APPL1 potentiates insulin secretion in pancreatic β cells by enhancing protein kinase Akt-dependent expression of SNARE proteins in mice.

Abstract

Insulin resistance and defective insulin secretion are the two major features of type 2 diabetes. The adapter protein APPL1 is an obligatory molecule in regulating peripheral insulin sensitivity, but its role in insulin secretion remains elusive. Here, we show that APPL1 expression in pancreatic β cells is markedly decreased in several mouse models of obesity and diabetes. APPL1 knockout mice exhibit glucose intolerance and impaired glucose-stimulated insulin secretion (GSIS), whereas transgenic expression of APPL1 prevents high-fat diet (HFD)-induced glucose intolerance partly by enhancing GSIS. In both pancreatic islets and rat β cells, APPL1 deficiency causes a marked reduction in expression of the exocytotic machinery SNARE proteins (syntaxin-1, synaptosomal-associated protein 25, and vesicle-associated membrane protein 2) and an obvious decrease in the number of exocytotic events. Such changes are accompanied by diminished insulin-stimulated Akt activation. Furthermore, the defective GSIS and reduced expression of SNARE proteins in APPL1-deficient β cells can be rescued by adenovirus-mediated expression of APPL1 or constitutively active Akt. These findings demonstrate that APPL1 couples insulin-stimulated Akt activation to GSIS by promoting the expression of the core exocytotic machinery involved in exocytosis and also suggest that reduced APPL1 expression in pancreatic islets may serve as a pathological link that couples insulin resistance to β-cell dysfunction in type 2 diabetes.

Full text

Page 1

APPL1 potentiates insulin secretion in pancreatic
β cells by enhancing protein kinase Akt-dependent
expression of SNARE proteins in mice
Kenneth K. Y. Chenga, Karen S. L. Lama, Donghai Wub, Yu Wangc, Gary Sweeneyd, Ruby L. C. Hooa, Jialiang Zhanga,
and Aimin Xua,c,1
Departments ofaMedicine andcPharmacology and Pharmacy, The University of Hong Kong, Hong Kong, China;bKey Laboratory of Regenerative Biology,
Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou 510000, China; anddDepartment of Biology, York University,
Toronto, ON, Canada M3J1P3
Edited by Thomas C. Südhof, Stanford University School of Medicine, Palo Alto, CA, and approved April 12, 2012 (received for review February 11, 2012)
Insulin resistance and defective insulin secretion are the two major
features of type 2 diabetes. The adapter protein APPL1 is an oblig-
atory molecule in regulating peripheral insulin sensitivity, but its
role in insulin secretion remains elusive. Here, we show that APPL1
expression in pancreatic β cells is markedly decreased in several
mouse models of obesity and diabetes. APPL1 knockout mice ex-
hibit glucose intolerance and impaired glucose-stimulated insulin
secretion (GSIS), whereas transgenic expression of APPL1 prevents
high-fat diet (HFD)-induced glucose intolerance partly by enhanc-
ing GSIS. In both pancreatic islets and rat β cells, APPL1 deficiency
causes a marked reduction in expression of the exocytotic machin-
ery SNARE proteins (syntaxin-1, synaptosomal-associated protein
25, and vesicle-associated membrane protein 2) and an obvious
decrease in the number of exocytotic events. Such changes are
accompanied by diminished insulin-stimulated Akt activation. Fur-
thermore, the defective GSIS and reduced expression of SNARE
proteins in APPL1-deficient β cells can be rescued by adenovirus-
mediated expression of APPL1 or constitutively active Akt. These
findings demonstrate that APPL1 couples insulin-stimulated Akt
activation to GSIS by promoting the expression of the core exo-
cytotic machinery involved in exocytosis and also suggest that re-
duced APPL1 expression in pancreatic islets may serve as a path-
ological link that couples insulin resistance to β-cell dysfunction in
type 2 diabetes.
T
actions and secretion of insulin. β-cell dysfunction is the major
contributor to the development of T2DM (1). Impaired first-
phase insulin secretion is often observed in the very early stage of
T2DM and also in impaired glucose tolerance (2). However, the
molecular mechanism underlying the pathogenesis of β-cell dys-
function is vaguely understood.
