Articles in PresS. Am J Physiol Endocrinol Metab (November 3, 2015). doi:10.1152/ajpendo.00296.2015
1
Enhanced Insulin Sensitivity Mediated by Adipose Tissue Browning Perturbs Islet
2
Morphology and Hormone Secretion in Response to Autonomic Nervous Activation in
3
Female Mice
4
Bilal Omar1*, Martina Kvist-Reimer1, Sven Enerbäck2 and Bo Ahrén1
5 6
1
Lund University, Department of Clinical Sciences, Lund Sweden SE22184
2
Gothenburg University, Sahlgrenska Hospital, Gothenburg Sweden SE40530
7
Correspondence:
[email protected], tel- +46 46 2220760, fax- +46 462224022
8
Running Title: Autonomic mediated islet adaptation to insulin sensitivity
9 10 11 12 13 14 15 16 17 18 19
1 Copyright © 2015 by the American Physiological Society.
20
ABSTRACT
21
Insulin resistance results in compensatory increase in insulin secretion to maintain
22
normoglycemia. Conversely, high insulin sensitivity results in reduced insulin secretion to
23
prevent hypoglycemia. The mechanisms for this inverse adaptation are not well understood.
24
We utilized highly insulin-sensitive mice, due to adipocyte-specific overexpression of the
25
FOXC2 transcription factor, to study mechanisms of the reversed islet adaptation to increased
26
insulin sensitivity. We found that Foxc2TG mice responded to mild hyperglycemia with
27
insulin secretion significantly lower than that of wild type mice, however when severe
28
hyperglycemia was induced, Foxc2TG mice demonstrated insulin secretion equal to or greater
29
than that of wild type mice. In response to autonomic nervous activation by 2-deoxyglucose,
30
the acute suppression of insulin seen in wild-type mice was absent in Foxc2TG mice
31
suggesting impaired sympathetic signaling to the islet. Basal glucagon was increased in
32
Foxc2TG mice but they displayed severely impaired glucagon responses to cholinergic and
33
autonomic nervous stimuli. These data suggest that the autonomic nerves contribute to the
34
islet adaptation to high insulin sensitivity which is compatible with a neuro-adipo regulation
35
of islet function being instrumental for maintaining glucose regulation.
36
Keywords: Insulin, Glucagon, White adipose tissue, Sympathetic nervous system,
37
Parasympathetic nervous system
38
INTRODUCTION
39
Blood glucose homeostasis is one of the most tightly regulated processes in mammalian
40
physiology. In healthy humans, glucose levels vary by only a few percent throughout a 24
41
hour period (40). During feeding this tight regulation is mainly mediated by interplay between
42
insulin levels and the induction of glucose uptake in insulin-sensitive tissues (9). In contrast,
43
during fasting, glucose regulation is mediated primarily by glucagon induced release of stored
2
44
glucose from the liver (21). In overweight and obese individuals the sensitivity of the liver,
45
muscle and adipose tissue to insulin decreases, however the pancreatic beta cells adapt by
46
secreting more insulin to maintain normal glucose uptake in these tissues and maintain normal
47
blood glucose (41). Consequently, the loss of the ability of the beta cells to adapt to systemic
48
insulin resistance is a primary defect causing type 2 diabetes (22). As tight glucose regulation
49
is such a fundamental factor for normal energy homeostasis, it would stand to reason that an
50
adaptation to high degrees of insulin sensitivity is required to continue to maintain normal
51
blood glucose levels. This would also imply a consequent inverse relationship between insulin
52
sensitivity and insulin secretion such that higher insulin sensitivity would be expected to
53
result in lower insulin secretion (3, 11, 22). This was indeed found in a study of elite athletes
54
with high degrees of insulin sensitivity showing that basal and stimulated levels of insulin
55
were decreased compared to sedentary individuals with lower insulin sensitivity when blood
56
glucose levels were equal (6). Similarly, the improved insulin sensitivity induced by weight
57
reduction after bariatric surgery is associated with reduced insulin secretion (18). The
58
mechanistic basis for this “reverse adaptation” to insulin sensitivity has not been
59
characterized.
60 61
In recent years there has been an interest in the enhancement of insulin sensitivity by
62
transitioning insulin-resistant white adipose tissue into energy-dissipating brown adipose
63
tissue, so called “browning” of white adipose tissue (8). In rodents and humans, the
64
transcription factor FOXC2 has been shown to be one of the key mediators of browning of
65
white adipose tissue (15, 30). Mice with an adipose tissue-specific overexpression of the
66
Foxc2 transcription factor show browning of white adipose tissue, with increased expression
67
of brown fat associated genes UCP1, B3AR and others as well as increased oxygen
68
consumption in mitochondria and multilocular lipid droplets in white adipocytes (15). The
69
adipose tissue-specific Foxc2 transgenic mouse is lean and resistant to high fat diet-induced 3
70
obesity (28). However, whether and by which mechanism the enhanced insulin sensitivity
71
affects islet function has not been defined in this model. We have therefore undertaken an in
72
depth study to understand the cellular and physiological mechanisms that drive adaptation to
73
increased insulin sensitivity to maintain normal glucose regulation, using multiple metabolic
74
challenges in this unique model. We found that the reverse adaptation to insulin sensitivity is
75
primarily mediated by altered neural signaling to the pancreatic islet such that enhanced
76
systemic insulin sensitivity causes increased sympathetic nervous input to the islet to suppress
77
insulin, enhance basal glucagon secretion and thus maintain normal glucose levels and avoid
78
hypoglycemia.
79
METHODS
80
Animals. The generation of transgenic mice overexpressing the Foxc2 gene in adipose tissue
81
under the direction of the AP2 promoter has been described previously (15). The AP2 gene
82
promoter is specific for mature adipocytes and the transgene is active from the time of
83
adipocyte maturity. Transgenic mice heterozygous for the Foxc2 transgene, on a C57BL6/J
84
background, were backcrossed for multiple generations onto the C57BL6/JBomTac
85
background. 4-5 month old female mice were used for all experiments. We have previously
86
shown that the glucagon response to 2-D-deoxyglucose (2-DG) is more pronounced in female
87
than in male mice (27). Ethical approval for all animal experiments was received from the
88
regional ethical committee in Lund Sweden.
