PhD Thesis

FACULTY OF SCIENCE
UNIVERSITY OF COPENHAGEN
PhD Thesis
Kamil Borkowski
The interplay between cyclic AMP and insulin during
obesity development
Academic advisers: Karsten Kristiansen and Lise Madsen
Submitted: 28/04/2013
Dissertation for the degree of Doctor of Philosophy
Copenhagen University, Department of Biology
Copenhagen, Denmark
Author: Kamil Borkowski
Title: The interplay between cyclic AMP and insulin during obesity development
Academic advisors:
Karsten Kristiansen, Department of Biology, Copenhagen University, Copenhagen, Denmark
Lise Madsen, National Institute of Nutrition and Seafood Research, Bergen, Norway
Submitted: April 2013
1
Acknowledgments
I would like to express my gratitude to both of my supervisors, Karsten Krisiansen and Lise
Madsen, for their continued support and knowledge throughout the dissertation process.
Special thanks to Alison Keenan, for constant support and motivation.
I would also like to thank all my colleges in Denmark, Norway and USA for help in
conducting experiments.
This work was supported by grants from the Danish Natural Research Council, The Novo
Nordisk Foundation, the Carlsberg Foundation and the Danish council for Strategic Research,
Grant No.: 09-065171.
2
1. List of Abbreviations
007
ATF6
BIP
BMI
8-pCPT-2´-O-Me-cAMP
activating transcription factor-6
glucose-regulated protein 78
Body mass index
BMP7
C/EBPα
Bone-morphogenetic protein 7
CCAAT/enhancer binding protein α
CAP
CREB
Cbl- associated protein
cAMP responding element binding
EPAC
GR
exchange factor directly activated by cAMP
glucocorticoid receptor
HSL
IBMX
Hormone sensitive lipase
isobutylmethylxanthine
IGF
IRE1a
IRS
JNK1
MB
MEFs
PDEB3
PERK
PI3 K
PIP3
PKA
PKC
PPARy
pref-1
ROCK
SOS
TNF-α
insulin like growth factor
inositol requiring enzyme-1
Insulin receptor ligand
Jun-N-terminal kinase 1
N6-monobutyrylcAMP
mouse embryo fibroblasts
phosphodiesterase 3B
PKR-like ER kinase
phosphatidylinositol 3 kinase
trisphosphorylated inositol
Protein kinase A
protein kinase C
proliferator-activated receptor γ
preadipocyte factor-1
Rho associated kinase
Son-of-sevenless
tumour necrosis factor α
3
Contents
Acknowledgments................................................................................................................................... 2
1. List of Abbreviations ........................................................................................................................... 3
2. Abstract ............................................................................................................................................... 5
3. Introduction ........................................................................................................................................ 6
3.1. Purpose of review ........................................................................................................................ 6
3.2. Obesity as a worldwide problem ................................................................................................. 6
3.3. Adipogenesis and obesity ............................................................................................................ 6
3.4. Insulin ........................................................................................................................................... 8
3.4.1. Insulin and insulin-like growth factor receptor signalling cascades ..................................... 8
3.4.2. Defects in insulin signalling ................................................................................................. 12
3.5. Cyclic AMP .................................................................................................................................. 18
3.5.1. Epac ..................................................................................................................................... 19
3.5.2. cAMP action in adipocytes .................................................................................................. 20
3.6. The role of insulin and cAMP in adipogenesis ........................................................................... 22
3.6.1. Transcription factors of adipogenesis ................................................................................. 22
4. The aim of the study ......................................................................................................................... 25
5. List of manuscripts: ........................................................................................................................... 26
6. Discussion of results.......................................................................................................................... 27
6.1. The role of cAMP in the adipocyte differentiation .................................................................... 27
6.2. The role of insulin in mediating the negative effect of sucrose in high fat diet ........................ 28
6.3. Combine action of insulin and cAMP accelerate adipocyte differentiation in vivo................... 29
6.4. Conclusions ................................................................................................................................ 30
7. Literature .......................................................................................................................................... 31
8. Annex ................................................................................................................................................ 37
4
2. Abstract
Insulin and cAMP signalling are related to two opposite metabolic responses. Insulin
secretion is elicited in response to food availability and trigger catabolic processes like
lipogenesis and glycogen synthesis with a purpose of energy storage. On the other hand
cAMP signalling is associated with stress and starvation and stimulates processes like
glycogen degradation and lipolysis to distribute the stored energy through the body.
Although in energy preserving tissues insulin inhibit cAMP signalling, in fat precursor cells
cAMP potentiates insulin action and promote adipogenesis. Activation of exchange factor
directly activated by cAMP (Epac) has been shown to be crucial for cAMP mediated
potentiation of insulin signaling. In the current study I am trying to answer the question as
to how cAMP accelerates adipogenesis in vivo as well as what is the role of Epac in cAMP
mediated potentiation of insulin signalling. Moreover, I am investigating how increased
insulin secretion caused by sucrose consumption, affects insulin signaling in peripheral
tissues.
5
3. Introduction
3.1. Purpose of review
The following review summarises current information regarding the metabolic and
molecular actions of insulin and cyclic AMP and their role in adipocyte metabolism
and differentiation.
3.2. Obesity as a worldwide problem
According to the World Health Organisation, “overweight and obesity are defined as
abnormal or excessive fat accumulation that presents a risk to health”. Body mass index
(BMI) is a commonly used indicator of obesity, which describes the relationship between
height and body mass of an individual. A person with BMI higher than 25 is considered
overweight, whereas a person with a BMI higher than 30 is considered obese. According to
the data from 2008, more than 1.4 billion adults above the age of 20 are overweight and
more than 500 million obese, which is almost 10% of the world population. The prevalence
of obesity differs by gender. There are 50% more obese women than men, and the obesity
ratio is usually much higher in Westernized countries. In 2008, countries like USA, Kuwait,
Czech Republic, and United Arab Emirates reached more than a 30% obesity ratio and are
placed in top 20 the most obese countries. On the other hand countries like Ethiopia, Nepal
and India are on the bottom of the list with the obesity ratio smaller than 2 [1]. Obesity is a
global health concern, as it increases the risk of several diseases, like type 2 diabetes,
coronary heart disease and stroke. Additionally, obesity has been related to gallbladder
disease, musculoskeletal disorders, respiratory problems and several types of cancers like
colon, prostate or pancreas cancers [2, 3].
3.3. Adipogenesis and obesity
There are several hypotheses surrounded the mechanisms underlying obesity and the
generation of new fat cells. Recently, hypertrophic and hyperplasic type of obesity has been
defined. In hypertrophic type of obesity, fat organ mass is increased due to enlargement of
the fat cell size. In hyperplasic type of obesity, fat organ mass is increased due to the
6
increase in fat cells number [4]. Obesity caused by hyperplasia reminds still an unexplored
area of research. A substantial amount of published data supports the hypothesis that high
energy diets can promote the development of new fat cells [5]. There is significant amount
of human studies showing very significant correlations between severity of obesity and fat
cell hyperplasia, as well as negative correlations between adipocyte size and adipocyte
number in obese patients [6-8]. The relationship between adipocyte size and number to
obesity state is illustrated in Figure 1. However, the physiology behind the origins of
development of adipose tissue, as well as new fat cell development throughout human life,
is not fully understood. A recent study demonstrated that the number of fat cells in humans
increases during childhood and adolescence (until approximately 20th year of life) and
reminds constant during adulthood in obese as well as in lean subjects [8]. A similar
situation has been observed in rats where high caloric diets can induce hyperplasia of fat
cells when rats are subjected to it in the early time of life. A high caloric diet had no effect
on adipose cell number in adult rats and caused only hypertrophic type of obesity [9].
Another human study demonstrated a strong negative correlation between fat tissue
hyperplasia and the age of onset of obesity [6, 10]. On the other hand, the same studies
demonstrated that even obesity developed during adulthood can be hyperplasic. One of the
arguments for adulthood adipogenesis can be the fact of limited size of the fat cells. The
relationship of the obesity severity and the fat cell size is a saturation curve [6]. This means
that at some point during weight gain, the adipose cell size reaches a plateau and the only
reasonable possibility of further increase of the fat depot mass is by increase of the cell
number.
Figure 1. The increase of fat cell size (A) and cell number (B) in relation to obesity ratio. Figure taken from [6].
7
3.4. Insulin
Insulin, together with the glucagon, is the key factor regulating the blood levels of glucose,
metabolism, energy expenditure and adipogenesis [11, 12]. It coordinates processes related
to glucose uptake, lipolysis and gluconeogenesis. Insulin levels increase very shortly after a
meal, due to the increase of glucose and free amino acids [13]. Skeletal muscle followed by
adipose tissue is the primary organ responsible for blood glucose clearance after the meal
followed by liver and adipose tissue [14]. Insulin stimulates glucose uptake through the
glucose transporter, GLUT4, in contrary to GLUT-1, insulin independent basal glucose
transport [15]. GLUT4 is abundant in skeletal muscle and adipose tissue whereas liver
expresses GLUT2, which is insulin insensitive [16]. Upon insulin stimulation, intracellular
vesicles containing GLUT4 fuse with the plasma membrane using the SNARE (soluble NSF
attachment protein receptor (where NSF is N-ethylmaleimide-sensitive fusion protein)
complex [17]. Apart from glucose uptake, insulin can increase synthesis of glycogen and
certain metabolic enzymes, as well as enhance DNA and RNA synthesis [18]. Insulin has
been reported to regulate the expression of more than 150 genes [19]. Among them is
glucose 6 phosphatase in the liver or glucagon in the alpha cells of pancreatic islets. In
adipose tissue and muscle, insulin stimulates the expression of hexokinase 2, glycogen
synthase, GLUT4, Insulin receptor ligands (IRS) 1 and p85 alpha regulatory subunit of
phosphatidylinositol 3 kinase (PI3K) [20]. Insulin has been shown to regulate multiple
processes related to homeostasis and acts on many tissues. Recently its action in the central
nervous system attracted attention. Insulin stimulation in the central nervous system
promotes satiety and decreases food intake. This action is thought to be mediated by
arcuate nucleuses of hypothalamus, where the insulin receptor is highly expressed. Here,
insulin has been demonstrated to act via stimulation of expression of causing loss of
appetite neuropeptides like proopiomelanocortin, which is a precursor of a-melanocyte–
stimulating hormone, and cocaine and amphetamine regulated transcript, as well as with
the orexigenic neuropeptide Y and the agouti-related peptide [21]. A recent study
highlighted a possible role for insulin in the promotion of an obese phenotype based on
carbohydrate quality in a high fat diet. Nifedipine, an insulin secretion inhibitor, was able to
protect mice from diet induced obesity from a high fat high sucrose diet [22]. Moreover,
Ins1+/-:Ins2-/- mice which display reduced insulin secretion and are protected against chronic
hyperinsulinemia, are also protected against high fat diet induced obesity [23]. Moreover,
elevated plasma insulin is correlated with insulin resistance in humans [24]
3.4.1. Insulin and insulin-like growth factor receptor signalling cascades
Insulin consists of two short peptides bound together by two disulphate bridges and binds
to the insulin receptor as its main target. The insulin receptor is necessary to maintain
insulin action. Both humans and mice lacking insulin receptor are borne alive, but cannot
survive, which suggests a critical role of insulin action during early development, but not for
8
foetal metabolism [25]. The insulin receptor is a heterotetrameric membrane glycoprotein,
consisting of two α and two ß subunits, bound together with disulfate bridges [26]. It
belongs to the receptor tyrosine kinase family. Alpha subunits create an extracellular part of
the receptor and binds insulin, whereas beta subunits create transmembrane and the
intracellular part of the receptor. From the many of phosphorylated tyrosines on the
intracellular part of the insulin receptor, the juxtamembrane autophosphorylation site plays
a critical role in interaction between the insulin receptor and the intracellular substrates
[26]. The insulin receptor can also bind insulin like growth factors (IGFs) andpeptide
hormones that have structures similar to insulin. The affinity of binding of IGF is
approximately 100 to 1,000 fold lower than for insulin; however, the physiological
concentration of IGF is approximately 100 folds higher than insulin. There are two splicing
variants of alpha subunit of insulin receptor A and B. Variant A lacks an amino acids
fragment of 12 residues on the C terminal [27]. There is no noticeable difference in affinity
for insulin between the two isoforms, however, variant A has higher affinity towards IGF-2.
