioscience & Chemical Engineering
Ulsan National Institute of Science and Technology
Ulsan 689-798 (Korea)
[**] This work was supported by the Global Frontier Project Grant
(2013M3A6A4044245), the Bio & Medical Technology Development
Program (2012M3A9C4048780), the Basic Research Laboratory
(2010-0019766), and Basic Science Research Program
(2012R1A1A1015407) funded by the National Research Foundation
of Korea (NRF). M.K., J.P., J.Y.K., D.L., and A.J. are grateful for
a BK21 Scholarship.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201310618.
5102
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 5102 –5106

Angewandte
Chemie
Figure 2. Target validation and structural optimization of primary hit
compounds. a) Transactivation profiles of rosiglitazone, P29C06, and
P29C07 using a PPARg-derived reporter gene in 293T cells at 24 h after
treatment (n=3). Error bars, s.d. b) Docking analysis of 2{E,B,P}
using crystal structure of PPARg ligand-binding domain (LBD) (PDB
ID: 2hfp; Discovery Studio 1.7 [Accelrys] was used; Cys313 was
intentionally tilted for better fit; hydrogen bonding is illustrated by the
light green dashed line). c) Design of optimized leads. Detailed
synthetic procedure is described in Supporting Information.
Figure 1. Target identification of small-molecule glucose uptake
enhancers. a) Fluorescence microscopic images of C2C12 myoblasts
and myotubes 24 h after treatment with each compound using the
GB2 procedure. Scale bar: 10 mm. b) Chemical structures of ampki-
none, P29C06, and P29C07. c) FITGE-based target identification of
P29C06 using photoaffinity probe, P29C06-Az. Two-dimensional gel
images were acquired in the Cy3 or Cy5 channel using a fluorescence
scanner. The merged image shows the target protein selectively labeled
in red (white arrow).
For the validation of the outcome of FITGE-based target
identification, we subjected the initial hits to a cell-based
PPARg-luciferase transactivation assay. As shown in Fig-
ure 2a, P29C06 and P29C07 exhibited partial agonism toward
PPARg with moderate efficacy (EC50: 1.5 and 1.7 mm,
respectively). The differential pattern of cellular glucose
uptake in myotubes and myoblasts upon treatment with hit
compounds might be due to the expression level of PPARg
(Supporting Information, Figure S4). Moderate activity of the
initial hits drove us to pursue a rational drug discovery
approach with the structural and functional information of
PPARg. Based on the docking simulation, the initial hit
compound P29C06 occupies the ligand-binding site[12] of
PPARg with some extra space that can be utilized to enhance
its efficacy as well as selectivity (Figure 2b; Supporting
Information, Figure S5). Therefore, we designed and synthe-
sized a series of analogues to improve their efficacy in cellular
glucose uptake (Figure 2c). First, we confirmed that the direct
analogues of initial hits, 1{R1,R2}, having ethyl (E) group at
the R1 position and either benzyl (B) or phenethyl (PE) group
at the R2 position showed better transcriptional activity of
PPARg (Supporting Information, Figure S6). Next, we dec-
orated 1{E,B} and 1{E,PE} with a second set of isocyanates to
generate a 1,3-dicarbonyl moiety, which was predicted to be
more tightly bound at the ligand-binding site of PPARg via an
additional hydrogen bonding with Cys313 (Figure 2b). To
confirm the selectivity as well as the efficacy, we examined the
resulting series of analogues in cell-based transactivation
and adipocytes (Supporting Information, Figure S1c,d). To
induce covalent tethering of P29C06-Az to its target proteins,
the probe was incubated with live myotubes for 3 h, followed
by UV irradiation for 30 min. The cells were then lysed and
treated with an azide-containing Cy5 reporter to visualize
P29C06-Az-protein complexes by click chemistry. To elimi-
nate false-positive signals owing to non-specific labeling with
photoaffinity probe, an excess of soluble competitor (P29C06)
was added as a negative control in an identical procedure,
where the Cy3 reporter was used for bioorthogonal labeling.
The following experiments and analysis were performed as
previously described.[5] Unfortunately, we were unable to
identify target protein candidates in C2C12 myotubes, which
might be due to the low expression levels of target proteins.