89
Euglycemic-hyperinsulinemic clamp studies. For euglycemic-hyperinsulinemic clamp
90
experiments, mice were anesthetized with a single i.p injection of fentanyl (0.78 mg/kg) –
91
fluanisone (25 mg/kg) [Hypnorm Vetpharma] and midazolam (12.5 mg/kg) and rested for 30
92
minutes prior to receiving a primed continuous infusion of insulin (30mU/kg/mouse; Actrapid
93
Novo Nordisk A/S, Bagsvaerd, Denmark) and variable glucose (40% D-glucose in saline) for
4
94
90 minutes. Blood glucose was monitored every five minutes with a hand-held glucose
95
monitor (HemoCue, Ängelholm Sweden). Plasma samples were taken at time 0, 60 and 90
96
minutes for the measurement of insulin.
97
Hyperglycemic clamp studies. Anesthetized mice were given a single intravenous injection of
98
D-glucose in the tail vein to elevate blood glucose to the target level and then infused with
99
variable glucose (30-40% D-glucose in saline) to maintain blood glucose at the target level for
100
90 minutes. Plasma samples were taken at 0, 1, 5, 20, 50, 70 and 90 minutes for measurement
101
of insulin.
102
Intravenous glucose, 2-deoxyglucose, arginine and carbachol challenges. Anesthetized
103
mice were fasted for five hours and then given a single intravenous injection of D-glucose
104
(0.35 mg/kg or 1.0 mg/kg), 2-D-deoxyglucose (50 mg/ml), L-arginine (5 mg/mouse) or the
105
muscarinic receptor agonist carbachol (30µg/kg) in the tail vein. Blood was sampled from the
106
intraorbital retrobulbar sinus plexus at 0, 1, 5, 10, 20, 30 and 50 minutes for glucose; 0, 2, 10
107
and 30 minutes for 2-deoxyglucose and 0, 1, 5, 10, 20 and 50 minutes for arginine and
108
carbachol injections. Plasma was separated by centrifugation and glucose, insulin and
109
glucagon were measured.
110
Ex vivo insulin and glucagon secretion from isolated islets. Pancreata were perfused in situ
111
with Hank’s Balanced Salt Solution (HBSS) containing 2.7U Collagenase P (Roche,
112
Mannheim, Germany), removed and incubated at 37° C for 19 minutes. Digested pancreata
113
were washed with HBSS and islets were hand-picked in an inverted microscope. Freshly
114
picked islets were pre-incubated in HEPES buffer containing 125 mmol/l MgCl2, 25 mmol/l
115
HEPES (pH 7.4), 5.6 mmol/l glucose and 0.1% fatty acid free BSA (Boehringer Mannheim,
116
GmbH, Germany). Islets were then incubated in batches of three per well in eight replicates
117
and incubated with varying glucose concentrations with or without parasympathetic
5
118
(carbachol) or sympathetic (clonidine) neurotransmitter analogues (Sigma-Aldrich, St. Louis,
119
MO, USA) for 60 minutes. Media samples are taken at 60 minutes for measurement of insulin
120
and glucagon.
121
Immunohistochemistry. Pancreata were extracted and fixed in 4% paraformaldehyde in 0.1
122
mmol/L PBS containing 0.1% picric acid, prior to cryopreservation in 20% (w/v) sucrose in
123
PBS. Pancreata were frozen in OCT compound and 10 µm sections were cut and mounted.
124
Sections were immunostained using antiserum against rat proinsulin (1:1280 dilution), and
125
rabbit glicentin for glucagon detection (1:10560 dilution) [both from Eurodiagnostica,
126
Malmö, Sweden]. Glucose transporter 2 (GLUT2) was detected using rabbit anti-rat GLUT2
127
antiserum (1:200 dilution) [Chemicon International, Inc., Temecula, CA, USA]. Islet areas
128
were quantified as described previously (36, 37). Briefly, Alexa-fluor coupled fluorescent
129
secondary antibodies (goat anti-guinea pig Alexa546 for insulin and goat anti-rabbit Alexa488
130
for glucagon) were used in double staining experiments. Fluorescent images were captured in
131
an Olympus BX 51 fluorescent microscope with an Olympus DP-50 digital camera. Islet size
132
was determined manually by interactively defining outline of each individual islet and the
133
area covered was determined in the program NIH image 1.06 after calibration. A total of five
134
to seven sections from each pancreas were analyzed.
135
Biochemical analyses. Plasma glucose was determined by the glucose oxidase method with
136
ABTS as the substrate. Plasma insulin was determined by radioimmunoassay (Linco, St.
137
Louis, MO, USA). Plasma glucagon was determined by radioimmunoassay (Linco).
138
Statistical Analyses. All data are presented as mean ± SEM. The acute insulin and glucagon
139
responses are calculated as the mean of the 1 and 5 minute values, after baseline subtraction,
140
during an IVGTT or hyperinsulinemic clamp, whereas ∆AIR is the baseline subtracted 1
141
minute value only. For 2-deoxyglucose challenges the acute insulin and glucagon responses
6
142
were calculated as the baseline-subtracted mean 2 and 10 minute values, whereas ∆AIR was
143
the baseline subtracted 2 minute value only. Area under the curve was calculated using the
144
trapezoidal rule. Statistical significance was determined by the non-parametric Mann Whitney
145
U-test, unless otherwise stated. Statistical calculations were made using the GraphPad Prism
146
6.0 software package.
147
RESULTS
148
Foxc2 transgenic mice have lower basal glucose and insulin and are highly insulin
149
sensitive. Body weight and basal plasma parameters were obtained to assess the basal
150
metabolic state of Foxc2 transgenic mice. Foxc2 transgenic mice had a slightly higher body
151
weight compared to wild type mice (Figure 1A). Four hour and 16 hour fasting plasma
152
glucose were significantly lower in Foxc2 transgenic mice (Figure 1B) as was four hour
153
fasting insulin (Figure 1C) suggesting an increased insulin sensitivity. There was, however,
154
no significant difference in plasma insulin after an overnight fast. To assess whole-body
155
insulin sensitivity, euglycemic-hyperinsulinemic clamps were performed. Blood glucose was
156
clamped at 6.5 mmol/l under a 30 minute steady state period (Figure 1D). During the steady
157
state period, the glucose infusion rate needed to maintain euglycemia, a direct measure of
158
insulin sensitivity, was 3-fold higher in Foxc2 transgenic mice compared to wild type mice
159
(Figure 1E, F).