Insulin can also bind to IGF receptors, which also belongs to receptor tyrosine kinase family
[28]. In pancreatic beta islets, the action of insulin is conducted by IGF receptor 1 which
require higher insulin concentration than insulin receptor [29]. Insulin receptors are
expressed at high levels predominantly in white and brown adipocytes and skeletal muscle
and heart [26].
Insulin receptor ligands are scaffolding proteins that bind to phosphorylated insulin receptor
or IGF receptors. IRS consists of a placstrin homolog domain, a phosphotyrosine binding
domain at the N terminal and a large C terminal domain. After phosphorylation by insulin or
IGF receptor, the large C terminal can bind multiple signalling proteins leading to activation
of multiple signalling cascades [30]. All IRS domains have several phosphorylation sites,
which can be targeted by multiple kinases, other than insulin or IGF. Those phosphorylations
can affect IRS function by promotion of its degradation or prevention of binding to the
receptor [31]. There are 6 IRS isomers, named IRS1 to IRS6. All isomers have different
functions as well as different tissue specificity. The IRS-1 deficient mice show peripheral
insulin resistance and some growth aberrations. In adipose tissue and skeletal muscles IRS-1
is the primary conductor of insulin receptor signalling, and is crucial for glucose transport,
glycogen, lipid and protein synthesis, mitogenesis and gene expression [32]. IRS-1 can
directly bind the regulatory subunit of PI3K, as well as protein kinase B and C and mitogen
activated protein kinase. IRS-2 deficient mice are insulin intolerant and have impaired ß
islets growth. IRS-2 is also associated with type 2 diabetes. IRS-3 has been shown to conduct
insulin signalling in mature brown fat cells [30]. Depletion of IRS-4 causes only modest
developmental disorder and insulin intolerance [26]. Regulation of IRS is the crucial point in
regulation of entire insulin or IGF receptor signalling cascade. IRS can be regulated in the
many different ways involving modification of the protein itself as well as the cytoplasmic
part of the receptor. While generally tyrosine phosphorylation of IRS positively stimulates
the receptor signalling, serine/threonine phosphorylation typically serves as a negative
9
regulator of the receptor signalling. There are two groups of the kinases that negatively
regulate IRS. One group, like mTOR/S6K1, MAP kinase and protein kinase C (PKC) are the
kinases activated by IRS signalling and working as a negative feedback loop. The second
group consists of kinases activated by other signalling pathways to control IRS stimulation.
Those are glycogen synthase kinase IKK ß, c-Jun NH2-terminal kinase, mouse Pelle-like
kinase and AMPK. Regulatory serines are distributed across entire IRS sequence.
Phosphoserines located on plakstrin homology domain inhibits plasma membrane – IRS
interaction. Phosphoserines located on phosphotyrosine binding domain disturb IRS
interaction with juxtamembrane domain of the receptor and can also promote degradation
of IRS. Phosphoserines on the C-terminal domain can both inhibit the interaction of IRS with
SH2 domains containing effecter proteins as well as promote the degradation of IRS. Not all
serine/threonine phosphorylations of IRS protein negatively regulate it function.
Phosphorylation of Ser1223 or Ser629 (human numbering) by PKB, results in enhanced Tyr
phosphorylation of IRS protein [31]. Apart from IRS proteins, at least 6 others factors, Gab-1,
three isoforms of Shc, p62dok and APS (adapter protein containing a PH and SH2 domain)
can directly bind to insulin receptor and activate signalling cascades [33].
4.4.1.1. Phosphatidylinositol 3-kinase pathway
The majority of nsulin- or IGF receptor mediated action is mediated through PI3K or the
MAP kinase pathways (figure 2). PI3K is composed of a regulatory subunit and a 110 kD
catalytic subunit. The regulatory subunit has four isoforms, (p85-a, p85-b, p55/AS53,
p55PIK, and p50) which can either stimulate of suppress some of the activity of PI3K. For
example, deletion of p85 alpha subunit causes hypoglycaemia by increase of the basal level
of insulin uptake in several insulin responsive tissues. PI3K is required for insulin stimulated
glucose uptake; however, it is not sufficient for insulin action. It is known that several
factors can activate PI3K; however, only insulin has the ability to stimulate GLUT4
translocation [26]. Studies conducted on 3T3 L1 adipocytes showed that Cbl- associated
protein (CAP) may also be necessary for insulin stimulated glucose transport. CAP has been
shown to bind directly to the insulin receptor [33]. The direct response to PI3K activation is
an increase in trisphosphorylated inositol (PIP3) concentration. PIP3 can further activate PIdependent protein kinase-1 and-2, Akt (a product of the akt protooncogene), salt- and
glucocorticoid-induced kinases, protein kinase C or wortmannin-sensitive and insulinstimulated serine kinase. Among those Akt seems to be crucial for and glycogen synthesis
and together with protein kinase C, for insulin stimulated glucose uptake. In the case of
glucose transport, Akt has been shown to act through AS160 protein [27].
4.4.1.2. MAP kinase pathway
In contrary to the PI3K pathway, the MAP kinase pathway mediates most of the insulin gene
expression regulation. This signalling cascade starts with recruitment of the Grb2 adaptor
10
protein by both IRS-1 and by the insulin/IGF receptor itself. Grb2 recruits Son-of-sevenless
(SOS) exchange factor to the plasma membrane, which in turns activates protein Ras.
Activation of protein phospatase SHP2 by the interaction with IRS 1/2 or Gab-1 is also
required for activation of Ras. Activated Ras initiates the cascade of kinases acting through
Raf, MEK and ERK. Phosphorylated ERK then translocates to the nucleuses where it can
perform action on transcription factors such as p62TCF. Activation of ERK generally leads to
stimulation of proliferation and differentiation. [34]
4.4.1.3. The CAP/Cbl pathway
As mentioned before, activation of the PI3 pathway is not sufficient to stimulate glucose
uptake. Based on a study using 3T3-L1 adipocytes, Cbl protein was proposed to work parallel
with PI3K signalling. Cbl can be directly phosphorylated by the tyrosine kinase of the insulin
receptor. Upon phosphorylation, Cbl binds adaptor protein CAP, which brings it to the
plasma membrane by binding flotilin, protein associated with the lipid rafts. Phosphorylated
Cbl binds another adaptor protein Crk2, which in turns binds exchange factor C3G. C3G
activates small G protein TC10. Activation of TC10 seems to cooperate with PI3K signalling
to stimulate GLUT4 translocation to the plasma membrane. The mechanism of how TC10
stimulate GLUT4 translocation is still unclear, but it has been suggested that TC10 action is
through stabilization of actin filaments. [34]
11
Figure 2. Signal transduction in insulin action. The insulin receptor is a tyrosine kinase that
undergoes autophosphorylation and catalyses the phosphorylation of cellular proteins such as
members of the IRS family, Shc and Cbl. Upon tyrosine phosphorylation, these proteins interact with
signalling molecules through their SH2 domains, resulting in a diverse series of signalling pathways,
including activation of PI(3)K and downstream PtdIns(3,4,5)P3-dependent protein kinases, Ras and
the MAP kinase cascade, and Cbl/CAP and the activation of TC10. These pathways act in a concerted
fashion to coordinate the regulation of vesicle trafficking, protein synthesis, enzyme activation and
inactivation, and gene expression, which results in the regulation of glucose, lipid and protein
metabolism Figure adapted from [34].
3.4.2. Defects in insulin signalling
The most common outcome of dysfunction of insulin receptor signalling is insulin resistance.
Insulin resistance is a state where insulin responsive tissue like liver, muscle and fatty tissue,
have lover response to insulin and therefor have lover ability to decrease blood glucose
level. Insulin resistance has been pointed as one of the major cause of development of type
2 diabetes [35, 36]. As was described previously, insulin signalling via PKB stimulates glucose
uptake by GLUT4 glucose transporter. It has been shown that reduced insulin stimulated
glucose transport/phosphorylation activity is an early event of type 2 diabetes [37]. Skeletal
muscle has been shown to have the biggest glucose storage capacity and to be the most
efficient organ in insulin stimulated glucose uptake process in human body [38]. The ability
12
of skeletal muscle to take up glucose under insulin stimulation decreases by 50% in patients
with type 2 diabetes [38]. On the other hand, liver is responsible for maintaining blood
glucose levels during fasting. Insulin signalling during the meal stimulates glycogen synthesis
in the liver and inhibits gluconeogenesis. During fasting, degradation of glucagon and
gluconeogenesis are promoted, and glucose is secreted from liver into the blood stream.
Therefore, one of the outcomes of insulin resistance in liver is increased fasting blood
glucose levels [16].
3.4.2.1. Fat Induced insulin resistance
Insulin resistance together with type 2 diabetes has been very highly correlated with
obesity, therefore triglycerides accumulation in insulin sensitive tissues has been proposed
as one of the causes of impaired insulin signalling (for review see [39]). It has been shown,
that elevated fasted blood fatty acid concentration as well as triglycerides accumulation in
muscle are good predictors for insulin resistance in humans [37]. Elevated fasted plasma
fatty acids has been observed in insulin resistant offspring of type two diabetic parents,
which otherwise did not display conditions like obesity or hyperglycaemia [37].
As described previously, insulin activates Pi3 kinase via IRS-1 which leads to GLUT4
translocation to the plasma membrane and glucose uptake in insulin sensitive tissues. It has
been demonstrated that increased plasma fatty acids leads to attenuation ofPI3 kinase
activation by insulin in skeletal muscle. This action has been demonstrated to be conducted
by blocking of insulin receptor conducted tyrosine phosphorylation of IRS-1 (see Figure 3)
[37, 39]. Decreased tyrosine phosphorylation of IRS-1 has been demonstrated to be
mediated by IRS-1 serine phosphorylation. Pharmacological inhibition or knock down of
certain protein kinases (c-Jun NH2-terminal kinase, inhibitor of nuclear factor kappa beta
kinase β subunit, S6 kinase 1, and protein kinase C-θ) have been demonstrated to prevent
high fat induced insulin resistance in certain rodent models [37]. Prevention of high fat
induced insulin resistance has been achieved also by muscle specific serine to alanine
mutation of 302, 307 and 612 IRS-1 residues [37]. Increase of serine phosphorylation of IRS1 in muscle has been shown to be secondary to intracellular increases in long-chain fatty
acids and diacylglycerol [37]. Triacylglycerol and acyl CoAs intracellular accumulation
however, does not cause insulin resistance. Therefore regulation of diacylglycerol content
via enzymes involved in diacylglycerol metabolism plays a significant role in insulin
resistance development [39]. Diacylglycerol can directly activate novel PKCs, which unlike
conventional PKCs does not require calcium ions as a co-activator. In skeletal muscles,
diacylglycerol action is mediated through PKC θ, which further phosphorylates serine
residues of IRS-1 (see Figure 3), whereas in the liver diacylglycerol activates PKC ε, which in
turns phosphorylates serine residues of insulin receptor (see Figure 4). Another member of
the atypical PKC family, PKC δ has been demonstrated to play a role in development of
hepatic insulin resistance, however unlike PKC ε, PKC δ has been suggested to be activated
13
by inflammatory signalling. PKC δ regulates expression of genes involved in lipogenesis,
therefore PKC δ involvement in development of hepatic insulin resistance has been
suggested to be conducted via increase hepatic lipid accumulation. [39]
Ceramides are another group of lipids associated with insulin resistance. Ceramides are
synthesized from serine and acyl-CoA.
Ceramides mediate only saturated fat induced insulin resistance, since inhibition of ceramid
synthesis pathway in rodents prevents insulin resistance development by perfiusion of
saturated fatty acids, but not by unsaturated fatty acids. There is a line of evidence
suggesting involvement of Toll-like receptor 4 in ceramide accumulation and ceramide
induced insulin resistance. Ceramide mediated inhibition of insulin signalling has been
shown to be downstream from IRS and does not impair IRS tyrosine phosphorylation.