Although the peripheral metabolic tissues are the main targets
of insulin, mounting evidence from animal and human studies
suggests that the β cell itself also possesses an insulin signaling
system,whichplaysacriticalroleinregulatingβ-cellmass,survival,
insulin biosynthesis, and secretion (3). β-cell–specific inactivation
of several components involved in insulin signaling, including in-
sulin receptor (IR), insulin receptor substrate (IRS)-2, class IA
phosphatidylinositol 3-kinase (PI3K), and Akt, leads to impaired
insulin secretion and/or decreased β-cell mass (4–7). By contrast,
transgenic expression of active Akt or IRS-2 in β cells increases
β-cell mass and enhances insulin secretion, thereby rendering the
mice resistant to experimental diabetes (8, 9). In humans, the ex-
pression levels of several insulin signaling molecules (IR, IRS-2,
and Akt2) are reduced in pancreatic islets isolated from patients
withT2DM(10).Ontheotherhand,earlyintensiveinsulintherapy
can improve β-cell function and glycemic control in patients with
T2DM, further supporting the beneficial effect of insulin on β-cell
functions (11). However, the molecular pathways that couple in-
sulin signaling to insulin secretion in the β cells remain enigmatic.
ype 2 diabetes mellitus (T2DM) is a heterogeneous meta-
bolic disease resulting from a combined defect in both
APPL1, an adapter protein containing an NH2-terminal Bin/
Amphiphysin/Rvs domain (BAR), a central pleckstrin homology
(PH) domain, and a COOH-terminal phosphotyrosine-binding
domain(PTB),is a critical regulatorof endocytoticpathways(12).
As a binding partner of Akt2 (13), this adapter protein has
emerged as an important mediator of insulin actions in the liver
(14),skeletalmuscles(15),adipocytes(16),andendothelium(17).
In the liver, APPL1 potentiates the inhibitory effects of insulin on
hepatic gluconeogenesis via Akt activation, and overexpression of
APPL1 in liver alleviates hyperglycemia in db/db diabetic mice
(14). In skeletal muscles and adipocytes, APPL1 mediates insulin-
stimulated glucose uptake via controlling the translocation of cy-
tosolic glucose transporter 4 to the plasma membrane (15, 16). In
the blood vessels, APPL1 counteracts obesity-induced vascular
insulinresistancebyshiftingthevasoconstrictoreffectofinsulinto
nitric oxide-dependent vasodilatation (17). Furthermore, APPL1
interacts with the adiponectin receptors, thereby mediating the
insulin-sensitizing effects of this adipokine (18, 19).
Although APPL1 is abundantly expressed in the pancreas (13),
its roles in insulin secretion and β-cell functions have never been
explored. During the metabolic phenotyping analysis of APPL1
knockout (KO) mice (17), we observed a marked impairment of
glucose-stimulated insulin secretion (GSIS) in this rodent model.
In this study, we further investigated the pathophysiological role of
APPL1 in modulating insulin secretion in both lean and obese
mice,andelucidatedthemolecularmechanismsunderlyingAPPL1
functions in pancreatic β cells.
Results
APPL1 Expression in Pancreatic Islets Is Decreased in both Dietary and
Genetic Obese Mice. A previous study has detected a high level of
pancreatic APPL1 mRNA expression in mice (13). We further
examined the expression and distribution pattern of APPL1 in
pancreatic tissues. Both immunofluorescence staining and cell
fractionation analysis detected a much higher level of APPL1
protein in islet fraction compared with nonislet fraction (exocrine
cells) in the pancreas of C57 lean mice (Fig. 1 A and B). In pan-
creatic islets, a large proportion of APPL1 colocalized with in-
sulin, suggesting that this adapter protein is abundantly expressed
in β cells. Notably, APPL1 expression was significantly decreased
Author contributions: K.K.Y.C., K.S.L.L., and A.X. designed research; K.K.Y.C., R.L.C.H., and
J.Z. performed research; D.W. contributed new reagents/analytic tools; K.K.Y.C., D.W.,
Y.W., G.S., and A.X. analyzed data; and K.K.Y.C., K.S.L.L., Y.W., G.S., and A.X. wrote
the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
See Commentary on page 8795.
1To whom correspondence should be addressed. E-mail: amxu@hkucc.hku.hk.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1202435109/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1202435109PNAS
| June 5, 2012
| vol. 109
| no. 23
| 8919–8924
BIOCHEMISTRY
SEE COMMENTARY

Page 2

in islet cells of both high-fat diet (HFD)-induced obese mice and
genetically inherited db/db obese/diabetic mice compared with
those of lean controls (Fig. 1 A–C), whereas there was no obvious
difference in APPL1 expression in nonislet cells or other meta-
bolic tissues, including liver and skeletal muscles (Fig. S1). In line
with previous reports (20, 21), we found that pancreatic islets
isolated from both diet-induced obese mice and db/db diabetic
micedisplayedasignificantreductioninGSIScomparedwithage-
matched lean controls (Fig. 1D).