160
Foxc2 transgenic mice respond to high insulin sensitivity with a glycemia dependent
161
reduction in insulin secretion in vivo. In order to determine the effect of the notable increase
162
in insulin sensitivity on insulin secretion, Foxc2 transgenic mice were subjected to
163
hyperglycemic clamp experiments at three different levels of hyperglycemia. The acute
164
insulin response to glucose during the hyperglycemic clamps showed a distinct pattern that
165
was dependent on the degree of hyperglycemia. When clamped at 11.1 mmol/l glucose, the
7
166
acute insulin response was significantly lower in Foxc2 transgenic mice and the second phase
167
of insulin secretion was nearly absent, suggesting adaptation to increased insulin sensitivity
168
with decreased insulin secretion (Figure 2A, D, G). When clamped at 16.7mmol/l glucose,
169
approximately 2-fold higher than normal nonfasted blood glucose, there was no difference in
170
the acute insulin response or second phase insulin secretion between the wild type and Foxc2
171
transgenic mice (Figure 2B, E, H). When clamped at 33 mmol/l glucose, to establish maximal
172
insulin secretory responses to glucose in vivo, Foxc2 transgenic mice exhibited significantly
173
greater acute insulin responses than wild type controls, with no differences in second phase
174
insulin apparent (Figure 2C, F, I).
175
Foxc2 transgenic mice have normal insulin responses to arginine and cholinergic agonists
176
but impaired response to autonomic nervous activation by 2-deoxyglucose. In addition to
177
glucose we also determined the insulin secretory responses to non-glucose secretagogues with
178
different mechanisms of action. It was noted that both intravenous arginine and carbachol had
179
demonstrable and opposing effects on plasma glucose in Foxc2 transgenic mice and wild type
180
mice (Figure 3A, B). Acute intravenous arginine was associated with significantly greater
181
increases in the plasma glucose excursion in Foxc2 transgenic mice compared to wild-type
182
(Figure 3A). In contrast, stimulation with carbachol or 2-dexoyglucose resulted in
183
significantly lower glucose excursions in Foxc2 transgenic mice than in wild type mice
184
(Figure 3B, C). Insulin secretion in response to acute intravenous injection of arginine was
185
similar between Foxc2 transgenic and wild type mice (Figure 3D, G). We next investigated
186
the effect of the cholinergic muscarine receptor agonist carbachol on insulin secretion in
187
Foxc2 transgenic mice. There were no differences in insulin secretion between Foxc2
188
transgenic and wild type mice in response to intravenous injection of carbachol (Figure 3E,
189
H).
8
190
2-Deoxyglucose (2-DG) causes activation of the autonomic nervous system and signaling
191
through sympathetic and parasympathetic nerve terminals of the endocrine pancreas as well as
192
secretion of epinephrine and norepinephrine from the adrenal glands (13). Immediately after
193
intravenous injection of 2-DG, there was a significant suppression of insulin in wild type mice
194
(Figure 3F), which has been previously attributed to activation of the adrenergic receptors on
195
the beta cell through the sympathetic nervous system (24). After this reduction of insulin,
196
there is a clear increase in insulin secretion, which has been shown to be due to
197
parasympathetic nerve activation (Karlsson and Ahrén, 1991). It was found that the acute
198
suppression of insulin by 2-DG was completely absent in Foxc2 transgenic mice (Figure 3F,
199
I). In contrast, the subsequent insulin secretion, from 2 minutes to 30 minutes, was not
200
different between wild type and Foxc2 transgenic mice (Figure 3F).
201
Foxc2 transgenic mice have enhanced basal glucagon but impaired glucagon secretion in
202
response to cholinergic and autonomic stimuli. As carbachol is a mimetic of acetylcholine
203
and 2-deoxyglucose stimulates both cholinergic and adrenergic nervous activity, the
204
diminished glucose excursions in response to these stimuli suggested impaired glucagon
205
secretion and/or action in response to autonomic nervous activation in Foxc2 transgenic mice.
206
We next investigated glucagon secretion in the basal and stimulated states in Foxc2 transgenic
207
mice. Foxc2 transgenic mice had two-fold greater basal plasma glucagon than wild type mice
208
after a five hour fast (Figure 4A). Interestingly, there was no increase in basal plasma
209
glucagon in response to a prolonged fast in Foxc2 transgenic mice (Figure 4B).
210
We then proceeded to investigate glucagon secretion in response to the non-glucose
211
secretagogues. Despite the higher basal plasma glucagon in Foxc2 transgenic mice (Figure
212
4C), the acute glucagon response to arginine was not significantly different between wild type
213
and Foxc2 transgenic mice (Figure 4F).
9
214
While adrenergic receptor activation in the islet is inhibitory for insulin secretion and
215
stimulatory for glucagon secretion, cholinergic receptor activation is stimulatory to both (1,
216
29). Carbachol activates cholinergic muscarine 3 receptors on alpha cells (17), as well as beta
217
cells, while 2-DG stimulates both sympathetic adrenergic and parasympathetic cholinergic
218
neurotransmission to the islet. Intravenous injection of carbachol resulted in a significantly
219
blunted acute glucagon response in Foxc2 transgenic mice compared to wild type mice,
220
despite having higher basal glucagon (Figure 4D, G). 2-DG administration resulted in
221
significantly impaired acute glucagon responses in Foxc2 transgenic mice (Figure 4E, H), as
222
well as impaired total glucagon secretion (Figure 4E).