Ceramides have been shown to impair Akt2 activation, however, the exact mechanism of
this inhibition is not known. There are several implications suggesting that inhibition of Akt2
by ceramics is mediated by activation of protein phosphatase 2 A, which in turns
dephosphorylates and by that deactivates Akt2. Another suggested mechanism is via PKC ζ
isoform, since ceramides prevent dissociation of PKC ζ and Akt2 complex. [39]
The role of hepatic fat accumulation in development of insulin resistance, besides the
mechanism described above, has been associated with activation of unfolded protein
response, known also as endoplasmic reticulum stress. Endoplasmic reticulum stress is
characterized with accumulation of unfolded proteins, which attracts glucose-regulated
protein 78 known as BIP. BIP in the basal state is bind to endoplasmic reticulum membrane
proteins inositol requiring enzyme-1 (IRE1a), PKR-like ER kinase (PERK), and activating
transcription factor-6 (ATF6), suppressing their activity. After activation, those proteins work
to reduce unfolded proteins with the ER lumen by increasing membrane biogenesis,
suppressing protein translation and enhancing expression of ER chaperones. IRE1a has been
shown also to activate Jun-N-terminal kinase 1 (JNK1), which can phosphorylate serine
residue of IRS -1, and therefore inhibit insulin signalling. Endoplasmic reticulum stress has
been demonstrated to increase hepatic lipid accumulation by upregulation of SREBP1c and
ChREBP, the main transcription factors of lipogenesis. The expression of lipogenic genes is
also enhanced by XBP1s transcription factor, which is an alternative splicing form of XBP1
gene. IRE1a has been shown to conduct splicing of XBP1s and by that increase lipogenesis.
On the other hand activated PERK increases gluconeogenesis which contributes to fasting
hyperglycemia. [39]
14
Figure 3. The mechanism of insulin resistance caused by fatty acids in muscle. Insulin activates the
insulin receptor (IR) tyrosine kinase, which subsequently tyrosine phosphorylates IRS1. Through a
series of intermediary steps, this leads to activation of Akt2. Akt2 activation, via AS160 and RabGTPase (not shown), promotes the translocation of GLUT4-containing storage vesicles (GSVs) to the
plasma membrane, permitting the entry of glucose into the cell, and promotes glycogen synthesis
via glycogen synthase (GS). This central signalling pathway is connected to multiple other cellular
pathways that are designated by numbers 1–3. (1) The green shaded areas represent mechanisms
for lipid induced insulin resistance, notably diacylglycerol (DAG)-mediated activation of PKCq and
subsequent impairment of insulin signalling, as well as ceramide-mediated increases in PP2A and
increased sequestration of Akt2 by PKCz. Impaired Akt2 activation limits translocation of GSVs to the
plasma membrane, resulting in impaired glucose uptake. Impaired Akt2 activity also decreases
insulin-mediated glycogen synthesis. (2) The yellow areas depict several intracellular inflammatory
pathways—notably, the activation of IKK, which may impact ceramide synthesis, and the activation
15
of JNK1, which may impair insulin signalling via serine phosphorylation of IRS1. (3) The pink area
depicts activation of the unfolded protein response (UPR), which under some instances (such as
acute extreme exercise) may lead to activation of ATF6 and a PGC1a-mediated adaptive response.
The endoplasmic reticulum membranes also contain key lipogenic enzymes and give rise to lipid
droplets. Proteins that regulate the release from these droplets (e.g., ATGL and PNPLA3) may
modulate the concentration of key lipid intermediates in discrete cell compartments. CS, ceramide
synthase; G3P, glycerol 3-phosphate; LPA, lysophosphatidic acid; SPT, serine palmitoyl transferase;
TAG, triacylglycerol. Figure and figure description adapted from [39].
16
Figure 4. The mechanism of insulin resistance caused by fatty acids in liver. Pathways Involved in
Hepatic Insulin Resistance Insulin activates the insulin receptor (IR) tyrosine kinase, which
subsequently tyrosine phosphorylates IRS1 and 2. Through a set of intermediary steps, this leads to
activation of Akt2. Akt2 can promote glycogen synthesis (not shown), suppress gluconeogenesis, and
activate de novo lipogenesis (DNL). This central signalling pathway is connected to multiple other
cellular pathways that are designated by numbers 1–3. Key lipid synthesis pathways are juxtaposed
within this domain and are regulated by them. (1) The green shaded areas represent mechanisms for
lipid-induced insulin resistance—notably, diacylglycerol-mediated activation of PKCε and subsequent
impairment of insulin signalling, as well as ceramide mediated increases in PP2A and increased
sequestration of Akt2 by PKCz. Impaired Akt2 activation limits the inactivation of FOXO1 and allows
for increased expression of key gluconeogenesis enzymes. Impaired Akt2 activity also decreases
insulin-mediated glycogen synthesis (not shown). (2) The yellow areas depict several intracellular
inflammatory pathways—notably, the activation of IKK, which may impact ceramide synthesis, and
the activation of JNK1, which may impair lipogenesis. (3) The pink area depicts activation of the UPR
that can lead to increased lipogenesis via XBP1s and also increased gluconeogenesis via C/EBP. The
ER membranes also contain key lipogenic enzymes and give rise to lipid droplets. Proteins that
regulate the release from these droplets (e.g., ATGL and PNPLA3) may modulate the concentration
of key lipid intermediates in discrete cell compartments. CS, ceramide synthase; DNL, de novo
lipogenesis; FA-CoA, fatty acyl CoA; G3P, glycerol 3-phosphate; LPA, lysophosphatidic acid; SPT,
serine palmitoyl transferase; TAG, triacylglycerol; TCA, tricarboxylic acid cycle; PEP,
phosphoenolpyruvate. Figure and figure description adapted from [39].
3.4.2.2. Glucose induced insulin resistance
Insulin resistance induced by glucose is less explored than fatty acid induced insulin
resistance. Skeletal muscle and adipose tissue are responsible for glucose clearance;
therefore most investigations were trying to link glucose with peripheral insulin resistance.
It has been demonstrated that hyperglycaemia impairs glucose transport in vivo [40]. The ex
vivo experiment with perfused rat skeletal muscle demonstrated that glucose is able to
inhibit insulin stimulated glucose uptake. The effect of glucose is highly potentiated when
glucose is administrated together with insulin, although insulin by itself does not impair
insulin stimulated glucose uptake [41]. A similar observation has been noticed with human
subcutaneous adipose tissue. Ex vivo treatment of human adipocytes with high
concentration of insulin, alone or in combination with high concentration of glucose,
decreased both basal and insulin stimulated glucose uptake. Moreover pretreatment of
human adipocytes with high concentration of insulin in combination with high
concentration of glucose decreased quantity of IRS-1 proteins and furthermore decreased
ability of insulin to activate PKB. Surprisingly, administration of high concentration of insulin
alone did not reduce IRS-1 quantity or insulin stimulated PKB activation, yet glucose uptake
was decreased [42]. Glucose and insulin induced insulin resistance has been proposed to be
mediated through increased intracellular protein modification by O-linked N-acetylglucosamine (process known as O-GlcNAcylation). However, a study linking O-GlcNAcylation
and insulin resistance seems to be contradictory. N-acetyl-glucosamine is produced from
17
glucose in hexosamine biosynthetic pathway. An increase in glucose intracellular content
has been reported to increase O-GlcNAcylation and increased O-GlcNAcylation has been
demonstrated to induce insulin resistance in skeletal muscle and adipose tissue [43]. In
primary adipocytes culture, inhibition of rate limiting enzyme in hexosamine biosynthetic
pathway (glutamine:fructose-6-phosphate amidotransferase) prevented glucose and insulin
induced insulin resistance. This effect was diminished by administration of glucosamine, a
carbohydrate that enters the hexosamine biosynthetic pathway downstream from
glutamine:fructose-6-phosphate amidotransferase [44]. However, in another study
inhibition of the rate limiting enzyme in N-acetyl-glucosamine synthesis did not prevent the
reduction in glucose intake caused by high glucose combined with high insulin treatment of
isolated rat muscle [45]. The same study demonstrated that inhibition of protein synthesis
completely abolished the inhibitory effect of glucose and insulin on glucose uptake. Taken
together, more studies are required in order to evaluate the possible role of OGlcNAcylation in glucose induced insulin resistance. Glucose as well as glucosamine can
induce endoplasmin reticulum stress, which, as described above, can induce insulin
resistance [46]. N-acetyl-glucosamine is covalently attached to serine or tyrosine residue.
One of the proteins undergoing O-GlcNAcylation is IRS. O-GlcNAcylated residues of IRS can
no longer be phosphorylated by the insulin receptor and this action can explain how OGlcNAcylation decrease insulin signalling [43].
3.5. Cyclic AMP
cAMP is a secondary messenger molecule produced by a membrane enzyme adenylyl
cylcase. cAMP is synthetized from ATP and is degraded to AMP by phosphodiesterase.
Adenylyl cylcasecyclase is attached to the plasma membrane by two hydrophobic domains.
Two cytoplasmic domains are responsible for the catalytic activity and binding of G protein.
There are at least 9 different adenylyl cyclase types encoded by at least 9 different genes,
and more than 40 types of phosphodiesterases. Adenylyl cylcases are expressed at different
levels and are differently regulated in different tissues. Almost all adenylyl cyclases can be
directly activated by Gs alpha protein coupled receptors, like beta adrenergic receptors or
by forskolin. Additionally, depending on the type, adenylyl cyclase can be regulated by Gαi
and Gβγ G proteins coupled receptors, calcium ions as well as by some kinases and
phosphatases like PKA, PKC calmodulin kinase and calcineurin. All Adenylyl cyclases are
expressed in some part of the brain [47]. Phosphodiesterases can target cGMP or cAMP or
both of them. As a result, cAMP activity of some phosphodiesterases can be regulated by
the cGMP concentration, in a competition like manner. Additionally, cAMP by itself can
regulate activity of cAMP specific phosphodiesterases. It has also been reported that insulin,
histamine and calcium ions regulates phosphodiesterase activity [48, 49]. cAMP is involved
18
in many processes in almost every tissue in our body. Many hormones, like vasopressin,
dopamine, histamine or prostacyclin works through this secondary messenger. cAMP plays a
crucial role in glucose homeostasis and energy balance by mediating both stress and hunger
signalling via adrenalin and glucagon receptors. Adrenalin and glucagon signalling have the
most robust effect in adipocytes and liver where they counteract the effect of insulin by
stimulating of lypolysis and gluconeogenesis [50, 51]. So far there are two known
mammalians proteins with cAMP binding domain: protein kinase A (PKA) and exchange
factor directly activated by cAMP (EPAC).
3.5.1. Epac
Epac was discovered in 1998 by the genome database search for new cAMP binding
proteins, and is a guanine nucleotide exchange factor for small G protein Rap [52]. There are
two Epac isoforms, Epac 1 and 2. Epac 2 has two splicing variants named Epac 2A and 2B.
Also Epac 1 exists in at least two splicing variants. Both Epac 1 and 2 contains dishevelled
egl-10 pleckstrin domain attaching the protein to the cellular membranes, followed by a
cAMP binding domain, a Ras exchanger motif, a Ras associated domain and a guanine
nucleotide exchange domain (CDC25-homology domain). Epac 2 has an additional cAMP
binding domain at the N terminal of the protein (see Figure 5). The cAMP binding domain of
Epac is approximately 40 times less sensitive to cAMP than the regulatory subunit of PKA
and PKA is therefore considered as the primary cAMP target. Epac is thus considered as a
sensor of strong cAMP signalling. Epac 1 is ubiquitously expressed whereas Epac 2 seems to
be specific for certain brain region and adrenal glands. Epac 2B is expressed in pancreatic
islets. Epac 1 has different cellular localisation and has been observed in both in nucleuses
and mitochondria as well as on mitotic spindle during the mitosis. Both Epac isoforms have
ability to activate Rab 1 and 2. The Epac Rab1/2 signalling is very poorly describes. Several
reports point phosphor lipase C, Akt, mTORC2, inositol 1-4-5 trisphosphate, ERK, Ca2+
/calmodulin-dependent protein kinase II and PKC as the Epac downstream targets [53-55].
There are several reports suggesting regulation of cell adhesion via integrin signalling by
Rab1. Epac 2 on the other hand was shown to play an important role in insulin secretion by
controlling insulin containing vesicles trafficking [56]. For 15 years after its discovery, Epac
has been demonstrated to be involved in regulation of leptin secretion and signalling in
hypothalamus [57], cardiac Ca2+ signalling [58], cancer cell migration [59], learning and
social interactions [60], exocitosis [61], neuronal differentiation and neurit outgrowth,
cardiac hypertrophy [55], prevention of tissue fibrosis [62], smooth muscles relaxation [63]
and neuronal apoptosis [64].