GeneticAblation of APPL1CausesGlucoseIntoleranceandImpairment
of Insulin Secretion in Mice. We next investigated the physiological
role of APPL1 in insulin secretion and glucose metabolism using
APPL1 KO mice as recently described (17). Body weight gain
(Fig. S2A) and food intake (Table S1) were comparable between
male APPL1 KO mice and wild-type (WT) littermates when they
were fed with a standard chow (STC) or HFD for a period of
20 wk. APPL1 KO mice on either STC or HFD had significantly
higher fasting glucose and insulin levels compared with WT con-
trols (Table S1). Glucose tolerance test (GTT) revealed a mild
but significant impairment in glucose disposal in STC fed-APPL1
KO mice relative to WT controls, whereas a much more severe
glucose intolerance was observed in APPL1 KO mice fed with
HFD (Fig. 2A). In APPL1 KO mice on both STC and HFD,
insulin secretion during i.p. glucose challenge was markedly at-
tenuated compared with their WT littermates (Fig. 2B). Fur-
thermore, insulin secretion induced by L-arginine, which directly
induces membrane depolarization of β cells, was compromised in
APPL1 KO mice (Fig. S2B), suggesting that the lack of APPL1
causes a generalized defect in insulin secretion. APPL1 KO mice
also developed more severe insulin resistance, as determined by
insulin tolerance test (ITT) (Fig. S2 C and D). Taken together,
these findings suggest that glucose intolerance in APPL1 KO
mice is attributed in part to impaired insulin secretion.
Although the role of APPL1 in peripheral insulin actions is now
welldocumented(14–16),whetheritregulatesinsulinsecretionhas
notbeenexplored.Therefore,wefurtherinvestigatedtheimpactof
APPL1deficiencyoninsulinsecretioninisolatedmousepancreatic
islets. The islets from STC fed-APPL1 KO mice exhibited a com-
parable level of basal insulin secretion, but significantly lower
GSIS compared with WT controls (Fig. 2C). In addition, insulin
secretion stimulated by potassium chloride (KCl), which directly
depolarizes the cell membrane and induces calcium influx, was
blunted in the islets of APPL1 KO mice (Fig. 2C). Impairment of
insulin secretion in response to both glucose and KCl stimulation
wasalso observed in islets ofAPPL1KOmiceon HFD(Fig.S3A).
Tofurtherexaminethedynamicsofinsulinsecretion,perifusion
experiment was performed in islets of APPL1 KO and WT mice.
ThisanalysisdemonstratedthatAPPL1deficiencyledtoamarked
reduction in the first phase (6–18 min) of GSIS, but had little
effect on the second phase (18–42 min) of GSIS (Fig. 2 D and E).
A
Insulin APPL1 Merged
Lean
(BL6)
HFD
(BL6)
db/db
(BKS)
Lean
(BKS)
B
Lean
BKS
db/db
BKS
HFD
BL6
Lean
BL6
APPL1
GAPDH
Islet Non-islet
Lean
BKS
HFD
BL6
Lean
BL6
db/db
BKS
0
10
20
30
40
Insulin secretion
(ng/ml/islet/hr)
Basal11 mM 25 mM
Glucose
*
*
*
*
Lean (BKS)
Lean (BL6)
db/db (BKS)
HFD (BL6)
D
Lean (BKS)
Lean (BL6)
db/db (BKS)
HFD (BL6)
N.S.
N.S.
Relative APPL1
expression
0.0
0.5
1.0
1.5
**
C
Islet Non-islet
Fig. 1.
both dietary and genetic obese mice. (A) Immunofluorescence staining of
APPL1 (green) and insulin (red) in pancreas sections of 12-wk-old male
C57BL/6J (BL6) lean mice on standard chow, or high-fat diet (HFD), or C57BKS
db/+ lean mice or C57BKS (BKS) db/db obese/diabetic mice. (B) Western blot
analysis of APPL1 in nonislet (exocrine cells) and islet fractions isolated from
the pancreas of C57BL/6J lean, or HFD-fed obese mice, or C57BKS db/+ lean
or db/db mice. (C) Densitometric analysis for the relative abundance of
APPL1 as in B. (D) Static GSIS in islets isolated from different mouse models as
specified. *P < 0.05 (n = 5). NS, not significant.
Decreased APPL1 expression in pancreatic islets and impaired GSIS in
A
B
0
0
1
2
3
4
10
Time (minute)
2030
Insulin level (fold change)
*
*
*
*
*
*
Glucose level (mmol/L)
015 30
Time (minute)
45 607590
0
5
10
15
20
25
30
35
*
*
*
*
*
*
*
*
Wt-STC
KO-STC
Wt-HFD
KO-HFD
Wt-STC
KO-STC
Wt-HFD
KO-HFD
C
D
Insulin secretion (fold change)
E
Wt-STC
KO-STC
Basal11
25 50 mM
KCl
( M)
(mM)
Glucose
0
10
20
30
40
Insulin secretion
(ng/ml/islet/hr)
*
*
*
Wt-STC
KO-STC
*
*
0 6 12 18 24 30 36 42
Time (minute)
0
4
8
12
**
Wt-STC
KO-STC
AUC (arbitrary unit)
0
5
10
15
*
6-18 18-42
Time (minute)
N.S
Glucose (25 mM)Basal
()
Fig. 2.
tion in mice. (A) GTT in 20-wk-old male APPL1 KO and WT littermates fed
with HFD or STC. (B) In vivo GSIS in APPL1 KO mice and WT littermates fed
with STC or HFD (fold change over basal insulin level). (C) Glucose- and KCl-
stimulated static insulin secretion in islets isolated from 20-wk-old male
APPL1 KO mice and WT controls on STC. Insulin concentration was measured
at 30 min after stimulation. (D) Dynamic insulin secretion of pancreatic islets
isolated from APPL1 KO and WT mice in response to glucose (25 mM) using
a perifusion system. (E) Area under curve (AUC) for the first phase (6–18 min)
and second phase (18–42 min) of insulin secretion during the perifusion
experiment as in D. *P < 0.05 (n = 6–8). NS, not significant.