On the other hand, we successfully identified a number of
potential target proteins by mass analysis in differentiated
3T3L1 adipocytes (Figure 1c; Supporting Information, Figur-
es S2, S3, Tables S1, S2). Among them, we identified PPARg
as the target protein. In fact, PPARg is a member of the
nuclear receptor family and has been shown to play important
roles in energy homeostasis and glucose metabolism[10]
associated with insulin sensitization and adipogenesis. Fur-
thermore, its agonistic ligand, rosiglitazone, has been used as
an oral therapeutic for the treatment of type 2 diabetes.[11]
Angew. Chem. Int. Ed. 2014, 53, 5102 –5106
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5103

.
Angewandte
Communications
Table 1: Transactivation profiles of WY14643, GW501516, rosiglitazone,
and compounds 1 and 2 using a PPAR-derived reporter gene in 293Tcells
at 24 h after treatment with each compound.[a]
Entry
Compounds
PPARa
PPARd
PPARg
1
2
3
1{E,B}
1{E,PE}
2{E,B,E}
>20000
IA
IA[c]
IA
IA
1510
1730
201
IA
4
2{E,B,P}
IA
IA
6.01
5
6
7
8
2{E,B,B}
2{E,B,PE}
2{E,PE,E}
2{E,PE,P}
364
3830[b]
IA
IA
51.6
42.2
>10000
180
>10000
479[b]
IA
IA
9
2{E,PE,B}
IA
IA
>10000
382
421
3.9
12.6
9.2
41.7
78.7
7.8
23.8
3.3
3.8
>10000
>10000
22
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2{E,PE,PE}
2{E,B,2MeOP}
2{E,B,3MeOP}
2{E,B,4MeOP}
2{E,B,4MP}
2{E,B,4CF3P}
2{E,B,4ClP}
2{E,B,4BP}
2{E,B,34DMeOP}
2{E,B,35DMeOP}
2{E,B,34MeDOP}
WY14643
7160[b]
IA
383[b]
17.2[b]
IA
IA
4290[b]
IA
IA
IA
IA
IA
IA
IA
IA
IA
1710[b]
IA
266[b]
IA
IA
IA
669
>10000
>10000
IA
1.4
>10000
GW501516
Rosiglitazone
[a] units: nm. [b] IC50. [c] inactive.
Figure 3. Identification of 2{E,B,35DMeOP} as novel ligands of PPARg.
a) Chemical structures of R35 and S35. b) Transactivation profiles of
rosiglitazone, R35, and S35 using a PPARg-derived reporter gene in
293T cells after 24 h treatment (n=3). Error bars, s.d. c),d) Sensor-
grams from surface plasmon resonance (SPR) assay of R35 (c) and
S35 (d) with murine PPARg ligand-binding domain (mPPARg-LBD).
e),f) Thermograms from isothermal titration calorimetry (ITC) assay of
R35 (e) and S35 (f) with mPPARg-LBD. g) Dissociation constants
related to ligand-PPARg-LBD complex formation determined with ITC
or SPR assay. R35, R-2{E,B,35DMeOP}; S35, S-2{E,B,35DMeOP}.
assays for all three subtypes of PPAR (PPARa, PPARd, and
PPARg). Among the initial 8 analogues (entries 3 to 10,
Table 1), 2{E,B,P} exhibited a drastically improved efficacy
toward PPARg from the micromolar to the nanomolar range
(1.5 mm to 6.0 nm in EC50) by the simple addition of a second
isocyanate moiety to 1{E,B}. We further diversified the phenyl
moiety of 2{E,B,P} at the R3 position with various substituents
(entries 11 to 20, Table 1) and identified 2{E,B,35DMeOP} as
the best candidate with excellent efficacy (3.3 nm in EC50) and
very high selectivity toward PPARg (entry 19, Table 1).