19
Figure 5. Domain structure of Epac proteins. CNB (cyclic nucleotide–binding). DEP (Disheveled, Egl-10, and
Pleckstrin), CDC25-HD (CDC25-homology domain), REM (Ras exchange motif), RA (Ras-association). Figure
taken from [56].
3.5.2. cAMP action in adipocytes
Adipocytes express multiple adrenergic receptors (β1, β2, β3, α1 and α2). As described
previously, activation of β adrenergic receptors stimulates cAMP signalling, whereas
stimulation of α adrenergic receptors decreases cAMP signalling [65]. Expression of
adrenergic receptors makes adipocytes sensitive to action of catecholamines, like
noradrenaline and adrenaline.
Adipocytes are the main fat storage cells in our body. The primary purpose of this organ is to
provide free fatty acids from stored energy when the food is limited and energy required.
Glucose is the primary cellular fuel, however when glucose is limited, free fatty acids are
secreted from adipose tissue to the blood stream and are oxidised by the liver, skeletal
muscle and heart. Moreover, fatty acids are used as a fuel in the situations of stress and
increased exercise. Therefore, processes responsible for lipid accumulation as well as for
lipolysis need to be strictly controlled by the mechanism regulated by energy status of the
organism. Hormone sensitive lipase (HSL) is the crucial enzyme in the lipolysis process. It
initiates triglycerides degradation by cleaving off the first fatty acids chain from
triacylglycerol. HSL upon activation translocates from the cytoplasm to the lipid droplets,
where the first step of lipolysis occurs. The best known mechanism promoting lipolysis is by
activation of cAMP pathway (see Figure 6). PKA is the main enzyme regulating HSL activity.
PKA phosphorylates HSL on serine Ser-563 (site 1) and Ser-565 (site 2) (numeration
according to human protein). Side 1 is phosphorylated upon lipolitic stimulation, whereas
side 2 is phosphorylated at the basal state and when phosphorylated, prevents
phosphorylation of the side 1 which decreases lipolysis. Other kinases, like glycogen
synthase kinase, AMP dependent protein kinase and Ca2+/calmodulin dependent protein
kinase II can also phosphorylate HSL on the side 2. However, phosphorylation of side 1 is not
crucial for enzyme activity. Further studies discovered that Ser-659 and 660 are
phosphorylation targets of cAMP signalling and regulators of the enzyme activity. Perlipins
are the lipofilic, lipid droplets associated proteins which are abundant on the surface of the
lipid droplets. The function of those proteins is prevention of interaction of lipolitic enzymes
with the lipids insight lipid droplets. Perlipins have multiple phosphorylation sides which are
20
also targets of PKA. Upon phosphorylation, perlipins lose their blocking ability, therefore
cAMP signalling enables HSL translocation from cytoplasm to the lipid droplets. cAMP is not
the only factor regulating perlipins and HSL in adipocytes. TNF-α also induces lipolysis in
adipocytes, regulating HSL and perlipins. [66]
All cAMP action described in this paragraph can ascribed to PKA activity. However, recent
studies indicated that Epac may be involved in regulation of AMP-activated protein kinase,
the main energy sensor in the cell. AMP-activated protein kinase is activated when
AMP/ATP ratio increases due to cell energy deprivation. Activated AMP-activated protein
kinase promotes energy delivering processes like for example adipocyte lipolysis. AMPactivated protein kinase is regulated by phosphorylation of tyrosine 172 (human protein
numeration) and specific Epac agonist has been shown to increase this phosphorylation and
activity of AMP-activated protein kinase. However, Epac specific agonist is not able to
increase glycolysis alone. [67]
cAMP signalling in adipocytes is regulated in multiple ways. Testosterone stimulates
expression of β adrenergic receptors, and thus sensitises adipocytes to adrenaline and
noradrenaline stimulation. TNF-α, on the other hand decreases the level of β-3 adrenergic
receptor, however, increases the level of β-2 receptors. Adenosine receptor is conjugated
with Gi protein activating phosphodiesterase, and it has been shown that TNF-α
downregulate the expression of Gi protein coupled with adenosine receptor. The insulin
action promotes utilisation of excess of nutrience in the blood stream, therefor insulin also
decreases lipolysis. Insulin inhibits lipolysis, stimulated by cAMP signalling, by activation of
phosphodiesterase 3B (PDEB3). Insulin activation of PDEB3 is conducted via phosphorylation
of Ser 302 residue (human protein sequence) by Akt. On the other hand thyroid hormones
and Tnf-α decrease activity of PDEB3. [66]
Increase in intracellular calcium level, like insulin, also increases activity of PDEB3 and
decreases lipolysis. Recent investigation demonstrated that the calcium-sensing receptor,
which plays a crucial role in regulation of plasma calcium level, is wildly expressed in
adipocytes. Activation of the calcium-sensing receptor leads to increase in intracellular
adipocyte calcium level, and by that inhibition of cAMP induced lipolysis. [68]
21
Figure 6. Action of cAMP and it regulation in adipocytes. Figure adapted form [66].
3.6. The role of insulin and cAMP in adipogenesis
3.6.1. Transcription factors of adipogenesis
While early events in adipocyte differentiation are still unclear, the final step of adipocyte
differentiation is far more characterised proliferator-activated receptor γ (PPARy). PPARy
activates expression of CCAAT/enhancer binding protein α (C/EBPα), which in turn feeds
back on PPARy. PPARy and C/EBPα activates genes necessary for the final development of
the mature, insulin-responsive adipocytes (see figure 7) [69, 70]. A few genes have been
reported to play a unique role for differentiation white or brown adipose tissue. Studies
conducted in mice have shown that C/EBPα is necessary for WAT origin and its lack does not
affect the BAT development. Bone-morphogenetic protein 7 (BMP7) has been reported to
induce BAT differentiation whereas two others members of BMP family, BMP 2 and 4
suppress brown adipocyte differentiation by inhibiting of UCP1 expression [69]. Most
knowledge about molecular mechanism of adipocyte differentiation comes from cell culture
studies. 3T3 L1 mouse fibroblasts, mouse embryo fibroblasts (MEFs) or 3T3 F442A are most
frequently used preadipocyte models. Adipocyte differentiation of those cells comes out
after 4 to 6 days of growth in medium containing foetal calf serum, high concentration of
insulin or physiological concentration of IGF-1 and glucocorticoids. Factors that elevate
22
cellular concentration of cAMP strongly enhance differentiation. In the standard
differentiation method, cells are treated with insulin, dexamethason and
isobutylmethylxanthine (IBMX) [71]. IBMX act as an inhibitor of phosphodiesterase and
when is present in the cytoplasm, elevates cAMP level [71]. Insulin or IGF-1 binds to IGF
receptor on the preadipocytes membrane surface. As it was described previously, activated
IGF receptor works through MAP and PI3K pathways, resulting with activation of PKB and
Erk. Both PKB and Erk can activate transcription factor cAMP responding element binding
(CREB), which initiate expression of adipocyte differentiation related genes [70, 71].
Glucocorticoids, like dexamethasone, are reported to be increased in several models of
obesity and to induce insulin resistance. In humans, excess of these hormones leads to
modification of fatty tissue distribution and to central obesity [72]. Unfortunately, it is still
unclear why dexamethasone is necessary for adipopogenesis. As others glucocorticoids,
dexamethasone diffuses through the cellular membrane and binds to the glucocorticoid
receptor (GR). GR is a nuclear transcription factor from the same superfamily like PPARγ.
One of the transcriptional targets of GR is C/EBPα, the main proadipogenic transcription
factor. It was reported, that operating through C/EBPα, dexamethasone promotes
transcription of insulin sensitive glucose transporter GLUT4 [73]. C/EBPα expression is
suppressed by TGF-β. Treatment of primary rat preadipocytes with TGF-β prevents
adipocyte differentiation, whereas co-treatment with dexamethasone counteracts effect of
TGF-β and reverses suppression of C/EBPα expression [74]. However, cells which
overexpress C/EBPα still require dexamethasone stimulation to undergo adipocyte
differentiation [70], what suggests other way of stimulation of adipogenesis by
dexamethasone. GR can also act as a genes expression inhibitor. It was reported, that
dexamethasone significantly reduces expression of tumour necrosis factor α (TNF-α) [72]
and preadipocyte factor-1 (pref-1) [70], the two potent antiadipogenic agents, and this
acting is believed to be crucial for dexamethasone proadipogenic properties.[75].
CCAAT/enhancer binding protein β and δ (C/EBPβ, C/EBPδ) are the earliest transcriptional
factors observed during adipocyte-like differentiation. One of the target genes of those two
factors encodes peroxisome
Elevation of the cAMP level activates PKA. PKA inactivates small G protein RhoA by
phosphorylation. RhoA is an activator of Rho associated kinase (ROCK), which is
serine/threonine specific. ROCK was found to regulate insulin/IGF receptor pathway via
phosphorylation of IRS protein. At high activity, ROCK phosphorylates serine residue 612 of
IRS-1 which cause inhibition of IGF receptor signalling. On the other hand, at low activity,
ROCK induces IGF receptor signalling by phosphorylation of serine residues 632 and 635.
However, blocking of ROCK, by specific inhibitor Y27632, can replace the action of PKB in
adipocyte differentiation, suggesting that ROCK activity is not necessary for adipocyte
differentiation [71]. In 2008, Petersen et al. [71] discovered that activation of both cAMP
targets is required for adipocyte like differentiation. The studies were carried out using 3T3
23
L1 mouse fibroblast preadipocytes and two cAMP analogues- 8-pCPT-2´-O-Me-cAMP (007)
and N6-monobutyrylcAMP (MB) which specifically activates only Epac or PKA respectively.
007 and MB have been used to induce adipocyte like differentiation instead of IBMX.
Figure 7. Molecular mechanism of adipogenesis induction in 3T3 L1 Mouse fibroblasts. IGF receptor activation,
as well as cAMP signalling and glucocorticoids stimulation is necessary for adipogenesis. Activation of IGF
receptor by insulin or IGF triggers two intracellular signalling cascades leading thru Pi-3 kinase pathway and
MAP kinase pathway. CREB is believed to stimulate expression of initial adipogenic transcription factors
C/EBPβ and C/EBPδ which afterwards induce expression of PPARγ. One of the PPARγ target genes is C/EBPα
which feeds back on PPARy. Products of genes activated by PPARγ and C/EBPα are necessary for fibroblasts
differentiation into adipocytes. CREP can be activated by three different enzymes: Erk, known also as a MAP
kinase, PKB, activated in PI-3 kinase pathway, and by PKA. However, in adipogenesis, CREB activation seems to
be Erk dependent. Glucocorticoids, also necessary for adipocyte differentiation, bind to their soluble receptor
which acts as a negative regulator of antiadipogenic Pref-1 and TNF-α genes. Schema based on [70-72].
24
This study demonstrated that stimulation of only one of the cAMP target proteins does not
potentiate dexamethasone and insulin induced adipocyte like differentiation in 3T3 L1 cells.
On the other hand, differentiation was strongly enhanced when both Epac1 and PKA were
activated. Those findings suggest that activation of Epac1 may counteract the negative
impact of PKA on insulin signalling transition in still unknown manner. Illustration of insulin,
cAMP and glucocorticoids signalling in preadipocytes is presented in Figure 7.
4. The aim of the study
The present thesis focuses on the role that in obesity play one of the major metabolism
regulating factor – insulin, and molecular regulator of cell metabolism – cAMP.
The purpose of the first part is to provide information about specific action of two cAMP
activated factors, Epac and PKA which together with insulin are the driving force of
adipogenesis in vitro.
The second part investigates how PTP1B, the phosphatase reported to regulate insulin
secretion, impacts body function in different types of diets.
The third part demonstrates how the model of in vitro preadipocyte cell differentiation,
driven by the combined action of insulin and cAMP, can be observed in vivo.
25
5. List of manuscripts:
1. Borkowksi K, Wrzesinski K, Rogowska-Wrzesinska A, Audouze K, Bakke J, Petersen RK,
Haj F, Madsen L, Kristiansen K. Proteomic analysis of cAMP signalling during adipocyte
differentiation of 3T3-L1 preadipocytes. (Manuscript)
2. Borkowski K, Bettaieb A, Bakke J, Xi Y, Myrmel LS, Haj F, Madsen L, Kristiansen K
Perturbation of first phase insulin secretion in mice with pancreas-specific ablation of
Ptpn1 attenuates sucrose induced peripheral insulin resistance. (Manuscript)
3. Dankel SN, Degerud EM, Borkowski K, Fjære E, Midtbø LK, Haugen C, Solsvik MH,
Lavigne AM, Kristiansen K, Liaset B, Sagen JV, Mellgren G, Madsen L (2013).Weight
cycling promotes circadian shift and fat gain in C57BL/6J mice. (Manuscript submitted to
American Journal of Physiology - Endocrinology and Metabolism)
26
6. Discussion of results
Insulin and cAMP signalling play significant roles in sensing energy condition in our body.