223
Foxc2 transgenic mouse islets have normal responses to parasympathetic and sympathetic
224
neurotransmitter analogues ex vivo. As the insulin and glucagon secretory responses to 2-
225
DG were impaired in Foxc2 transgenic mice in vivo, but the insulin secretory response to
226
carbachol was normal, it was unclear as to whether the defect in hormone secretion was due to
227
altered transmission of neural signals to the islet, altered response of the islets to
228
neurotransmitter stimuli or both. If neural signaling to the islet is impaired in vivo, then
229
isolated islets should respond normally to stimulation by neurotransmitters in vitro if post
230
receptor signaling and function is unaltered. We therefore next performed in vitro insulin
231
secretion experiments in islets isolated from wild type and Foxc2 transgenic mice. Similar to
232
the insulin secretion in vivo during hyperglycemic clamps, islets from Foxc2 transgenic mice
233
had slightly, but not significantly, lower insulin secretion in response to 11.1 mmol/l glucose
234
(Fig. 5A). However, more importantly, islets from wild type and Foxc2 transgenic mice
235
displayed similar increases in insulin secretion in response to the parasympathomimetic
236
carbachol (Fig. 5B). Furthermore, glucose-stimulated insulin secretion was suppressed to
237
similar extents by the α2-receptor agonist clonidine in wild type and transgenic mice (Fig.
238
5B). Arginine stimulates insulin secretion independently of any receptor activation and, in the 10
239
presence of arginine, carbachol stimulated and clonidine inhibited insulin secretion to similar
240
extents in both wild type and Foxc2 transgenic mice (Fig. 5B).
241
Foxc2 transgenic mice have smaller islet size compared to wild type mice. Although it was
242
clear that there was low basal and glucose stimulated insulin secretion and impairments in
243
glucagon secretion in response to autonomic neurotransmission, it was unclear whether this
244
was due to alterations in islet morphology. We therefore next investigated islet morphology in
245
Foxc2 transgenic mice. Immunohistochemical analysis of pancreatic sections showed that the
246
core-mantle structure, typical of rodent islets, was unaltered in Foxc2 transgenic mice (Figure
247
6A). However, evaluation of islet areas in wild type and Foxc2 transgenic mice revealed
248
significantly smaller total areas of islets from Foxc2 transgenic mice (Figure 6B). This was
249
associated with a significantly lower insulin positive area of islets of Foxc2 transgenic mice
250
(Figure 6C) but not the glucagon positive area (Figure 6D).
251
DISCUSSION
252
It has been over 30 years since a number of studies demonstrated that insulin resistant
253
subjects have a compensatory increase in insulin secretion (10, 11, 23). Despite much
254
research on this area since then, relatively few studies have demonstrated the logical reverse
255
that increased insulin sensitivity is compensated for by decreased insulin secretion. A few
256
studies have been performed on elite athletes as a model of increased insulin sensitivity, each
257
showing reduced fasting and stimulated insulin secretion in response to oral or intravenous
258
glucose or arginine (6, 35). After bariatric surgery an increase in insulin sensitivity after
259
weight loss is met with a decrease in insulin secretion (18). This inverse adaptation would
260
functionally be a safeguard against increased risk for hypoglycemia in high insulin sensitivity.
261
It is known that a degree of expansion of functional beta cell mass partly explains the
262
mechanism by which insulin resistance is compensated for by increased insulin secretion. A
11
263
mechanism for how the body adapts to increased insulin sensitivity has not been described to
264
date.
265
The findings of this study suggest that the metabolic response to increased systemic insulin
266
sensitivity is a decrease in basal and glucose stimulated insulin secretion at physiological
267
glucose levels. This decreased insulin secretion is therefore appropriate for the ambient level
268
of insulin sensitivity. The islets have smaller insulin positive areas in response to increased
269
insulin sensitivity just as insulin positive islet area increases in response to insulin resistance
270
(7). When challenged with high glucose there was no difference in insulin secretion between
271
wild type and transgenic mice but when challenged with supra-physiological amounts of
272
glucose, there was a greater insulin secretory response in transgenic mice suggesting that
273
there is no functional impairment of the beta cell population resulting from the high level of
274
insulin sensitivity and the beta cells are “rested” therefore primed to give a greater response to
275
an extreme glucose challenge.
276
The regulation of islet hormone secretion by the nervous system involves autonomic nerves,
277
sensory nerves and neuropeptides (1). Sympathetic nervous signaling results in decreased
278
insulin secretion and increased glucagon secretion, as energy in the form of glucose needs to
279
be rapidly mobilized from stores under conditions of acute stress (24). Parasympathetic
280
nervous signaling to the islet occurs via the vagal nerve (5). Parasympathetic vagal
281
stimulation results in the secretion of the major islet hormones insulin, glucagon, somatostatin
282
and pancreatic polypeptide (4, 12). In addition to autonomic signals, there are sensory nerves
283
innervating the islet which affect islet hormone secretion (38). Similar to sympathetic nervous
284
signaling, sensory nervous activation and the release of calcitonin gene-related peptide
285
(CGRP) within the islet is inhibitory to insulin secretion and stimulatory to glucagon secretion
286
(2).
12
287
As tools for investigating autonomic stimuli in vivo, we used the cholinergic agonist
288
carbachol and the autonomic activator 2-DG, which is a non-metabolized glucose analogue
289
which through competition with glucose uptake and intracellular phosphorylation creates
290
intracellular glycopenia (14, 42). Since central nervous system is highly dependent on glucose
291
availability, systematic administration of 2-DG results in neuroglycopenia (31) which
292
activates the autonomic nervous system, as evidenced by increased plasma catecholamine
293
levels and pancreatic sympathetic activation (20). This in turn results in increased glucagon
294
and insulin secretion as we previously have demonstrated in mice (24, 26), and since these
295
responses are inhibited by both muscarinic blockade and alpha-adrenergic blockade (25), it is
296
clear that both the parasympathetic and sympathetic branches of the autonomic nervous
297
system are activated by 2-DG in mice. We observed increased basal glucagon secretion but
298
impaired glucagon secretion in response to cholinergic or autonomic stimuli. There appeared
299
to be a dysregulation between autonomic nervous signaling and the islet, as the suppressive
300
action of adrenergic signals on insulin secretion and the stimulatory effect of adrenergic and
301
cholinergic signals on glucagon secretion, were impaired in Foxc2 transgenic mice. This
302
could be due to an increase in the basal sympathetic activity in response to the extreme insulin
303
sensitivity seen in Foxc2 transgenic mice. Foxc2 transgenic mice had decreased basal insulin
304
and increased basal glucagon which would be expected if sympathetic signaling to the islet
305
were increased. In fact, in response to 2-DG, plasma insulin in wild-type mice was initially
306
suppressed to the same level as that of Foxc2 transgenic mice at baseline, whereas plasma
307
insulin was completely unchanged immediately after 2-DG injection in Foxc2 transgenic
308
mice. This suggests that Foxc2 transgenic mice already had a high level of sympathetic
309
nervous signaling to the islet. When islets from Foxc2 transgenic mice were stimulated with
310
parasympathetic or sympathetic neurotransmitter analogues in vitro, the insulin and glucagon
311
secretory responses were normal. This confirms that altered neural signaling in vivo, not
13
312
altered islet function, is the main contributor to the secretory phenotype of Foxc2 transgenic
313
mice.