Effects of APPL1 deficiency on glucose tolerance and insulin secre-
8920
| www.pnas.org/cgi/doi/10.1073/pnas.1202435109Cheng et al.

Page 3

Consistent with the above findings in isolated mouse islets, rat
INS-1E β cells infected with adenovirus encoding APPL1 RNAi
exhibited an ∼75% reduction in APPL1 expression (Fig. S3B),
and this change was accompanied by a significantly decreased
insulin secretion induced by glucose and KCl (Fig. S3C). Taken
together, these in vivo and in vitro findings support an obligatory
role of APPL1 in insulin secretion.
Transgenic Expression of APPL1 Prevents Obesity-Induced Glucose
Intolerance and Defective Insulin Secretion. To test whether over-
expression of APPL1 alleviates HFD-induced decrement of glu-
cose tolerance and insulin secretion, we next performed GTT and
GSIS in APPL1 transgenic mice (17) and WT controls on both
STC and HFD. The expression level of APPL1 in pancreatic islets
of the transgenic mice was increased by ∼3.5-fold compared with
WT controls (Fig. 3A). Transgenic expression of APPL1 had no
obvious impact on food intake and body weight gain, but caused
a significant alleviation in HFD-induced fasting hyperglycemia
and hyperinsulinemia (Table S1). In response to glucose chal-
lenge, glucose excursion curves were similar between APPL1
transgenic mice and WT controls fed with STC, whereas HFD-
induced glucose intolerance was significantly reduced by trans-
genicexpression of APPL1(Fig.3B). Furthermore,HFD-induced
impairment in GSIS was partially reversed by the transgenic ex-
pression of APPL1 (Fig. 3C). The islets isolated from STC-fed
APPL1transgenicmicedisplayedatrendtowardenhancedinsulin
secretion in response to glucose (11 mM) or KCl (50 mM) stim-
ulation (Fig.3D), andsuch changes becamemuch moreevidentin
the islets of HFD-fed APPL1 transgenic mice compared with the
islets isolated from WT controls (Fig. 3E). Insulin sensitivity as
assessed by ITT was similar between APPL1 transgenic mice and
WT controls, whereas HFD-induced insulin resistance was sig-
nificantly alleviated by transgenic expression of APPL1 (Fig. 3F).
APPL1 KO Mice Exhibit Decreased Expression of SNARE Proteins and
Impaired Exocytosis. To further delineate the mechanism that
causesdefectiveinsulinsecretioninAPPL1KOmice,weanalyzed
β-cell mass, insulin content, ATP production, and calcium influx
in response to glucose challenge. The pancreatic β-cell mass was
significantly increased in both STC- and HFD-fed APPL1 KO
mice compared with the WT controls (Fig. S4A). The total insulin
content in the isolated pancreatic islets was not different between
APPL1KOmiceandWTlittermates(Fig.S4B).Glucose-induced
ATP production and calcium influx in pancreatic islets were
comparable between the two groups (Fig. S4 C and D), suggesting
thatdefectiveinsulinsecretioninAPPL1KOmiceoccursatastep
downstream of calcium influx. The expression levels of key genes
involved in the regulation of insulin synthesis (insulin and pan-
creatic and duodenal homeobox 1, PDX-1) and glucose metabo-
lism (glucose kinase, GCK and glucose transporter 2, GLUT2)
were also similar between the islets of APPL1 KO mice and WT
A
APPL1
ol/L)
Glucose level (mm
B
Wt-HFD
Tg-HFD
Wt-STC
Tg-STC
25
20
Wt Tg
ld change)
Insulin level (fo
Wt-HFD
Tg-HFD
Wt-STC
Tg-STC
4
CD
30
40
cretion
Insulin se
let/hr)
(ng/ml/is
Wt-STC
Tg-STC
p=
0.080
p=
0.055
GAPDH
Relative APPL1
*
0 10 20 30 40 50 60 70 80 90
Time (minute)
0
5
10
15
*
*
*
0
1
2
3
4
expression
WtTg
*
*
*
0
0
1
2
3
10
Time (minute)
2030
0
10
20
Basal11 mM 25 mM 50 mM
Glucose
KCl
secretion
Insulin
l/islet/hr)
(ng/m
15
10
20
25
Wt-HFD
Tg-HFD
*
*
*
Wt-HFD
Tg-HFD
Wt-STC
Tg-STC
EF
0 40.4
0.8
1.2
ose level
Gluc
baseline)
(% of
*
*
*
*
0
5
Basal 11 mM 25 mM 50 mM
Glucose
KCl
020
Time (minute)
406080
0.0
Fig. 3.