Insulin is associated with food availability and promotes processes of food storage such as
glucose uptake, lipogenesis and glycogen synthesis. On the other hand, cAMP signalling is
associated with nutrition deprivation, promoting release of stored energy by activation of
lipolysis and glycogen degradation. gluneogenesis is not release of energy, it consumes
energy [76]. More importantly, both signalling cascades directly interfere with each other. In
insulin sensitive tissue, like adipocytes, muscle cells and liver, Insulin inhibits cAMP signalling
by activation of phosphodiesterase. On the other hand, in differentiating adipocytes and
muscle, the Epac branch of cAMP signalling tends to potentiate insulin/Insulin like growth
factor signalling
6.1. The role of cAMP in the adipocyte differentiation
Our group has demonstrated that activation of Epac is necessary for cAMP-induced
acceleration of adipocyte differentiation [71]. The microarray analysis (data not published)
demonstrated that Epac by itself is unable to induce major transcriptional responses. Epac
has been demonstrated to play an important role in regulation of leptin signalling in
hypothalamus [57], cardiac Ca2+ signalling [58], cancer cell migration [59], learning and
social interactions [60], exocytosis [61], neuronal differentiation and neurit outgrowth,
cardiac hypertrophy [55], prevention of tissue fibrosis [62], smooth muscles relaxation [63]
and neuronal apoptosis [64]. In the first manuscript we have identified 7 novel targets of
Epac, which have not been previously reported. Only one of the target proteins displayed
changes at the mRNA level as well, demonstrating that the majority of the observed
differences could not be detected by gene expression analyses.
As was pointed previously, Epac is involved in calcium signalling in β-cells, and high levels of
intracellular calcium are known to inhibit adipocyte differentiation [77]. Three of the
proteins down-regulated in our analysis are calcium regulated, and changes of two of them
are Epac specific. One of them is an enzyme, lysosomal alpha glucosidase, which is involved
in degradation of glycogen to glucose. In mature adipocytes, cAMP signalling is involved in
activation of energy deriving processes, like glycogen degradation, however, this action is
associated with the PKA branch of cAMP signalling.
In vitro studies have shown that steroid hormones, like glucocorticoids, accentuate
adipogenesis in many settings [78]. All steroid hormones are derived from cholesterol. In
our analysis we have demonstrated that Epac stimulates the expression of a key enzyme in
cholesterol synthesis - Hydroxymethylglutaryl-CoA synthase at both the transcriptional and
the translational level. This action is not counteracted but rather potentiated by PKA.
Moreover, PKA alone does not induce Hydroxymethylglutaryl-CoA synthase expression.
27
cAMP is known to regulate testosterone production [79], however, this regulation appears
downstream from cholesterol synthesis. In vitro studies have demonstrated that cholesterol
rich very low density lipoprotein stimulates adipogenesis. Surprisingly, other reports have
demonstrated a suppressive role of cholesterol on mouse adipose-derived stromal cells
differentiation [80]. This action was accompanied by down-regulation of PPARgamma2
expression.
6.2. The role of insulin in mediating the negative effect of sucrose in high fat
diet
There is a considerable body of evidence that sucrose potentiates the obesogenic effect of
high fat diets. Sucrose supplementation of high fat diets causes an increase in body weight
together with an augmentation of fat depots in mice models, when compared to protein
supplementation [81-83]. Studies with chemical inhibitors of insulin secretion linked insulin
to the obesogenic effect of sucrose in high fat diets [22]. It was recently demonstrated, that
glucose-stimulated insulin secretion is highly dependent on communication between β cells
in the pancreas and this phenomenon is mediated by ephrins [84]. Both ephrin A5 and
EphA5 can differentially effect insulin secretion. Signalling from EphA5 attenuates insulin
secretion, while signalling by ephrin A5 enhances insulin secretion [84]. EphA5 is active
when phosphorylated and protein tyrosine phosphatase 1B (PTP1B) is responsible for its
dephosphorylation (unpublished data). Therefore, PTP1B was proposed to regulate insulin
secretion in pancreatic beta cells. In the study presented in second manuscript we have
used Ptpn1 pancreas specific deletion to further investigate the role of insulin in the
obesogenic effect of sucrose supplementation in high fat diet.
Sucrose has been demonstrated to induce peripheral insulin resistance in animal models
[85]. Series of ex vivo experiments have demonstrated that insulin is necessary for glucose
to induce insulin resistance in skeletal muscle and fat tissue [41, 42]. Increased insulin
secretion has been previously suggested to be a response of β cells to the development of
peripheral insulin resistance [86]. On the other hand, some studies have demonstrated that
hyperinsulinemia can induce insulin resistance [87]. High sucrose diets are known to
increase the first phase insulin secretion in animal models [88]. It has been previously
suggested that developing insulin resistance is counteracted by an increase in first phase
insulin secretion [86]. On the other hand elevated insulin level combined with high glucose
has been shown to induce insulin resistance in human adipose tissue [89].
In the second manuscript we show as expected that animals fed with a high fat diet
supplemented with sucrose displayed an increased first phase insulin secretion when
compared to the low fat fed animals. The first phase insulin secretion was not elevated
when sucrose was replaced with protein in the high fat diet. Of note, the effect of sucrose
on first phase insulin secretion was blunted in mice with pancreas specific Ptpn1 deletion,
supporting the hypothesis of an involvement of PTP1B in insulin secretion. On the other
28
hand, pancreas specific Ptpn1 deletion influenced neither fasting nor fed, plasma glucose
and insulin levels in mice fed sucrose or protein supplemented high fat diets. This suggest
that PTP1B may be involved in the rapid first phase insulin secretion, but has no or minor
impact on insulin secretion caused by prolonged glucose stimulation of β cells.
Pancreatic specific Ptpn1 deletion prevented induction of peripheral insulin resistance in
high fat high sucrose fed mice, linking first phase insulin secretion with sucrose induced
peripheral insulin resistance.
It is important to mention that sucrose supplementation of high fat diets significantly
increased the total mass of white fat depots, as well as the mass of the intrascapular brown
adipose depot. Although the difference in adipose tissue mass of wild type and KO mice was
not significantly different, pancreas specific Ptpn1 deletion blunted the difference in fat
depots mass between the mice fed a high fat diet supplemented with sucrose and mice fed
a high fat diet supplemented with protein.
6.3. Combine action of insulin and cAMP accelerate adipocyte
differentiation in vivo.
In vitro adipogenesis is induced by the combined action of insulin and cAMP [11]. Insulin and
cAMP signalling are usually stimulated in cell under different conditions. Insulin is and food
availability signal, promoting glucose intake, lipogenesis and glucagon synthesis. On the
other hand cAMP signalling is associated with glucagon and adrenalin stimulation, meaning
that it mediates responses to food deprivation and stress [76]. In the third manuscript we
hypothesized that so-called yoyo dieting (circles of caloric restriction and refeeding) would
promote an increase in fat tissue mass by stimulation of adipogenesis.
The results demonstrated that mice subjected to yoyo diet gained significantly more weight
compared to the mice fed ad libitum. Moreover, the total caloric intake was not different
between those two groups suggesting that yoyo dieting increases feed efficiency.
Calculation of the epididymal and inguinal fat cellularity demonstrated a very strong
tendency towards an increase of the total cell number in the inguinal adipose depot in the
yoyo dieting group when compared to the ad libitum fed group. The same tendency,
however, less significant, was observed in the epididymal adipose depot. The increase in
adipose cell mass could then explain the difference in the weight gain between yoyo dieting
and ad libitum fed mice. It is worth to mention that the masses of neither liver nor muscle
were different in the two groups. Despite the increased body weight, yoyo dieting mice
remained glucose tolerant. Also, plasma triglycerides and fatty acids were lower and
adiponectin levels were higher in the yoyo dieting group compared to the ad libitum group.
These finding may suggest an increase capacity of fat storage, which can be explained by an
increase in adipocyte cell number in the fat depots.
29
There is emerging evidence for a key role of clock genes in energy homeostasis [90]. The
aberration of circadian rhythm by night shift work is associated with an increased risk of
obesity and metabolic syndrome [91]. Of interest in this context, we describe in the third
manuscript a possible link between cAMP signalling and expression of circadian rhythm
genes in adipose tissue. Thus, we show that mice undergoing yoyo dieting have decreased
expression of several circadian genes in white adipose tissue, when compare to the ad
libitum fed mice. In our experiment, two circadian genes, albumin D box-binding protein
(Dbp) and thyrotroph embryonic factor (Tef) has been found to be down-regulated in
inguinal and epididymal adipose tissue of the yoyo group compare to the ad libitum fed
mice. Moreover, we have demonstrated that in 3T3-L1 cell model of adipocyte, cAMP
decrease dramatically expression of Tef and Dbp, which may suggests that similar
mechanism is activated in vivo in preadipocytes during conditions where cAMP signalling is
elevated adding another aspect to the many physiological roles of cAMP-dependent
processes.
6.4. Conclusions
In the current study we have demonstrated that cAMP signalling can regulate the circadian
rhythm genes, and we hypothesized that by down-regulation of circadian genes, cAMP
stimulates adipogenesis in vivo. Moreover, we have identified a number of new Epac
regulated proteins. The nature of those regulations prevented them from being identified
using microarray data analysis. Our results provide a solid base for further investigations of
the involvement of Epac in biological processes previously not related to this cAMPactivated factor.
We have demonstrated that pancreas specific Ptpn1 deletion blunts the increase in first
phase insulin secretion induced by sucrose in the context of a high fat diet, however, but has
no effect on insulin secretion caused by prolonged glucose stimulation of b cells. Moreover,
we suggest that a decrease in the first phase insulin secretion is linked to the prevention of
peripheral insulin resistance.
Moreover, we have shown that yoyo dieting increases feed efficiency and promotes obesity
in rodents partially by increasing fat cell number. However, the increase in obesity is not
accompanied with a deterioration of metabolic parameters.
30
7. Literature
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Team EE: WHO launches the World Health Statistics 2012. Eurosurveillance 2012, 17(20):1515.
Obesity epidemic puts millions at risk from related diseases. Indian journal of medical
sciences 1997, 51(12):482.
Giovannucci E, Michaud D: The role of obesity and related metabolic disturbances in
cancers of the colon, prostate, and pancreas. Gastroenterology 2007, 132(6):2208-2225.
Hausman DB, DiGirolamo M, Bartness TJ, Hausman GJ, Martin RJ: The biology of white
adipocyte proliferation. Obes Rev 2001, 2(4):239-254.
Spiegelman BM, Flier JS: Adipogenesis and obesity: rounding out the big picture. Cell 1996,
87(3):377-389.
Hirsch J, Batchelor B: Adipose tissue cellularity in human obesity. Clin Endocrinol Metab
1976, 5(2):299-311.
Salans LB, Cushman SW, Weismann RE: Studies of human adipose tissue. Adipose cell size
and number in nonobese and obese patients. The Journal of clinical investigation 1973,
52(4):929-941.
Spalding KL, Arner E, Westermark PO, Bernard S, Buchholz BA, Bergmann O, Blomqvist L,
Hoffstedt J, Naslund E, Britton T et al: Dynamics of fat cell turnover in humans. Nature
2008, 453(7196):783-787.
Knittle JL, Hirsch J: Effect of early nutrition on the development of rat epididymal fat pads:
cellularity and metabolism. The Journal of clinical investigation 1968, 47(9):2091-2098.
Brook CGD, Lloyd JK, Wolf OH: Relation between Age of Onset of Obesity and Size and
Number of Adipose Cells. Brit Med J 1972, 2(5804):25-&.
Guller S, Corin RE, Mynarcik DC, London BM, Sonenberg M: Role of insulin in growth
hormone-stimulated 3T3 cell adipogenesis. Endocrinology 1988, 122(5):2084-2089.