314
The finding of altered islet responses to nervous inputs in the lean, highly insulin sensitive
315
Foxc2 transgenic mice should not be a complete surprise. Nearly 30 years ago Peterson and
316
colleagues demonstrated an inverse relationship between body fat and autonomic nervous
317
activity (34). Subsequent studies have demonstrated that caloric restriction and/or exercise
318
result in increased insulin sensitivity in parallel with an increase in autonomic nervous
319
activity (32).
320
Mechanistically, the overexpression of the FOXC2 transcription factor in adipose tissue
321
results in increases in brown adipose tissue mass and white adipose tissue with brown adipose
322
characteristics (15). There are a number of other transgenic mouse models that show
323
increased brown adipose tissue characteristics and energy expenditure and these are also
324
associated with increased insulin sensitivity (16, 19, 33). However the dramatic increase in
325
insulin sensitivity seen in Foxc2 transgenic mice and other models cannot be explained by
326
increased energy dissipation in brown and white adipose tissue alone. Bartness and colleagues
327
have recently described the neural innervation and signaling not only to the brown adipose
328
tissue but also from the brown adipose tissue to centers in the brain (39). The signaling from
329
the brown adipose tissue was mediated via sensory nerves and fed back to the autonomic
330
nervous system at the level of the hindbrain (39). This suggests that the increased mass and
331
metabolic activity of brown adipose tissue of Foxc2 transgenic mice could itself be activating
332
autonomic nervous signaling independent of the insulin sensitivity status. There are some
333
limitations to the study that are worth mentioning. This study was only carried out in female
334
mice, as previous studies have demonstrated a significantly greater islet hormone response to
335
autonomic nervous stimulus by 2-DG in females than in males (27). Furthermore, the high
336
degree of insulin sensitivity in the Foxc2 transgenic mouse model is much less pronounced in 14
337
males (data not shown), making them less suitable for a study on islet adaptation to high
338
degrees of insulin sensitivity. The reasons for this gender difference are not known at present.
339
Another limitation of our study is the lack of definitive information about the absolute beta
340
cell mass in this model of extreme insulin sensitivity.
341
To our knowledge, our study is the first to demonstrate that chronically increased insulin
342
sensitivity affects islet hormone secretion by a mechanism involving alterations to islet
343
morphology and the basal activation state of the branches of the autonomic nervous system.
344
Future studies will need to be performed to determine whether autonomic nervous signaling
345
affects islet hormone secretion in highly insulin sensitive humans.
346
ACKNOWLEDGEMENTS
347
The authors would like to acknowledge Lena Kvist, Lillian Bengtsson, Kristina Andersson
348
and Catharina Blennow of Lund University for their excellent technical assistance.
349
GRANTS
350
This work was supported financially by the Swedish research council, Lund University
351
Medical Faculty and Region Skåne. The authors have nothing to disclose with regards to this
352
study.
353
DISCLOSURES
354
The authors have nothing to disclose with regards to this study.
355
AUTHOR CONTRIBUTIONS
356
B.O. designed and performed experiments, data analysis and wrote the manuscript. S.E.
357
generated and provided the Foxc2 transgenic mouse line and wrote the manuscript. MR
15
358
performed the immunohistochemistry B.A. designed experiments, performed data analysis
359
and wrote the manuscript. All of the authors approved of the final version of the manuscript.
360
REFERENCES
361
1. Ahren B. Autonomic regulation of islet hormone secretion--implications for health and disease.
362
Diabetologia 43: 393-410, 2000.
363
2. Ahren B, Martensson H, and Nobin A. Effects of calcitonin gene-related peptide (CGRP) on islet
364
hormone secretion in the pig. Diabetologia 30: 354-359, 1987.
365
3. Ahren B and Pacini G. Importance of quantifying insulin secretion in relation to insulin sensitivity
366
to accurately assess beta cell function in clinical studies. Eur J Endocrinol 150: 97-104, 2004.
367
4. Ahren B, Paquette TL, and Taborsky GJ, Jr. Effect and mechanism of vagal nerve stimulation
368
on somatostatin secretion in dogs. The American journal of physiology 250: E212-217, 1986.
369
5. Ahren B, Taborsky GJ, Jr., and Porte D, Jr. Neuropeptidergic versus cholinergic and adrenergic
370
regulation of islet hormone secretion. Diabetologia 29: 827-836, 1986.
371
6. Ahren B and Thorsson O. Increased insulin sensitivity is associated with reduced insulin and
372
glucagon secretion and increased insulin clearance in man. J Clin Endocrinol Metab 88: 1264-1270,
373
2003.
374
7. Ahren J, Ahren B, and Wierup N. Increased beta-cell volume in mice fed a high-fat diet: a
375
dynamic study over 12 months. Islets 2: 353-356, 2010.
376
8. Bartelt A and Heeren J. Adipose tissue browning and metabolic health. Nature reviews
377
Endocrinology 10: 24-36, 2014.
378
9. Basu R, Dalla Man C, Campioni M, Basu A, Klee G, Toffolo G, Cobelli C, and Rizza RA.
379
Effects of age and sex on postprandial glucose metabolism: differences in glucose turnover, insulin
380
secretion, insulin action, and hepatic insulin extraction. Diabetes 55: 2001-2014, 2006.
381
10. Beard JC, Ward WK, Halter JB, Wallum BJ, and Porte D, Jr. Relationship of islet function to
382
insulin action in human obesity. J Clin Endocrinol Metab 65: 59-64, 1987.
16
383
11. Bergman RN, Phillips LS, and Cobelli C. Physiologic evaluation of factors controlling glucose
384
tolerance in man: measurement of insulin sensitivity and beta-cell glucose sensitivity from the
385
response to intravenous glucose. J Clin Invest 68: 1456-1467, 1981.