sulin sensitivity and enhances insulin secretion in HFD-fed obese mice. (A)
Western blot analysis for APPL1 expression in isolated pancreatic islets of 12-
wk-old male APPL1 transgenic (Tg) mice and WT littermates. (B) GTT in 16-
wk-old APPL1 Tg mice and WT littermates fed with either STC or HFD. (C) In
vivo GSIS in 16-wk-old male mice, and plasma insulin level was expressed as
fold change over the baseline. (D and E) Static insulin secretion in response
to glucose and KCl stimulation in islets from 20-wk-old APPL1 Tg and WT
littermates on STC or HFD, respectively. (F) ITT in 20-wk-old APPL1 Tg mice
and WT littermates fed with either STC or HFD. *P < 0.05 (n = 6–7).
Transgenic expression of APPL1 improves glucose tolerance and in-
e mRNA level
Relative
0 50
0.25
0.75
.
1.00
1.25
*
*
*
SNAP25
VAMP2
APPL1
STX1
Wt
KO
Wt
KO
SNAP25 STX1 VAMP2
0.00
ß-tubulin
0.0
0.4
0.8
1.2
Protein expression
SNAP25 STX1 VAMP2 STX1 VAMP2
*
*
*
Wt
KO
0
2
4
6
8
No. of docked granule
Wt KO
Wt
KO
AB
CD
*
EFG
No. of exocytotic events
2 /5 min
KO
0.0
0.5
1.0
1.5
2.0
KO
Wt
Latency (second)
N.S.
Wt
KO
10 µm
0
5
10
15
20
Wt
25
*
Fig. 4.
reduces the number of docked insulin granules and exocytotic events in
pancreatic islets. (A) Real-time PCR analysis for mRNA expression levels of the
threeSNARE proteins(SNAP25, STX1,andVAMP2) inpancreatic isletsisolated
from 20-wk-old male APPL1 KO mice or WT controls on STC. (B) Western blot
analysis for the SNARE proteins in the isolated islets. (C) Densitometric anal-
ysis for the relative abundance of SNARE proteins as in B. (D) Representative
electron microscopic images of docked insulin granules (Left, denoted with
arrows), and quantification of the number of docked granules (Right) in the
isolated islets. (E) Distribution of exocytotic events in isolated islets of APPL1
KO mice and WT controls as measured by two-photon excitation imaging.
White dots represent sites where exocytotic events were observed during
5 min of glucose (20 mM)-stimulated insulin secretion. The underlying red
image is sulforhodamine B fluorescence. (F) Average number of glucose-
stimulated exocytotic events inacell surface area of1,000 μm2. (G) Latency (in
seconds) for the onset of staining with sulforhodamine B relative to that of
staining with 10 kDa dextran. *P < 0.05 (n = 5–6). NS, not significant.
APPL1 deficiency decreases the expression of SNARE proteins and
Cheng et al.PNAS
| June 5, 2012
| vol. 109
| no. 23
| 8921
BIOCHEMISTRY
SEE COMMENTARY

Page 4

controls (Fig. S4E). On the other hand, APPL1 KO mice exhib-
ited a significant down-regulation in mRNA expression of several
soluble N-ethylmaleimide-sensitive factor attachment protein re-
ceptor (SNARE) proteins, including syntaxin-1 (STX1), synap-
tosomal-associated protein 25 (SNAP25), and vesicle-associated
membrane protein 2 (VAMP2) (Fig. 4A), which are the core
components of the exocytotic machinery of eukaryotic cells (22).
The mRNA expression level of Sec1/Munc18 (Munc18), Rab3a,
andsynaptotagminVII,whicharealsoinvolvedintheregulationof
insulin secretion through SNARE proteins (23, 24), was compa-
rable between the two groups of mice (Fig. S4E).
Consistent with the aforementioned changes in mRNA ex-
pression, Western blot analysis demonstrated a marked reduction
in protein levels of the three SNARE proteins in the islets of
APPL1 KO mice, compared with their WT controls (Fig. 4 B and
C). Notably, decreased expression of these SNARE proteins was
accompanied by a reduced number of docked insulin granules at
the plasma membrane in the islets of APPL1 KO mice (Fig. 4D).
By contrast, the islets from APPL1 transgenic mice displayed an
increase in docked insulin granules (Fig. S5). We next examined
the role of APPL1 in insulin granule exocytosis and fusion pore
dynamics with two-photon excitation imaging (25). In response to
glucosestimulation(20mM),theexocytoticeventsduring thefirst
5 min were significantly reduced in the islets of APPL1 KO mice
compared with WT controls (Fig. 4 E and F), further supporting
anindispensableroleofAPPL1inthefirst-phaseinsulinsecretion.