Klaman LD, Boss O, Peroni OD, Kim JK, Martino JL, Zabolotny JM, Moghal N, Lubkin M, Kim
YB, Sharpe AH et al: Increased energy expenditure, decreased adiposity, and tissue-specific
insulin sensitivity in protein-tyrosine phosphatase 1B-deficient mice. Molecular and cellular
biology 2000, 20(15):5479-5489.
Ingle DJ, Prestrud MC, Nezamis JE: The effect of insulin upon the level of blood amino acids
in the eviscerated rat as related to the level of blood glucose. The American journal of
physiology 1947, 150(4):682-685.
Reynolds TH, Supiano MA, Dengel DR: Regional differences in glucose clearance: effects of
insulin and resistance training on arm and leg glucose clearance in older hypertensive
individuals. Journal of Applied Physiology 2007, 102(3):985-991.
Karnieli E, Armoni M: Transcriptional regulation of the insulin-responsive glucose
transporter GLUT4 gene: from physiology to pathology. American journal of physiology
Endocrinology and metabolism 2008, 295(1):E38-45.
Shepherd PR, Kahn BB: Mechanisms of disease - Glucose transporters and insulin action Implications for insulin resistance and diabetes mellitus. New Engl J Med 1999, 341(4):248257.
Jagadish MN, Fernandez CS, Hewish DR, Macaulay SL, Gough KH, Grusovin J, Verkuylen A,
Cosgrove L, Alafaci A, Frenkel MJ et al: Insulin-responsive tissues contain the core complex
protein SNAP-25 (synaptosomal-associated protein 25) A and B isoforms in addition to
syntaxin 4 and synaptobrevins 1 and 2. The Biochemical journal 1996, 317 ( Pt 3):945-954.
Kusari AB, Byon J, Bandyopadhyay D, Kenner KA, Kusari J: Insulin-induced mitogen-activated
protein (MAP) kinase phosphatase-1 (MKP-1) attenuates insulin-stimulated MAP kinase
activity: a mechanism for the feedback inhibition of insulin signaling. Mol Endocrinol 1997,
11(10):1532-1543.
31
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
O'Brien RM, Streeper RS, Ayala JE, Stadelmaier BT, Hornbuckle LA: Insulin-regulated gene
expression. Biochemical Society transactions 2001, 29(Pt 4):552-558.
Ducluzeau PH, Perretti N, Laville M, Andreelli F, Vega N, Riou JP, Vidal H: Regulation by
insulin of gene expression in human skeletal muscle and adipose tissue. Evidence for
specific defects in type 2 diabetes. Diabetes 2001, 50(5):1134-1142.
Plum L, Belgardt BF, Bruning JC: Central insulin action in energy and glucose homeostasis.
The Journal of clinical investigation 2006, 116(7):1761-1766.
Hao Q, Lillefosse HH, Fjaere E, Myrmel LS, Midtbo LK, Jarlsby RH, Ma T, Jia B, Petersen RK,
Sonne SB et al: High-glycemic index carbohydrates abrogate the antiobesity effect of fish
oil in mice. American journal of physiology Endocrinology and metabolism 2012,
302(9):E1097-1112.
Mehran AE, Templeman NM, Brigidi GS, Lim GE, Chu KY, Hu X, Botezelli JD, Asadi A, Hoffman
BG, Kieffer TJ et al: Hyperinsulinemia drives diet-induced obesity independently of brain
insulin production. Cell metabolism 2012, 16(6):723-737.
Mayer-Davis EJ, Levin S, Bergman RN, D'Agostino RB, Karter AJ, Saad MF: Insulin secretion,
obesity, and potential behavioral influences: results from the Insulin Resistance
Atherosclerosis Study (IRAS). Diabetes Metab Res 2001, 17(2):137-145.
Accili D, Nakae J, Kim JJ, Park BC, Rother KI: Targeted gene mutations define the roles of
insulin and IGF-I receptors in mouse embryonic development. Journal of pediatric
endocrinology & metabolism : JPEM 1999, 12(4):475-485.
Kido Y, Nakae J, Accili D: Clinical review 125: The insulin receptor and its cellular targets.
The Journal of clinical endocrinology and metabolism 2001, 86(3):972-979.
Siddle K: Signalling by insulin and IGF receptors: supporting acts and new players. Journal
of molecular endocrinology 2011, 47(1):R1-10.
de la Monte SM, Wands JR: Review of insulin and insulin-like growth factor expression,
signaling, and malfunction in the central nervous system: relevance to Alzheimer's disease.
Journal of Alzheimer's disease : JAD 2005, 7(1):45-61.
Leibiger IB, Leibiger B, Berggren PO: Insulin signaling in the pancreatic beta-cell. Annual
review of nutrition 2008, 28:233-251.
Escribano O, Arribas M, Valverde AM, Benito M: IRS-3 mediates insulin-induced glucose
uptake in differentiated IRS-2(-/-) brown adipocytes. Molecular and cellular endocrinology
2007, 268(1-2):1-9.
Boura-Halfon S, Zick Y: Phosphorylation of IRS proteins, insulin action, and insulin
resistance. American journal of physiology Endocrinology and metabolism 2009,
296(4):E581-591.
Friedrichsen M, Poulsen P, Richter EA, Hansen BF, Birk JB, Ribel-Madsen R, Stender-Petersen
K, Nilsson E, Beck-Nielsen H, Vaag A et al: Differential aetiology and impact of
phosphoinositide 3-kinase (PI3K) and Akt signalling in skeletal muscle on in vivo insulin
action. Diabetologia 2010, 53(9):1998-2007.
Bjornholm M, Zierath JR: Insulin signal transduction in human skeletal muscle: identifying
the defects in Type II diabetes. Biochemical Society transactions 2005, 33(Pt 2):354-357.
Saltiel AR, Kahn CR: Insulin signalling and the regulation of glucose and lipid metabolism.
Nature 2001, 414(6865):799-806.
Warram JH, Martin BC, Krolewski AS, Soeldner JS, Kahn CR: Slow glucose removal rate and
hyperinsulinemia precede the development of type II diabetes in the offspring of diabetic
parents. Ann Intern Med 1990, 113(12):909-915.
Kitabchi AE, Temprosa M, Knowler WC, Kahn SE, Fowler SE, Haffner SM, Andres R, Saudek C,
Edelstein SL, Arakaki R et al: Role of insulin secretion and sensitivity in the evolution of
type 2 diabetes in the diabetes prevention program: effects of lifestyle intervention and
metformin. Diabetes 2005, 54(8):2404-2414.
32
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
Morino K, Petersen KF, Shulman GI: Molecular mechanisms of insulin resistance in humans
and their potential links with mitochondrial dysfunction. Diabetes 2006, 55:S9-S15.
Shulman GI, Rothman DL, Jue T, Stein P, Defronzo RA, Shulman RG: Quantitation of Muscle
Glycogen-Synthesis in Normal Subjects and Subjects with Non-Insulin-Dependent Diabetes
by C-13 Nuclear Magnetic-Resonance Spectroscopy. New Engl J Med 1990, 322(4):223-228.
Samuel VT, Shulman GI: Mechanisms for insulin resistance: common threads and missing
links. Cell 2012, 148(5):852-871.
Rossetti L, Smith D, Shulman GI, Papachristou D, Defronzo RA: Correction of Hyperglycemia
with Phlorizin Normalizes Tissue Sensitivity to Insulin in Diabetic Rats. Journal of Clinical
Investigation 1987, 79(5):1510-1515.
Richter EA, Hansen BF, Hansen SA: Glucose-Induced Insulin Resistance of Skeletal-Muscle
Glucose-Transport and Uptake. Biochemical Journal 1988, 252(3):733-737.
Renstrom F, Buren J, Svensson M, Eriksson JW: Insulin resistance induced by high glucose
and high insulin precedes insulin receptor substrate 1 protein depletion in human
adipocytes. Metabolism-Clinical and Experimental 2007, 56(2):190-198.
McLarty JL, Marsh SA, Chatham JC: Post-translational protein modification by O-linked Nacetyl-glucosamine: Its role in mediating the adverse effects of diabetes on the heart. Life
sciences 2013, 92(11):621-627.
Marshall S, Bacote V, Traxinger RR: Discovery of a Metabolic Pathway Mediating GlucoseInduced Desensitization of the Glucose-Transport System - Role of Hexosamine
Biosynthesis in the Induction of Insulin Resistance. Journal of Biological Chemistry 1991,
266(8):4706-4712.
Kawanaka K, Han DH, Gao J, Nolte LA, Holloszy JO: Development of glucose-induced insulin
resistance in muscle requires protein synthesis. The Journal of biological chemistry 2001,
276(23):20101-20107.
Copeland RJ, Bullen JW, Hart GW: Cross-talk between GlcNAcylation and phosphorylation:
roles in insulin resistance and glucose toxicity. Am J Physiol-Endoc M 2008, 295(1):E17-E28.
Hanoune J, Defer N: Regulation and role of adenylyl cyclase isoforms. Annual review of
pharmacology and toxicology 2001, 41:145-174.
Bender AT, Beavo JA: Cyclic nucleotide phosphodiesterases: Molecular regulation to clinical
use. Pharmacol Rev 2006, 58(3):488-520.
Kass DA, Takimoto E, Nagayama T, Champion HC: Phosphodiesterase regulation of nitric
oxide signaling. Cardiovascular research 2007, 75(2):303-314.
Torii N: [The role of Ca ions and cAMP in ACTH lipolysis]. Nihon Naibunpi Gakkai zasshi
1985, 61(9):859-871.
Zhang C, Wang G, Zheng Z, Maddipati KR, Zhang X, Dyson G, Williams P, Duncan SA, Kaufman
RJ, Zhang K: Endoplasmic reticulum-tethered transcription factor cAMP responsive
element-binding protein, hepatocyte specific, regulates hepatic lipogenesis, fatty acid
oxidation, and lipolysis upon metabolic stress in mice. Hepatology 2012, 55(4):1070-1082.
de Rooij J, Zwartkruis FJT, Verheijen MHG, Cool RH, Nijman SMB, Wittinghofer A, Bos JL:
Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature
1998, 396(6710):474-477.
Misra UK, Pizzo SV: Upregulation of mTORC2 activation by the selective agonist of EPAC, 8CPT-2Me-cAMP, in prostate cancer cells: assembly of a multiprotein signaling complex.
Journal of cellular biochemistry 2012, 113(5):1488-1500.
Mayati A, Levoin N, Paris H, N'Diaye M, Courtois A, Uriac P, Lagadic-Gossmann D, Fardel O,
Le Ferrec E: Induction of intracellular calcium concentration by environmental
benzo(a)pyrene involves a beta2-adrenergic receptor/adenylyl cyclase/Epac-1/inositol
1,4,5-trisphosphate pathway in endothelial cells. The Journal of biological chemistry 2012,
287(6):4041-4052.
33
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
Laurent AC, Breckler M, Berthouze M, Lezoualc'h F: Role of Epac in brain and heart.
Biochemical Society transactions 2012, 40(1):51-57.
Bos JL: Epac: a new cAMP target and new avenues in cAMP research. Nat Rev Mol Cell Bio
2003, 4(9):733-738.
Yan J, Mei FC, Cheng H, Lao DH, Hu Y, Wei J, Patrikeev I, Hao D, Stutz SJ, Dineley KT et al:
Enhanced leptin sensitivity, reduced adiposity and improved glucose homeostasis in mice
lacking of Exchange protein directly activated by cAMP isoform 1. Molecular and cellular
biology 2012.
Ruiz-Hurtado G, Morel E, Dominguez-Rodriguez A, Llach A, Lezoualc'h F, Benitah JP, Gomez
AM: Epac in cardiac calcium signaling. Journal of molecular and cellular cardiology 2012.
Almahariq M, Tsalkova T, Mei FC, Chen H, Zhou J, Sastry SK, Schwede F, Cheng X: A Novel
EPAC-Specific Inhibitor Suppresses Pancreatic Cancer Cell Migration and Invasion.
Molecular pharmacology 2013, 83(1):122-128.
Yang Y, Shu X, Liu D, Shang Y, Wu Y, Pei L, Xu X, Tian Q, Zhang J, Qian K et al: EPAC null
mutation impairs learning and social interactions via aberrant regulation of miR-124 and
Zif268 translation. Neuron 2012, 73(4):774-788.
van Hooren KW, van Agtmaal EL, Fernandez-Borja M, van Mourik JA, Voorberg J, Bierings R:
The Epac-Rap1 signaling pathway controls cAMP-mediated exocytosis of Weibel-Palade
bodies in endothelial cells. The Journal of biological chemistry 2012, 287(29):24713-24720.