386
12. Bloom SR and Edwards AV. Pancreatic endocrine responses to stimulation of the peripheral ends
387
of the vagus nerves in conscious calves. The Journal of physiology 315: 31-41, 1981.
388
13. Bloom SR, Edwards AV, and Hardy RN. The role of the autonomic nervous system in the
389
control of glucagon, insulin and pancreatic polypeptide release from the pancreas. The Journal of
390
physiology 280: 9-23, 1978.
391
14. Brown J. Effects of 2-deoxyglucose on carbohydrate metablism: review of the literature and
392
studies in the rat. Metabolism 11: 1098-1112, 1962.
393
15. Cederberg A, Gronning LM, Ahren B, Tasken K, Carlsson P, and Enerback S. FOXC2 is a
394
winged helix gene that counteracts obesity, hypertriglyceridemia, and diet-induced insulin resistance.
395
Cell 106: 563-573, 2001.
396
16. Chiang SH, Bazuine M, Lumeng CN, Geletka LM, Mowers J, White NM, Ma JT, Zhou J, Qi
397
N, Westcott D, Delproposto JB, Blackwell TS, Yull FE, and Saltiel AR. The protein kinase
398
IKKepsilon regulates energy balance in obese mice. Cell 138: 961-975, 2009.
399
17. Duttaroy A, Zimliki CL, Gautam D, Cui Y, Mears D, and Wess J. Muscarinic stimulation of
400
pancreatic insulin and glucagon release is abolished in m3 muscarinic acetylcholine receptor-deficient
401
mice. Diabetes 53: 1714-1720, 2004.
402
18. Guldstrand M, Ahren B, and Adamson U. Improved beta-cell function after standardized weight
403
reduction in severely obese subjects. Am J Physiol Endocrinol Metab 284: E557-565, 2003.
404
19. Hansen JB, Jorgensen C, Petersen RK, Hallenborg P, De Matteis R, Boye HA, Petrovic N,
405
Enerback S, Nedergaard J, Cinti S, te Riele H, and Kristiansen K. Retinoblastoma protein
406
functions as a molecular switch determining white versus brown adipocyte differentiation. Proc Natl
407
Acad Sci U S A 101: 4112-4117, 2004.
408
20. Havel PJ, Veith RC, Dunning BE, and Taborsky GJ, Jr. Pancreatic noradrenergic nerves are
409
activated by neuroglucopenia but not by hypotension or hypoxia in the dog. Evidence for stress-
17
410
specific and regionally selective activation of the sympathetic nervous system. J Clin Invest 82: 1538-
411
1545, 1988.
412
21. Hendrick GK, Frizzell RT, Williams PE, and Cherrington AD. Effect of hyperglucagonemia
413
on hepatic glycogenolysis and gluconeogenesis after a prolonged fast. The American journal of
414
physiology 258: E841-849, 1990.
415
22. Kahn SE. The relative contributions of insulin resistance and beta-cell dysfunction to the
416
pathophysiology of Type 2 diabetes. Diabetologia 46: 3-19, 2003.
417
23. Kahn SE, Prigeon RL, McCulloch DK, Boyko EJ, Bergman RN, Schwartz MW, Neifing JL,
418
Ward WK, Beard JC, Palmer JP, and et al. Quantification of the relationship between insulin
419
sensitivity and beta-cell function in human subjects. Evidence for a hyperbolic function. Diabetes 42:
420
1663-1672, 1993.
421
24. Karlsson S and Ahren B. Contribution of adrenergic nerves and the adrenals to 2-deoxy-D-
422
glucose-induced insulin and glucagon secretion in the mouse. International journal of pancreatology :
423
official journal of the International Association of Pancreatology 10: 207-215, 1991.
424
25. Karlsson S and Ahren B. Inhibition of 2-deoxy-glucose-induced glucagon secretion by
425
muscarinic and alpha-adrenoceptor blockade in the mouse. Diabetes research and clinical practice 3:
426
239-242, 1987.
427
26. Karlsson S, Bood M, and Ahren B. The mechanism of 2-deoxy-glucose-induced insulin
428
secretion in the mouse. Journal of autonomic pharmacology 7: 135-144, 1987.
429
27. Karlsson S, Scheurink AJ, and Ahren B. Gender difference in the glucagon response to
430
glucopenic stress in mice. American journal of physiology Regulatory, integrative and comparative
431
physiology 282: R281-288, 2002.
432
28. Kim JK, Kim HJ, Park SY, Cederberg A, Westergren R, Nilsson D, Higashimori T, Cho YR,
433
Liu ZX, Dong J, Cline GW, Enerback S, and Shulman GI. Adipocyte-specific overexpression of
434
FOXC2 prevents diet-induced increases in intramuscular fatty acyl CoA and insulin resistance.
435
Diabetes 54: 1657-1663, 2005.
18
436
29. Kimura H, Katagiri K, Ohno T, Harada N, Imanishi H, Iwasaki M, Ito M, and Takeuchi T.
437
Effect of acetylcholine and new cholinergic derivative on amylase output, insulin, glucagon, and
438
somatostatin secretions from perfused isolated rat pancreas. Horm Metab Res 14: 356-360, 1982.
439
30. Lidell ME, Seifert EL, Westergren R, Heglind M, Gowing A, Sukonina V, Arani Z, Itkonen
440
P, Wallin S, Westberg F, Fernandez-Rodriguez J, Laakso M, Nilsson T, Peng XR, Harper ME,
441
and Enerback S. The adipocyte-expressed forkhead transcription factor Foxc2 regulates metabolism
442
through altered mitochondrial function. Diabetes 60: 427-435, 2011.
443
31. Muller EE, Cocchi D, and Forni A. A central site for the hyperglycemic action of 2-deoxy-d-
444
glucose in mouse and rat. Life sciences Pt 1: Physiology and pharmacology 10: 1057-1067, 1971.
445
32. Nakano Y, Oshima T, Sasaki S, Higashi Y, Ozono R, Takenaka S, Miura F, Hirao H,
446
Matsuura H, Chayama K, and Kambe M. Calorie restriction reduced blood pressure in obesity
447
hypertensives by improvement of autonomic nerve activity and insulin sensitivity. Journal of
448
cardiovascular pharmacology 38 Suppl 1: S69-74, 2001.