On the other hand, expansion of the fusion pore as detected by
measuring latency for the onset of staining with fluorescent
markers of different sizes was comparable between the islets of
APPL1 KO mice and WT littermates (Fig. 4G), implying that
APPL1 deficiency has no effect on the fusion dynamics.
In INS-1E β cells, adenovirus-mediated knockdown of APPL1
expression resulted in a significant reduction in expression of
SNARE proteins (Fig. S6A), whereas adenovirus-mediated
expression of APPL1 augmented the expression of SNAP25,
STX1, and VAMP2 in primary β cells isolated from APPL1 KO
mice (Fig. 5A). The latter change was associated with increased
glucose- and KCl-induced insulin secretion (Fig. 5 B–D). Taken
together, these findings suggest that defective insulin secretion in
APPL1 KO mice is attributed at least in part to decreased ex-
pression of the SNARE proteins in β cells.
Defective Insulin Secretion in APPL1-null β Cells Is Rescued by Akt
Activation. The Akt/FoxO1 signaling cascade plays a pivotal role
in the regulation of SNARE protein expression and insulin se-
cretion (4, 26). Because APPL1 has been shown to potentiate Akt
signaling in adipocytes (16), hepatocytes (14), and muscle cells
(15), we next investigated whether APPL1 deficiency has any
impact on this signaling cascade in pancreatic islets. Insulin-
stimulated phosphorylation of Akt at Ser-473 and FoxO1 at Ser-
256 in the islets of APPL1 KO mice was significantly decreased
compared with those of WT controls, whereas there was no ob-
vious difference in insulin-induced phosphorylation of ERK1/2
(Thr-202/Tyr-204) between the two groups of mice (Fig. 6A).
A
APPL1
STX1
STX1
SNAP25
VAMP2
GAPDH
KO+
Luci
KO+
APPL1
KO+Luci
KO+APPL1
*
*
2.0
1.5
2.5
2.0
ession
Protein expre
*
20
30
ecretion
Insulin se
slet/hr)
(ng/ml/i
KO+Luci
KO+APPL1
*
*
B
0.0
0.5
1.0
STX1SNAP25 VAMP2
BasalGlucose
11 mM
KCl
50 mM
0
10
CD
KO+Luci
KO+APPL1
N.S.
*
KO+Luci
KO+APPL1
10
5
15
20
arbitrary unit)
AUC (a
*
*
*
5.0
2.5
7.5
10.0
12.5
n secretion
Insulin
d change)
(fold
*
Time (minute)
6-18 18-42
0
06 12 18 24 30 36 42
Time (minute)
0.0
Basa
l
Glucose (25 mM)
Fig. 5.
expression and insulin secretion in the islets of APPL1 KO mice. (A) Islets
isolated from 20-wk-old male APPL1 KO mice on STC were infected with
recombinant adenovirus expressing APPL1 or luciferase (Luci) for 36 h, fol-
lowed by Western blot analysis to detect the expression levels of the three
SNARE proteins. Bar chart on the Right is the densitometric analysis for the
relative abundance of the three SNARE proteins in the infected islets. (B)
Static insulin secretion in the infected islets measured after stimulation with
glucose or KCl for 30 min. (C) First phase (6–18 min) and second phase (18–42
min) of GSIS of the infected islets using the perifusion system. (D) AUC for
the first and second phase of insulin secretion during the perifusion exper-
iment as in C. *P < 0.05 (n = 8). NS, not significant.
Adenovirus-mediated expression of APPL1 enhances SNARE protein
Phospho/total
0.0
0.5
1.0
1.5
2.0
2.5
N.S.
*
Wt
KO
*
3.0
AInsulin 0 10 0 10 min
pAkt
Akt
Wt
KO
pFoxO1
FoxO1
pERK
ERK
GAPDH
p-FoxO1
Insulin0 10 0 10 0 10 min
p-Akt
p-ERK
Wt+Luci
KO+Luci
KO+CA-Akt
KO+L
KO+Luci
KO+CA-Akt
N.S.
N.S.
i
Wt+Luci
BC
STX1
SNAP25
VAMP2
GAPDH
KO
Myc
Wt
CA
-Akt
LuciLuci
STX1 SNAP25 VAMP2
0.0
0.5
1.0
1.5
Relative mRNA level
**
* *
*
*
KO+Luci
KO+CA-Akt
N.S.
Wt+Luci
N.S.
N.S.
Glucose
Basal 11 mM
25 mM
0
10
20
30
40
Wt+Luci
Insulin secretion
(ng/ml/islet/hr)
*
*
*
*
E
STX1 SNAP25 VAMP2
0.0
0.5
1.0
1.5
Protein expression
*
*
*
*
*
*
N.S.