Prior to the in-depth biological evaluation of this com-
pound, we wanted to address the issue regarding the
stereocenter adjacent to the isoxazole ring. To this end, we
synthesized two enantiomers of the most active compound,
2{E,B,35DMeOP} (Figure 3a; Supporting Information,
Scheme S2). Interestingly, each enantiomer exhibited differ-
ent activity in the PPARg transactivation assay, with the R
enantiomer (R35) being more potent than its S counterpart
(S35; Figure 3b). Moreover, the EC50 of R35 for PPARg
transactivation was in the sub-nanomolar range, whereas that
of S35 was 480-times lower, which suggested that the binding
of R35 favors the recruitment of co-activators with proper
conformational changes of PPARg.[13] On the basis of two
independent biophysical studies using surface plasmon reso-
nance (SPR) spectroscopy and isothermal titration calorim-
etry (ITC), we confirmed that the binding affinity of R35 and
S35 to the ligand binding domain (LBD) of PPARg was
significantly enhanced from that of P29C06 (Figure 3c–g;
Supporting Information, Figures S7, S8, Tables S3, S4). These
results indicated that the marginal transcriptional activity of
P29C06 was due to its moderate binding affinity to the
PPARg LBD. Along with this, R35 with an EC50 of 0.38 nm
showed a higher binding affinity to PPARg, compared to S35
and rosiglitazone. Collectively, we postulated the direct
correlation of binding affinity with transactivation ability.
However, we also observed that S35 has a comparable
binding affinity to that of rosiglitazone, even though S35 is
a much less potent agonist than rosiglitazone. Therefore, we
suspected the mode of action of S35 to PPARg might be
different from that of rosiglitazone. To test this hypothesis, we
systematically compared our novel PPARg ligands, R35 and
S35, for lead triage. First, both R35 and S35 increased glucose
uptake to similar levels, which was measured using [14C]-2-
deoxy-d-glucose in adipocytes (Figure 4a). We then sought to
determine whether R35 and S35 might serve as inhibitors of
the Cdk5-mediated phosphorylation of PPARg,[14] which is an
emerging mode of action for the development of novel anti-
diabetic agents without the side effects of rosiglitazone, a full
PPARg agonist (Supporting Information, Figure S9).[11a,14,15]
Small-molecule ligands that inhibit the phosphorylation of
PPARg at Ser273, such as MRL24, showed an insulin-
sensitizing effect in in vitro and in vivo systems by the
5104
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 5102 –5106

Angewandte
Chemie
examined the ligand-mediated expression of genes related to
fatty acid metabolism, including aP2, CD36, and ACO, which
are known to be increased upon treatment with PPARg
agonists.[17] As expected, the treatment of 3T3L1 adipocytes
with R35 increased the expression of aP2, CD36, and ACO,
whereas S35 exerted only a marginal effect. On the other
hand, both R35 and S35 caused increases in the expression of
Adiponectin and Rybp, the genes that are most sensitive to the
phosphorylation of PPARg at Ser273.[15] Most importantly,
only S35 increased the expression level of Adipsin and
Selenbp1, whereas R35 and rosiglitazone reduced the expres-
sion of these genes (Figure 4e). These secondary biochemical
evaluations imply that stereoisomeric differences in R35 and
S35 may have different effects on gene expression, adipo-
genesis, and PPARg phosphorylation. R35 behaves as a con-
ventional PPARg agonist while S35 deviated from agonism
with high correlation in the phosphorylation inhibitory event
unlike conventional PPARg agonists. Therefore, S35 may
serve as a key structural clue for the development of new
PPARg-related therapeutic agents to treat glucose homeo-
stasis-related diseases without side effects.
In summary, we demonstrated the hit-to-lead optimiza-
tion of new small-molecule enhancers of cellular glucose
uptake, discovered from image-based phenotypic screening in
myotubes. Our initial hit compounds were subjected to
FITGE-guided target identification, which revealed PPARg
as the target protein in adipocytes. The subsequent rational
optimization of the initial hits generated a lead compound
with high potency (4000-fold enhancement from initial hit).