449
33. Pan D, Fujimoto M, Lopes A, and Wang YX. Twist-1 is a PPARdelta-inducible, negative-
450
feedback regulator of PGC-1alpha in brown fat metabolism. Cell 137: 73-86, 2009.
451
34. Peterson HR, Rothschild M, Weinberg CR, Fell RD, McLeish KR, and Pfeifer MA. Body fat
452
and the activity of the autonomic nervous system. The New England journal of medicine 318: 1077-
453
1083, 1988.
454
35. Pratley RE, Hagberg JM, Rogus EM, and Goldberg AP. Enhanced insulin sensitivity and lower
455
waist-to-hip ratio in master athletes. The American journal of physiology 268: E484-490, 1995.
456
36. Reimer MK and Ahren B. Altered beta-cell distribution of pdx-1 and GLUT-2 after a short-term
457
challenge with a high-fat diet in C57BL/6J mice. Diabetes 51 Suppl 1: S138-143, 2002.
458
37. Reimer MK, Holst JJ, and Ahren B. Long-term inhibition of dipeptidyl peptidase IV improves
459
glucose tolerance and preserves islet function in mice. Eur J Endocrinol 146: 717-727, 2002.
460
38. Rosenfeld MG, Mermod JJ, Amara SG, Swanson LW, Sawchenko PE, Rivier J, Vale WW,
461
and Evans RM. Production of a novel neuropeptide encoded by the calcitonin gene via tissue-specific
462
RNA processing. Nature 304: 129-135, 1983.
19
463
39. Ryu V, Garretson JT, Liu Y, Vaughan CH, and Bartness TJ. Brown adipose tissue has
464
sympathetic-sensory feedback circuits. The Journal of neuroscience : the official journal of the Society
465
for Neuroscience 35: 2181-2190, 2015.
466
40. Service FJ, O'Brien PC, and Rizza RA. Measurements of glucose control. Diabetes Care 10:
467
225-237, 1987.
468
41. Wagenknecht LE, Langefeld CD, Scherzinger AL, Norris JM, Haffner SM, Saad MF, and
469
Bergman RN. Insulin sensitivity, insulin secretion, and abdominal fat: the Insulin Resistance
470
Atherosclerosis Study (IRAS) Family Study. Diabetes 52: 2490-2496, 2003.
471
42. Wick AN, Drury DR, Nakada HI, and Wolfe JB. Localization of the primary metabolic block
472
produced by 2-deoxyglucose. The Journal of biological chemistry 224: 963-969, 1957.
473 474
FIGURE LEGENDS
475
Figure 1. Bodyweight, fasting glucose and insulin and insulin sensitivity measured by
476
euglycemic-hyperinsulinemic clamp in wild type and Foxc2 transgenic mice
477
(A) Body weight in four month old female wild type (WT) and Foxc2 transgenic (Foxc2TG)
478
(n = 26-29). Mice were fasted for 4 (n = 14-20) or 16 (n = 8-9) hours and plasma glucose (B)
479
and insulin (C) were determined. Euglycemic-hyperinsulinemic clamp experiments were
480
carried out on non-fasted anesthetized mice (n = 5-7 mice per group). (D) Insulin was infused
481
at a constant rate and glucose at a variable rate to obtain a steady state blood glucose value of
482
6.5 mmol/l between 60 and 90 minutes post infusion start. (E, F) The rate of glucose infusion
483
was recorded for all time points and the glucose infusion rate (GIR) was calculated during the
484
30 minute steady state period between 60 and 90 minutes. Data are presented as mean ± SEM.
485
Asterisks indicate the probability level of random difference between the wild type and Foxc2
486
transgenic mice * p < 0.05, ** p < 0.01, *** p < 0.001.
487
Figure 2. Dynamic insulin secretion in vivo determined by hyperglycemic clamp 20
488
Plasma glucose (A, B, C) and insulin (D, E, F) in mice fasted five hours and given a bolus
489
injection of D-glucose followed by variable infusion of glucose to reach target blood glucose
490
of 11.1 mmol/l (A, D, G), 16.7 mmol/l (B, E, H) or 33.3 mmol/l (C, F, I) and which was
491
maintained at a steady state for 90 minutes. The acute insulin response to glucose (∆AIR) was
492
calculated as described in Experimental procedures. n = 7-11 per group. Data are presented as
493
mean ± SEM. Asterisks indicate the probability level of random difference between the wild
494
type and Foxc2 transgenic mice * p < 0.05, ** p < 0.01, *** p < 0.001.
495
Figure 3. Plasma glucose and insulin in response to non-glucose secretagogues
496
Plasma glucose and insulin in mice fasted five hours and given single intravenous injections
497
of L-Arginine (A, D, G) (n =4), Carbachol (B, E, H) (n = 6-10) or 2-deoxy D-glucose (C, F, I)
498
(n = 11-12). The acute insulin responses (∆AIR) (G, H, I) to each secretagogue were
499
calculated as described in Experimental procedures. Data are presented as mean ± SEM.
500
Asterisks indicate the probability level of random difference between the wild type and Foxc2
501
transgenic mice * p < 0.05, ** p < 0.01, *** p < 0.001.
502
Figure 4. Plasma glucagon after fasting or injection of non-glucose secretagogues
503
Plasma glucagon in mice fasted for 4 (A) or 16 (B) hours (n = 11-14 per group). Plasma
504
glucagon in mice fasted for five hours and given single intravenous injections of (C) L-
505
Arginine (n = 4), (D) Carbachol (n = 6) or (E) 2-deoxy D-glucose (n = 11). The acute
506
glucagon responses to each secretagogue were calculated as described in Experimental
507
procedures. Data are presented as mean ± SEM. Asterisks indicate the probability level of
508
random difference between the wild type and Foxc2 transgenic mice * p < 0.05, ** p < 0.01,
509
*** p < 0.001.
510
Figure 5. Insulin and glucagon secretion in isolated islets from wild type and Foxc2
511
transgenic mice 21
512
(A) Insulin secretion in response to varying glucose concentrations. (B) The fold-change in
513
insulin secretion in response to neural (carbachol 100 µM, clonidine 1 µM) and non-neural
514
(arginine 10 mM) factors in the presence of 11.1 mmol/l glucose. (C) Glucagon secretion in
515
response to glucose alone and in the presence of neural and non-neural factors. (D) The fold-
516
change in glucagon secretion in response to neural and non-neural factors in the presence of
517
11.1 mmol/l glucose. n = 3 mice per genotype.