D
Fig. 6.
activation. (A) Islets from 20-wk-old male APPL1 KO mice and WT controls on
STC were treated with insulin (50 nM) for various time points as indicated,
followed by Western blot analysis using antitotal or phospho-FoxO1 (Ser-256),
antitotal or phospho-Akt (Ser-473), antitotal or phospho-ERK1/2 (Thr-202/Tyr-
204), or anti-GAPDH antibody as a loading control. Bar chart on the Right
represents the relative fold change of phosphorylation as quantified by densi-
tometry. (B) Real-time PCR analysis for mRNA expression levels of the three
SNARE proteins in the islets infected with adenovirus encoding the constitu-
tively active form of Akt (CA-Akt) or luciferase (Luci) for 48 h. (C) Western blot
analysisforproteinexpressionlevelsofthethreeSNAREproteinsoftheinfected
islets. Note that CA-Akt is tagged with a Myc epitope at the NH2 terminus and
can be detected by an anti-Myc antibody. (D) Densitometric analysis for the
relative abundanceof the three SNARE proteins in the infected islets asinC. (E)
Static GSIS in the infected islets (n = 4). *P < 0.05 (n = 6). NS, not significant.
APPL1 induces SNARE protein expression and insulin secretion via Akt
8922
| www.pnas.org/cgi/doi/10.1073/pnas.1202435109 Cheng et al.

Page 5

Likewise, RNAi-mediatedreduction in APPL1 expression in INS-
1EcellssignificantlyenhancedtheinteractionbetweenAktandits
endogenous inhibitor tribble homolog 3 (TRB3) (Fig. S6B) and
blunted insulin-stimulated phosphorylation of Akt and FoxO1
(Fig.S6C).Ontheotherhand,adenovirus-mediatedexpressionof
APPL1 reversed the impairment in insulin-stimulated Akt/FoxO1
phosphorylation in APPL1-null β cells (Fig. S7A). Similarly,
transgenic expression of APPL1 augmented the phosphorylation
of Akt and FoxO1 induced by insulin in the islets of APPL1
transgenic mice on HFD (Fig. S7B).
To test whether the activation of Akt could reverse defective
insulin secretion in APPL1-null β cells, islets isolated from
APPL1 KO mice were infected with adenovirus encoding lu-
ciferase or a constitutively active form of Akt (CA-Akt). This
analysis demonstrated that decreased expression of the three
SNARE proteins in islets of APPL1 KO mice was largely re-
versed by adenovirus-mediated expression of CA-Akt but not by
the luciferase control (Fig. 6 B–D), and this change was ac-
companied by the restoration of GSIS in the APPL1-null β cells
(Fig. 6E).
Discussion
Defective insulin secretion is a major feature of β-cell dysfunc-
tion and a primary contributor to hyperglycemia in T2DM.
However, because insulin secretion is a highly dynamic process
regulated by complex mechanisms, the pathological pathways
leading to β-cell dysfunction remain poorly characterized. In the
present study, we provide both in vitro and in vivo evidence
showing that APPL1, a multidomain adapter protein involved in
the peripheral actions of insulin (14–16), is a physiological reg-
ulator of insulin secretion in pancreatic β cells. GSIS is blunted in
APPL1 KO mice, but is augmented by transgenic expression of
APPL1. Despite impaired GSIS, it is noteworthy that basal
plasma insulin levels in APPL1 KO mice are much higher than
those in WT littermates. Increased basal insulin secretion in
APPL1 KO mice is perhaps due to the compensatory function of
β cells, and may also explain why glucose intolerance is not ob-
vious in young APPL1 KO mice (27).
Loss of first-phase insulin secretion occurs in the very early
stage of type 2 diabetes (2). In isolated islets from APPL1 KO
mice, we found that the impaired GSIS is mainly attributed to
defective first-phase insulin secretion. This conclusion is also
supported by our observation that the number of docked insulin
granules, a major pool for first-phase insulin secretion (2), is
decreased in APPL1-null islets but increased in APPL1 trans-
genic islets. Furthermore, our data from both gain-of-function
and loss-of-function studies suggest that APPL1 may promote
insulin exocytosis by increasing the expression of SNARE pro-
tein complex, a core component involved in calcium-induced
docking and exocytosis of secretory granules (22, 28).
The three main components of SNARE proteins (STX1,
SNAP25, and VAMP2) promote exocytosis of insulin granules by
forming a heterotrimeric complex, thereby facilitating membrane
fusion, priming and docking of secretory vesicles (22, 28). The
importance of SNARE-mediated exocytosis in the first-phase in-
sulin secretion has been documented in both in vitro and animal
studies (29), and decreased expression of the three SNARE pro-
teins has been suggested as a possible cause of defective insulin
secretion in rodents and patients with obesity and T2DM (30, 31).