Secondary biophysical and biochemical studies revealed the
stereoisomeric difference in our optimized PPARg ligands
(R35 and S35) dictating their transcriptional activity, binding
affinity, and inhibitory activity toward Cdk5-mediated phos-
phorylation of PPARg at Ser273. It turned out that S35 is
a PPARg phosphorylation inhibitor with promising glucose
uptake potential, while R35 is a highly potent conventional
PPARg agonist; therefore, our PPARg ligands, especially S35,
can be utilized for the development of new anti-diabetic
agents without side effects. In light of high demands for a new
class of therapeutic agents, the novel scaffolds identified from
unbiased phenotypic screening and target identification can
be a powerful resource for the development of first-in-class
drugs.
Figure 4. Differential phenotypes upon treatment with R35 and S35.
a) Glucose uptake profiles of fully differentiated 3T3L1 cells treated
with rosiglitazone, R35, or S35 for 24 h, quantified using [14C]-2-deoxy-
d-glucose (n=4). b),c) In vitro Cdk5 assay with MRL24, R35, and S35
using PPARg or Rb as a substrate. IB, immunoblot; NT, not treated;
pPPARg, phosphorylated PPARg; pRb, phosphorylated Rb. d) Differ-
entiation status of 3T3L1 cells treated with rosiglitazone, R35, or S35
monitored by lipid accumulation with Oil Red O staining; scale bar,
10 mm. e) Relative gene expression patterns by quantitative real-time
polymerase chain reaction with RNA extracted from fully differentiated
3TL1 cells treated with rosiglitazone, R35, or S35 (1 mm each) for 24 h
(n=4). Error bars, s.e.m.; *P<0.05, **P<0.001, ***P<0.0001; ns,
not significant.
regulation of gene sets related to the phosphorylation of
PPARg.[15] To evaluate their biochemical function, we mea-
sured the inhibition of Cdk5-mediated phosphorylation of
PPARg at Ser273 upon treatment with various doses of either
R35 or S35 using an in vitro enzymatic assay with purified
Cdk5 and PPARg with MRL24 as a positive control.[14] As
shown in Figure 4b, R35 effectively blocked Cdk5-mediated
phosphorylation of PPARg with an IC50 between 10 and
100 nm, while S35 showed a slightly lower activity than R35
(IC50 between 100 and 1000 nm). Given that neither R35 nor
S35 affected Cdk5-mediated phosphorylation of the Rb
protein, a known substrate of Cdk5 (Figure 4c),[16] we were
confident that the phosphorylation of PPARg was specifically
inhibited by ligand-mediated alteration of the interaction
between Cdk5 and PPARg, rather than an inhibitory effect on
general Cdk5 function.
In fact, the full agonism on PPARg triggers adipogenesis
in fibroblasts,[11b,d] which is a major adverse effect of
rosiglitazone.[14,15] As shown in Figure 4d, R35 potentiated
adipocyte differentiation to a similar level of rosiglitazone,
confirmed by the monitoring of cellular lipid accumulation
with Oil Red O staining. However, in the case of S35, we
observed extremely low levels of lipid accumulation, which
was comparable to that of the vehicle control. Finally, we
Received: December 7, 2013
Revised: February 21, 2014
Published online: April 1, 2014
Keywords: drug discovery · glucose uptake ·
.
phenotypic screening · PPARg · target identification
[1] a) E. Cohen, A. Dillin, Nat. Rev. Neurosci. 2008, 9, 759 – 767;
b) D. M. Nathan, J. B. Buse, M. B. Davidson, E. Ferrannini,
R. R. Holman, R. Sherwin, B. Zinman, A. American Diabetes,
D. European Association for Study of, Diabetes Care 2009, 32,
193 – 203; c) C. F. Semenkovich, J. Clin. Invest. 2006, 116, 1813 –
1822.
Angew. Chem. Int. Ed. 2014, 53, 5102 –5106
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5105

.
Angewandte
Communications
[2] M. Kars, L. Yang, M. F. Gregor, B. S. Mohammed, T. A. Pietka,
B. N. Finck, B. W. Patterson, J. D. Horton, B. Mittendorfer, G. S.
Hotamisligil, S. Klein, Diabetes 2010, 59, 1899 – 1905.
[3] A. Jo, J. Park, S. B. Park, Chem. Commun. 2013, 49, 5138 – 5140.