518
Figure 6. Islet morphology and cellular composition in wild-type and Foxc2 transgenic
519
mice
520
(A) Islet fluorescent immunohistochemistry from Foxc2 transgenic (TG) and wild-type (WT)
521
mice fed normal kilocaloric diets (NK). Top panel- staining with antibodies against GLUT2
522
(green). Bottom panel- staining with antibodies against insulin (red) and glucagon (green).
523
Scale bar represents 50 µm. (B) Total islet area (C) insulin positive islet area (D) glucagon
524
positive islet area. n = 4 mice per group. Data are presented as mean ± SEM. Asterisks
525
indicate the probability level of random difference between the wild type and Foxc2
526
transgenic mice * p < 0.05, ** p < 0.01
527
22
10
WT
***
6 4 2
4
15
WT Foxc2 TG
10
5
0 0
20
40
60
Time (min)
80
100
250 200
WT Foxc2 TG
**
150 100 50 0
16
E Glucose infusion rate (mg/kg/min)
D Blood Glucose (mmol/l)
*
8
0
Foxc2 TG
WT Foxc2 TG
Insulin (pmol/l)
20
0
C
10
Glucose (mmol/l)
30
Body weight (g)
B
**
4
16
F 40
WT Foxc2 TG
30 20 10 0 0
20
40
60
Time (min)
80
100
**
40
GIR (mg/min/kg)
A
30 20 10 0
WT
Foxc2 TG
B 40
WT Foxc2 TG
30
Glucose (mmol/l)
20
10 -20
0
20
40
60
80
30
20
10
100 120 140
-20
Time (min)
20
40
60
80
** *
**
500
WT Foxc2 TG
40
60
80
1000 500
100
120
20
40
60
80
100
120
1000
140
WT
Foxc2 TG
20
40
I
1000
500
0
80
100
120
1500 1000 500 0
WT
60
*
2000
AIR (pmol/l)
AIR (pmol/l)
AIR (pmol/l)
200 0
500
0
800
*
100 120 140
1500
1500
400
80
Time (min)
H
1000
60
WT Foxc2 TG
Time (min)
G
40
0 0
140
Time (min)
600
20
2000
0
20
0
F
1500
0 0
10
Time (min)
2000
Insulin (pmol/l)
WT Foxc2 TG
1000
20
0 -20
100 120 140
E
2000
Insulin (pmol/l)
0
WT Foxc2 TG
30
Time (min)
D 1500
40
WT Foxc2 TG
Insulin (pmol/l)
Glucose (mmol/l)
40
C Glucose (mmol/l)
A
Foxc2 TG
WT
Foxc2 TG
140
WT Foxc2 TG
9 8 7 6 5 -10
0
10
20
30
40
50
20
WT Foxc2 TG
15
10
5 -10
60
***
** ** 0
10
Time (min)
40
50
800 600 400 200 0 -10
0
*
*
** 10
20
30
40
50
10
WT Foxc2 TG
2000
1000
*** 0 -10
60
0
10
20
30
40
50
WT Foxc2 TG
600 400
**
200 0 -10
60
0
I
200
200
Foxc2 TG
10
20
*** 100
AIR (pmol/l)
800
AIR (pmol/l)
250
400
30
800
Time (min)
H
1000
600
20
1000
Time (min)
G
WT
0
F
Time (min)
0
***
10
Time (min)
3000
Insulin (pmol/l)
WT Foxc2 TG
WT Foxc2 TG
***
15
5 -10
60
E
1000
Insulin (pmol/l)
30
20
Time (min)
D
AIR (pmol/l)
20
*
***
25
Insulin (pmol/l)
Glucose (mmol/l)
10
C Glucose (mmol/l)
B Glucose (mmol/l)
A
150
0
-100
100
-200
50 0
WT
Foxc2 TG
-300
WT
Foxc2 TG
30
B Glucagon 16 hr fast (pg/mL)
40 20 0
WT
Foxc2 TG
C Glucagon pg/mL
WT Foxc2 TG
* *
100
0 -10
**
* 0
10
20
30
40
50
Time (min)
F
20 0
Foxc2 TG
E WT Foxc2 TG
150
*
100 50 0 -10
60
** 0
10
G 50
40
40
30 20 10
20
30
40
50
Time (min)
50
0
WT
200
AGR (pg/mL)
AGR (pg/mL)
40
D
150
50
60
Foxc2 TG
*
150 100 50 0 -10
60
* 0
10
20
Time (min) 100
**
30 20
0
*
200
H
10
WT
250
Glucagon pg/mL
60
80
80
AGR (pg/mL)
***
Glucagon pg/mL
Glucagon 5 hr fast (pg/mL)
A 80
*
60 40 20
WT
Foxc2 TG
0
WT
Foxc2 TG
30
WT Foxc2 TG
2.8 11.1
10
8
WT Foxc2 TG
6
4
2
0
C A A rg rg in in in in e+ e C ar A ba rg in ch in ol e+ C lo ni di ne
C ar ba ch ol C lo ni di ne
Glucose (mmol/l) Insulin Secretion (fold change)
100
A A rg rg in in in in e+ e C ar A ba rg in ch in ol e+ C lo ni di ne
200
WT Foxc2 TG
Glucagon Secretion (fold change)
300
C ar ba ch ol C lo ni di ne
Insulin Secretion (pg/hr/islet) 0
C ar ba ch ol C lo ni di ne A rg A in rg in in e+ in e C A ar rg ba in in ch e+ ol C lo ni di ne
Glucagon Secretion (pg/hr/islet)
A B
25
20
WT Foxc2 TG
15
10 5
0
6
4
WT Foxc2 TG
2
0
A
C
Total Islet Area (m2)
20000
*
15000 10000 5000 0
WT
Foxc2 TG
WT
Foxc2 TG
Glucagon Positive Area (m2)
D 20000 15000 10000 5000 0
Insulin Positive Area (m2)
B
20000 15000
*
10000 5000 0
WT
Foxc2 TG