Notably, β cells from STX1 KO mice display impaired first-phase
insulin secretion and defective docking and fusion of insulin
granules (29), a pattern similar to APPL1-null islets observed in
our study. In both the pancreatic islets of APPL1 KO mice and rat
INS-1E β cells with decreased APPL1 expression, we found that
the expression levels of all three SNARE proteins are markedly
decreased, whereas this change is reversed by adenovirus-medi-
ated replenishment of APPL1. These findings suggest that de-
creased APPL1 expression in islet β cells may contribute to
decreased expression of the SNARE proteins in obesity and di-
abetes, which in turn causes defective insulin secretion.
APPL1has been reported tomodulate bothsubstratespecificity
andactivityofAkt(14–17,32),aproteinkinasethatplaysacentral
role in mediating the peripheral actions of insulin. In addition, the
PI3K/Akt signaling pathway has been shown to control insulin
secretion at the step of exocytosis, partly via enhancing the ex-
pression of SNARE proteins (4, 7). The promoter regions of the
SNAREcomplex genescontainthebindingsitesforFoxO1,which
suppresses the transcription of these genes (4). Akt activation in β
cellsinduces thetranscriptional activation of theSNARE complex
genes by phosphorylating and inactivating FoxO1, but has no ob-
vious effect on expression of several other exocytotic proteins
(Rab3a, Munc18, and synaptotagmin) (4, 7). Pancreatic β-cell–
specificablationofPI3Kleadstoglucoseintoleranceanddefective
insulin secretion in mice, and PI3K-null β cells exhibit defective
exocytosis of insulin granules due to the reduced expression of
SNARE proteins and loss of cell–cell synchronization (4). These
defectscanberescuedbyforcedactivationofAktorinactivationof
the Akt downstream target FoxO1 in PI3K-null cells (4). Likewise,
transgenic mice with diminished Akt activity in β cells (7) display
glucose intolerance due todefective insulin secretionatthe level of
exocytosis,whereasglucose-stimulatedcalciuminfluxisnotaltered.
By contrast, transgenic mice with β-cell–specific activation of Akt
can protect against streptozotocin-induced diabetes by preserving
β-cell mass and GSIS (8). In line with these findings, several pieces
of evidence from the present study demonstrated that APPL1-
mediated expression of the three SNARE proteins is attributed to
theabilityofthisadapterproteininpotentiatinginsulin-evokedAkt
activation. First, in the islets of APPL1 KO mice and rat INS-1E β
cells with RNAi-mediated knockdown of APPL1, decreased ex-
pressionofthe threeSNAREproteinsisaccompaniedby a marked
attenuation in insulin-elicited phosphorylation of Akt and its
downstream target FoxO1. Second, adenovirus-mediated expres-
sion of constitutively active Akt is sufficient to rescue the reduction
in expression of the three SNARE proteins in APPL1-null islet
cells.Takentogether,thesedatasupporttheroleofAPPL1asakey
regulatorthatcouplesinsulinsignalingtoinsulinsecretioninβcells.
As an interacting partner of Akt (13), APPL1 has been shown
to enhance Akt activity by competing with the pseudokinase
TRB3 for Akt binding sites in hepatocytes and endothelial cells
(14, 17). Likewise, the present study found that the impaired
insulin-evoked Akt activation by knockdown of APPL1 expres-
sion is accompanied by enhanced interaction between Akt and
TRB3 in β cells. TRB3 acts as an endogenous Akt inhibitor by
trapping Akt within the cytosol and preventing its translocation
to the plasma membrane (14). By contrast, APPL1 releases Akt
trapped by TRB3 and promotes Akt translocation to the plasma
membrane and endosomes for further activation (14). Notably,
TRB3 expression is markedly elevated in islets from patients
with T2DM as well as insulin receptor-deficient mice (26).
Overexpression of TRB3 in both mouse β cells and human islet
cells blocks docking and exocytosis of insulin granules by down-
regulating the expression of the SNARE proteins (26). There-
fore, APPL1 and TRB3 display opposite changes in the islets of
obesity and diabetes and exert opposing effects in modulating
the expression of SNARE proteins and insulin exocytosis, sug-
gesting that these two Akt-binding proteins may act as a “yin-
and-yang” pair to fine-tune insulin secretion, by tightly control-
ling the expression of key exocytotic genes in β cells.
In conclusion, the present study has identified APPL1 as an
indispensable component that controls insulin secretion in pan-
creatic β cells, by modulating the expression of SNARE proteins
via an Akt-dependent pathway. Our data, together with the pre-
vious findings showing that APPL1 mediates the metabolic effects
of insulin in its peripheral targets, support the role of APPL1 as
a master coordinator controlling both secretion and actions of
insulin. These findings further highlight the importance of insulin
Cheng et al.PNAS
| June 5, 2012
| vol. 109
| no. 23
| 8923
BIOCHEMISTRY
SEE COMMENTARY