[4] a) H. Y. Lee, J. J. Lee, J. Park, S. B. Park, Chem. Eur. J. 2011, 17,
143 – 150; b) J. Park, H. Y. Lee, M. H. Cho, S. B. Park, Angew.
Chem. 2007, 119, 2064 – 2068; Angew. Chem. Int. Ed. 2007, 46,
2018 – 2022; c) Y. S. Tian, H. Y. Lee, C. S. Lim, J. Park, H. M.
Kim, Y. N. Shin, E. S. Kim, H. J. Jeon, S. B. Park, B. R. Cho,
Angew. Chem. 2009, 121, 8171 – 8175; Angew. Chem. Int. Ed.
2009, 48, 8027 – 8031.
[5] J. Park, S. Oh, S. B. Park, Angew. Chem. 2012, 124, 5543 – 5547;
Angew. Chem. Int. Ed. 2012, 51, 5447 – 5451.
[6] S. Oh, S. B. Park, Chem. Commun. 2011, 47, 12754 – 12761.
[7] S. Oh, S. J. Kim, J. H. Hwang, H. Y. Lee, M. J. Ryu, J. Park, S. J.
Kim, Y. S. Jo, Y. K. Kim, C. H. Lee, K. R. Kweon, M. Shong, S. B.
Park, J. Med. Chem. 2010, 53, 7405 – 7413.
Albrektsen, K. S. Frederiksen, W. E. Holmes, E. Boel, K. Taylor,
J. Fleckner, Diabetes 2002, 51, 1042 – 1051; c) R. K. Semple, V. K.
Chatterjee, S. OꢀRahilly, J. Clin. Invest. 2006, 116, 581 – 589; d) P.
Tontonoz, E. Hu, B. M. Spiegelman, Cell 1994, 79, 1147 – 1156.
[12] R. T. Nolte, G. B. Wisely, S. Westin, J. E. Cobb, M. H. Lambert,
R. Kurokawa, M. G. Rosenfeld, T. M. Willson, C. K. Glass, M. V.
Milburn, Nature 1998, 395, 137 – 143.
[13] B. C. Kallenberger, J. D. Love, V. K. Chatterjee, J. W. Schwabe,
Nat. Struct. Biol. 2003, 10, 136 – 140.
[14] J. H. Choi, A. S. Banks, J. L. Estall, S. Kajimura, P. Bostrom, D.
Laznik, J. L. Ruas, M. J. Chalmers, T. M. Kamenecka, M. Blꢁher,
P. R. Griffin, B. M. Spiegelman, Nature 2010, 466, 451 – 456.
[15] J. H. Choi, A. S. Banks, T. M. Kamenecka, S. A. Busby, M. J.
Chalmers, N. Kumar, D. S. Kuruvilla, Y. Shin, Y. He, J. B.
Bruning, D. P. Marciano, M. D. Cameron, D. Laznik, M. J.
´
Jurczak, S. C. Schurer, D. Vidovic, G. I. Shulman, B. M. Spiegel-
man, P. R. Griffin, Nature 2011, 477, 477 – 481.
[16] X. GraÇa, A. De Luca, N. Sang, Y. Fu, P. P. Claudio, J.
Rosenblatt, D. O. Morgan, A. Giordano, Proc. Natl. Acad. Sci.
USA 1994, 91, 3834 – 3838.
[17] M. Schupp, M. Clemenz, R. Gineste, H. Witt, J. Janke, S.
Helleboid, N. Hennuyer, P. Ruiz, T. Unger, B. Staels, U.
Kintscher, Diabetes 2005, 54, 3442 – 3452.
[8] M. W. Rajala, P. E. Scherer, Endocrinology 2003, 144, 3765 –
3773.
[9] J. V. Staros, Trends Biochem. Sci. 1980, 5, 320 – 322.
[10] K. K. Ryan, B. Li, B. E. Grayson, E. K. Matter, S. C. Woods, R. J.
Seeley, Nat. Med. 2011, 17, 623 – 626.
[11] a) M. Ahmadian, J. M. Suh, N. Hah, C. Liddle, A. R. Atkins, M.
Downes, R. M. Evans, Nat. Med. 2013, 19, 557 – 566; b) T.
5106
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 5102 –5106