Insel PA, Murray F, Yokoyama U, Romano S, Yun H, Brown L, Snead A, Lu D, Aroonsakool N:
cAMP and Epac in the regulation of tissue fibrosis. British journal of pharmacology 2012,
166(2):447-456.
Roscioni SS, Maarsingh H, Elzinga CR, Schuur J, Menzen M, Halayko AJ, Meurs H, Schmidt M:
Epac as a novel effector of airway smooth muscle relaxation. Journal of cellular and
molecular medicine 2011, 15(7):1551-1563.
Suzuki S, Yokoyama U, Abe T, Kiyonari H, Yamashita N, Kato Y, Kurotani R, Sato M, Okumura
S, Ishikawa Y: Differential roles of Epac in regulating cell death in neuronal and myocardial
cells. The Journal of biological chemistry 2010, 285(31):24248-24259.
Fain JN, Garciasainz JA: Adrenergic Regulation of Adipocyte Metabolism. Journal of lipid
research 1983, 24(8):945-966.
Carmen GY, Victor SM: Signalling mechanisms regulating lipolysis. Cellular signalling 2006,
18(4):401-408.
Omar B, Zmuda-Trzebiatowska E, Manganiello V, Goransson O, Degerman E: Regulation of
AMP-activated protein kinase by cAMP in adipocytes: Roles for phosphodiesterases,
protein kinase B, protein kinase A, Epac and lipolysis. Cellular signalling 2009, 21(5):760766.
He YH, Zhang HQ, Teng JH, Huang LN, Li Y, Sun CH: Involvement of calcium-sensing receptor
in inhibition of lipolysis through intracellular cAMP and calcium pathways in human
adipocytes. Biochemical and biophysical research communications 2011, 404(1):393-399.
Hallenborg P, Feddersen S, Madsen L, Kristiansen K: The tumor suppressors pRB and p53 as
regulators of adipocyte differentiation and function. Expert opinion on therapeutic targets
2009, 13(2):235-246.
Rosen ED, Spiegelman BM: Molecular regulation of adipogenesis. Annu Rev Cell Dev Bi
2000, 16:145-171.
Petersen RK, Madsen L, Pedersen LM, Hallenborg P, Hagland H, Viste K, Doskeland SO,
Kristiansen K: Cyclic AMP (cAMP)-mediated stimulation of adipocyte differentiation
requires the synergistic action of Epac- and cAMP-dependent protein kinase-dependent
processes. Molecular and cellular biology 2008, 28(11):3804-3816.
Zilberfarb V, Siquier K, Strosberg AD, Issad T: Effect of dexamethasone on adipocyte
differentiation markers and tumour necrosis factor-alpha expression in human PAZ6 cells.
Diabetologia 2001, 44(3):377-386.
34
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
Hernandez R, Teruel T, Lorenzo M: Insulin and dexamethasone induce GLUT4 gene
expression in foetal brown adipocytes: synergistic effect through CCAAT/enhancer-binding
protein alpha. Biochemical Journal 2003, 372:617-624.
Shin SM, Kim K, Kim JK, Yoon SR, Choi I, Yang Y: Dexamethasone reverses TGF-betamediated inhibition of primary rat preadipocyte differentiation. FEBS letters 2003, 543(13):25-30.
Cristancho AG, Lazar MA: Forming functional fat: a growing understanding of adipocyte
differentiation. Nature reviews Molecular cell biology 2011, 12(11):722-734.
Madsen L, Kristiansen K: The importance of dietary modulation of cAMP and insulin
signaling in adipose tissue and the development of obesity. Annals of the New York
Academy of Sciences 2010, 1190:1-14.
Silva CM, Sato S, Margolis RN: No time to lose: workshop on circadian rhythms and
metabolic disease. Genes & development 2010, 24(14):1456-1464.
Lin YC, Hsiao TJ, Chen PC: Persistent Rotating Shift-Work Exposure Accelerates
Development of Metabolic Syndrome among Middle-Aged Female Employees: A Five-Year
Follow-Up. Chronobiology international 2009, 26(4):740-755.
Neal JW, Clipstone NA: Calcineurin mediates the calcium-dependent inhibition of adipocyte
differentiation in 3T3-L1 cells. Journal of Biological Chemistry 2002, 277(51):49776-49781.
Zhang P, O'Loughlin L, Brindley DN, Reue K: Regulation of lipin-1 gene expression by
glucocorticoids during adipogenesis. Journal of lipid research 2008, 49(7):1519-1528.
Shimizu-Albergine M, Tsai LC, Patrucco E, Beavo JA: cAMP-specific phosphodiesterases 8A
and 8B, essential regulators of Leydig cell steroidogenesis. Molecular pharmacology 2012,
81(4):556-566.
Li H, Li H, Guo H, Liu F: Cholesterol suppresses adipocytic differentiation of mouse adiposederived stromal cells via PPARgamma2 signaling. Steroids 2013, 78(5):454-461.
Ma T, Liaset B, Hao Q, Petersen RK, Fjaere E, Ngo HT, Lillefosse HH, Ringholm S, Sonne SB,
Treebak JT et al: Sucrose Counteracts the Anti-Inflammatory Effect of Fish Oil in Adipose
Tissue and Increases Obesity Development in Mice. PloS one 2011, 6(6).
Hao Q, Lillefosse HH, Fjaere E, Myrmel LS, Midtbo LK, Jarlsby RH, Ma T, Jia BB, Petersen RK,
Sonne SB et al: High-glycemic index carbohydrates abrogate the antiobesity effect of fish
oil in mice. Am J Physiol-Endoc M 2012, 302(9):E1097-E1112.
Madsen L, Pedersen LM, Liaset B, Ma T, Petersen RK, van den Berg S, Pan J, Muller-Decker K,
Dulsner ED, Kleemann R et al: cAMP-dependent signaling regulates the adipogenic effect of
n-6 polyunsaturated fatty acids. Journal of Biological Chemistry 2008, 283(11):7196-7205.
Konstantinova I, Nikolova G, Ohara-Imaizumi M, Meda P, Kucera T, Zarbalis K, Wurst W,
Nagamatsu S, Lammert E: EphA-Ephrin-A-mediated beta cell communication regulates
insulin secretion from pancreatic islets. Cell 2007, 129(2):359-370.
Grimditch GK, Barnard RJ, Hendricks L, Weitzman D: Peripheral Insulin Sensitivity as
Modified by Diet and Exercise Training. American Journal of Clinical Nutrition 1988,
48(1):38-43.
Chicco A, D'Alessandro ME, Karabatas L, Pastorale C, Basabe JC, Lombardo YB: Muscle lipid
metabolism and insulin secretion are altered in insulin-resistant rats fed a high sucrose
diet. Journal of Nutrition 2003, 133(1):127-133.
Shanik MH, Xu Y, Skrha J, Dankner R, Zick Y, Roth J: Insulin resistance and hyperinsulinemia:
is hyperinsulinemia the cart or the horse? Diabetes care 2008, 31 Suppl 2:S262-268.
Ferreira Mdel R, Lombardo YB, Chicco A: beta-Cell adaptation/dysfunction in an animal
model of dyslipidemia and insulin resistance induced by the chronic administration of a
sucrose-rich diet. Islets 2010, 2(6):367-373.
Renstrom F, Buren J, Svensson M, Eriksson JW: Insulin resistance induced by high glucose
and high insulin precedes insulin receptor substrate 1 protein depletion in human
adipocytes. Metabolism: clinical and experimental 2007, 56(2):190-198.
35
36
8. Annex
37
2
Proteomic analysis of cAMP-mediated signaling during
adipocyte differentiation of 3T3-L1 preadipocytes
3
4
Kamil Borkowksi1, Krzysztow Wrzesinski2, Adelina Rogowska-Wrzesinska2, Karine Audouze3, Jesse
Bakke4, Rasmus Koefoed Petersen1,Fawaz Gaj Haj4, Lise Madsen1,5, Karsten Kristiansen1
1
5
6
7
1 - Department of Biology, University of Copenhagen, Ole Maaløes Vej 5
8
DK-2200 Copenhagen N, Denmark
9
2 –Department of Biochemistry and Molecular Biology, University of Southern Denmark, Campusvej 55
10
DK-5230 Odense M, Denmark
11
12
3-Center for Biological Sequence Analysis, Department of Systems Biology, Technical University of Denmark,
DK-2800 Kgs Lyngby, Denmark
13
4- Departments of Nutrition, University of California Davis, Davis, CA 95616, USA
14
5- National Institute of Nutrition and Seafood Research (NIFES), N-5817 Bergen, Norway
15
16
17
18
19
20
Corresponding authors:
21
22
23
Karsten Kristiansen, Department of Biology, Copenhagen University, Ole Maaløes Vej 5, 2200
København N, Copenhagen, Denmark. Phone: +45 353-24443, E-mail: [email protected]
24
25
Lise Madsen, National Institute of Nutrition and Seafood Research (NIFES), Nordnesboder 2,
26
5005 Bergen, Norway. E-mail [email protected]
27
28
29
30
ABSTRACT
31
Initiation of adipocyte differentiation is promoted by the synergistic action of insulin/insulin-
32
like growth factor and cAMP-dependent signaling. The action of cAMP is mediated via PKA
33
and Epac, where PKA acts by repression of tRho-kinase activity, whereas Epac counteracts
34
the PKA-Rho-kinase dependent reduction in insulin/insulin-like growth factor signaling by
35
restoring and enhancing insulin/insulin-like growth factor signaling. The action of PKA per
36
se in adipocyte differentiation is well described, but detailed knowledge of the Epac-
37
dependent branch and the interplay with PKA in the context of adipocyte differentiation is
38
still scares. In the present study we present a comprehensive study of Epac-mediated
39
processes and their interplay with PKA during the initiation of adipocyte differentiation of
40
3T3-L1 preadipocytes using a combination of proteomics, molecular approaches and
41
bioinformatics. Proteomic analyses revealed 7 proteins regulated uniquely in response to
42
Epac activation, 4 in response to PKA activation, and 11 in response to the combined
43
activation by Epac and PKA. Network analyses indicated that these proteins are involved in
44
pathways of importance for glucose metabolism, inositol metabolism and of calcium-
45
dependent signaling adding a novel facet to our understanding of cAMP-mediated signaling
46
during initiation of adipocyte differentiation.
47
48
INTRODUCTION
49
The development of obesity is related not only to increased fat cell mass, but also increased
50
fat cell number as the result of fat cell differentiation [1]. Much data on preadipocyte
51
differentiation has been acquired from cell culture studies, where 3T3-L1 and 3T3-F442A
52
mouse fibroblasts as well as mouse embryo fibroblasts have been used as models. In the
53
standard differentiation procedure, mouse fibroblasts are treated with high concentrations of
1
54
insulin, glucocorticoid and a cAMP elevating agents such as 3-isobutyl-1-methylxanthine
55
(IBMX) or forskolin [2, 3]. Factors that elevate cellular concentrations of cAMP strongly
56
enhance differentiation. In the cytoplasm, cAMP activates protein kinase A (PKA) and
57
exchange factor directly activated by cAMP (Epac). PKA inactivates Rho kinase by
58
phosphorylation, and this action has been previously shown to be crucial for preadipocyte
59
differentiation [3, 4]. Rho kinase was found to regulate the insulin/ insulin like growth factor
60
receptor (IGFR) signaling pathway via phosphorylation of the IRS protein. When Rho kinase
61
is highly active, it phosphorylates serine residue 612 of IRS-1, which causes inhibition of
62
IGFR signaling. On the other hand, at low activity, Rho kinase induces IGFR signaling by
63
phosphorylation of serine residues 632 and 635 [5, 6]. Epac has been reported to activate the
64
G protein Rap, which induces important changes in cytoskeleton organization and cell
65
adhesion [7]. In 2008, Petersen et al. [3] reported that activation of both of these cAMP
66
targets is required for differentiation of mouse fibroblasts by employing the two cAMP
67
analogues 8-pCPT-2´-O-Me-cAMP (007) and N6-monobutyryl cAMP (MB) which
68
specifically activate Epac or PKA, respectively. Similarly, it was show that synergistic
69
activation of PKA and Epac also is required for adipocyte differentiation of human hMADS
70
cells [8].
71
The aim of the present study was to examine the role of cAMP-mediated signaling via Epac
72
and PKA during adipocyte differentiation of 3T3-L1 preadipocytes using proteomics in
73
combination with molecular approaches and network analyses. The cells were induced to
74
differentiate using 007 or MB or mixture of 007 and MB. Samples were collected at 0, 8, 24
75
and 48 hours of the treatment and at day 8. The samples were submitted to proteomic analysis
76
using two dimensional gel electrophoresis combined with protein identification by mass
77
spectrometry.
78
2
79
MATERIALS AND METHODS
80
Cell culture and differentiation
81
3T3 L1 mouse fibroblasts were obtained from the Eukaryotic Gene Expression and
82
Differentiation Group, Department of Biochemistry and Molecular Biology, University of
83
Southern Denmark. Cells were cultured up to the third passage before start of differentiation.
84
Cells were maintained and induced for differentiation as described (Petersen 2008 [3]). In
85
short 3T3-L1 cells were cultured to confluence in Dulbecco’s Modified Eagle’s Medium
86
(DMEM, Invitrogen™: Cat. No. 41966029) supplemented with 10% calf serum. Two-days
87
postconfluent (designated day 0) cells were induced to differentiate with DMEM (Dulbecco’s
88
modified Eagle’s medium) supplemented with 10% fetal bovine serum (FBS), 1 μM
89
dexamethasone
90
isobutylmethylxanthine (IBMX) (Sigma). For analyses of the roles of PKA and Epac, IBMX
91
was replaced with 200 μM of 8-pCPT-2´-O-Me-cAMP (Biolog) (007 treatment), 100 μM of
92
N6-monobutyryl-cAMP (Biolog) (MB treatment) or 200 μM of 8-pCPT-2´-O-Me-cAMP +
93
100μM of N6-monobutyryl-cAMP (007MB treatment). After 48 h media were replaced with
94
DMEM supplemented with 10% FBS and 1 μg/ml insulin. The cells were subsequently
95
reefed every 48 h with DMEM supplemented with 10% fetal bovine serum. The cells were
96
cultured up to day 8 of differentiation. At day 8 triglycerides presents in differentiated cells
97
were visualized by Oil-Red-O-staining [9].
(dex)(Sigma)
1
μg/ml
insulin
(ins)
(Sigma),
and
0.5
mM
98
99
Samples collection for two dimensional gel analysis
100
Cells were washed three times with HANK’s buffer and collected by scraping. After
101
collection cells were immediately frozen in liquid nitrogen. The cell lysates were prepared by
3
102
addition of IPG lysis buffer and shaking overnight. Each sample was centrifuged at 1500 rcf
103
for 15min at 10oC and the supernatants were transferred to new tubes. Protein concentration
104
was determinate by the Bradford method.
105
106
Two dimensional gel electrophoresis
107
For the first dimension 18cm IPG strips, covering a pH gradient from 4 to 7 (IPG 4-7) were
108
used (Ge Healthcare Cat. No. 17-1233-01). In-gel rehydration was performed in two steps.
109
First step strips were rehydrated overnight in 200 μl of sample diluted in IPG lysis buffer
110
(300μg of protein per gel were loaded). Next, strips were rehydrated for 6 hour in 100 μl of
111
lysis buffer. Isoelectric focusing was performed using a linearly increasing profile: 0 V to
112
600 V for 2:15 h, f600 V to 3500 for 8 h, and 3500 V for 9:25 h. After focusing the strips
113
were incubated in equilibration buffer for 15 min. and then frozen at -80°C. After thawing
114
gels were incubated in equilibration buffer for 15 min. The Second dimension SDS PAGE
115
was performed using a vertical electrophoresis system Protean II™ (BioRad) and 12.5%
116
(w/v) laboratory made acrylamide gels (acrylamide: N,N’-ethylene-bis-acrylamide ratio
117
200:1). The gels were run overnight at 20oC. 6 mA per gel was applied for the first 2 hours of
118
electrophoresis, then 12 mA per gel. The running buffer contained 0.67% Tris-Base, 1.44%
119
of glycine and 1% SDS. Additionally, buffer recirculation was applied. The gels were stained
120
with Sypro Ruby (Invitrogen™ Cat. No. S21900) and scanned on the Typhoon scanner with
121
an excitation wave length of 488nm and emission filter at 610nm.
122
123
124
4
125
Gels images analysis with DECODON Delta 2D, version 3.6 software
126
Gel images were analyzed using the DECODON Delta 2D software, version 3.6. Differences
127
were considered as significant when the average intensity ratio of a spot from the
128
experimental treatment group compared to the average intensity ratio of a spot of the control
129
group was bigger than 1.5 or smaller than 0.67. All found differences were evaluated by a t-
130
test with a level of confidence p<0.05.
131
132
Mass spectrometry
133
In gel digestion. Spots exhibiting significant changes were cut out from the gel. The gel
134
plugs were washed with 100 μl of sterile water for 5 min. Water was removed and the gel
135
plugs were washed with 100 μl of 100% acetonitrile (ACN) for 20 min. ACN was removed
136
and gel plugs were dried in a Speed Vac. Dry gel plugs were rehydrated with 10μl of trypsin
137
solution (10ng/ml, dissolved in 50 mM NH4HCO3, pH 7.8) for 20 min on the ice.
138
Unabsorbed trypsin was removed and the gel plugs were covered with 20 μl of 50 mM
139
NH4HCO3, pH 7.8. Digestion was carried out overnight at 37°C.
140
Target loading. Desalting and concentration of peptides mixtures were done using laboratory
141
made microcolumns, prepared from GELoader micropipette tips, packed with POROS R2.
142
The columns were pre-washed with 10 μl of 0.1% trifluoroacetic acid (TFA). 10 μl of
143
peptides mixture from the digested protein were loaded onto microcolumns. The columns
144
were then washed with 10 μl of 0.1% TFA. Peptides were released from the column with 3μl
145
of CHCA diluted in 70% ACN/0.1% TFA and placed on the MALDI plate. Trypsin digested
146
β-galactosidase was used as a standard for instrument calibration and was placed on every
147
third spot on the MALDI plate.
5
148
Sample analysis. A 4800 MALDI TOF/TOF Analyzer from Applied Biosystems was used
149
for recording of positive ion MS and MS/Ms spectra. For Ms analysis the mass range 700 to
150
3500 Da was selected. Total laser shots number was set to 800 and a fixed laser intensity of
151
3100 was selected. For Ms/Ms analysis, the total number of shots was set to 1280. For each
152
Ms spectrum at least three of the most intensive peaks were selected for Ms/Ms analysis. MS
153
and associated MS/MS data were combined into a single mass list in the mgf. file format
154
using an in-house developed script developed by Jakob Bunkenborg, University of Southern
155
Denmark. Mass lists were searched in Swiss-Prot database using an in-house MASCOT
156
server. The MASCOT server settings were: maximal number of missed cleavages: 1;
157
significant threshold p<0.05; MS mass accuracy: 50 ppm; MS/MS 0.6 Da: partial
158
modification: methionine oxidation.
159
160
Data analysis
161
The Ingenuity pathway analysis software was used for network analysis.
162
To perform gene enrichment and facilitate further data integration, the selected mouse genes
163
were mapped to their human homologs (EntrezGene identifiers). A total of 19 genes were
164
identified. The list of genes was expanded by including their known first-order protein-
165
protein interaction partners using a high confidence human interactome containing 428,429
166
unique protein-protein interactions consisting of refined experimental proteomics data (Lage).
167
This step result in a second gene set containing 81 genes. Both gene sets were investigated
168
for functional annotation based on the three Gene Ontology categories (molecular function,
169
biological process and cellular components). We further explored pathways integrating
170
KEGG pathway information in both sets. All p-values obtained were calculated using
171
hypergeometric testing, and were corrected for multiple testing with Bonferroni correction.
6
172
The significance cutoff for the corrected p-values was set to 0.05. Cytoscape open source
173
software was used for data visualization.
174
175
Quantitative PCR
176
For mRNA isolation cells were harvested in TRIsol reagent (Invitrogen) followed by
177
chloroform/isopropanol extraction and precipitation. cDNA synthesis was performed using
178
the First Strand cDNA Synthesis Kit (Ferments). Quantitative PCR reactions were performed
179
using SYBR green assay with FAST qPCR MasterMix Plus Kit (Eurogenetec) and a
180
Mx3000P™ Real-Time PCR System from Stratagene. TATA box binding protein was used
181
as a reference gene. Data were analyzed using the MxPro software from Stratagene.
182
183
Short hairpin RNA transfection
184
shRNA against Isyna1 was transduced into 3T3-L1 cells using the lentivirus pGIPZ vector.
185
293FT cells were used for virus production. In short, 80% confluent 293FT cells were using
186
Lipofectamine 2000 (Invitrogen) transfected with a pGIPZ vector containing shRNA or with
187
a pGIPZ empty vector, packaging plasmid psPAX2 and envelope plasmid 2.G. Two days
188
after transfection, the medium containing viruses was collected and added to 40% confluent
189
3T3-L1 cells. Two days after transfection, the cells were selected using 2µg/ml puromycin
190
(InviveGen). After selection cells were maintained in medium containing 1µg/ml puromycin.
191
192
193
7
194
Western blot analysis
195
The Mini Format 1-D Electrophoresis Systems (Bio-Rad) and homemade precast 8%
196
acrylamide gels were used for western blot analyses. The Chemiluminescence Detection Kit
197
for HRP (Biological Industries) and a Fusion FX5 Chemiluminescence imaging system
198
(Montreal Biotech. Inn) were used. The acquired data was quantified using the ImageJ open
199
source software.
200
201
RESULTS
202
Two dimensional gel analysis
203
In order to determine protein expression and post translational alterations in 3T3-L1 cells
204
following treatment with cyclic AMP (cAMP) analogues targeting either PKA or Epac, we
205
analyzed cells during different time points of differentiation using two dimensional (2D) gel
206
electrophoresis. As described in Materials and Methods, two cAMP analogues were used to
207
replace IBMX used as cAMP-elevating agents in the standard differentiation protocol.
208
Results from 2D gels were analyzed to identify alterations in protein expression and/or
209
modification. The most conspicuous changes in protein expression during the differentiation
210
were observed at 48 hours (graphs summarizing expression of all proteins, where alterations
211
were observed are presented in Figure S1). In order to obtain statistically significant results,
212
we extensively analyzed the 48 hour time point in terms of differences between individual
213
treatments and dexamethasone plus insulin stimulation (reference treatment) (see Materials
214
and Methods). The experimental design together with a list of proteins exhibiting significant
215
changes at the 48 hour time point are presented in Figure 1.
8
216
Twenty-five spots were significantly different from the reference treatment at 48 hours.
217
Nineteen proteins were identified from the original 25 spots and two spots remained
218
unidentified. Eight proteins (Hydroxymethylglutaryl-CoA synthase, Pyruvate carboxylase,
219
Phosphoglycerate mutase, Transaldolase, Laminin subunit gamma-1, Endoplasmin, Ras-
220
related protein Rap and Guanine deaminase) reached a 2 fold up-regulation and one protein,
221
Importin subunit beta, was more than 2-folds down-regulated (see table 1). Four proteins,
222
Inositol-3-phosphate synthase 1, Phosphoglycerate mutase, Pyruvate carboxylase and
223
Laminin subunit gamma-1 were identified in more than one spot. The Spots identified as
224
inositol-3-phosphate synthase 1, phosphoglycerate mutase 1/2, guanine deaminase,
225
hydroxymethylglutaryl-CoA synthase and hippocalcin-like protein 1 are shown in the figure
226
2. Compared with the reference treatment, Epac activation induced the change of 10 proteins
227
(8 up and 2 down – regulations), from which 9 was also observed in companied Epac and
228
PKA stimulation (Figure 3). In addition, activation of PKA changed intensity of 7 proteins (6
229
up and 1 down – regulations), from which 5 were common with combined Epac and PKA
230
activation. Finally in response to the simultaneous activation of Epac and PKA we observed a
231
significant change of 22 different proteins (17 up and 5 down – regulations) from which 11
232
were unique for the combined stimulation of Epac and PKA. Surprisingly, 3 protein changes
233
were common for both Epac and PKA activation. All differences have been summarized in
234
the protein identification table (Table 1).
235
236
Analysis of mRNA expression
237
We examined mRNA expression of proteins identified by proteomics using quantitative PCR
238
to investigate whether or not observed changes involved transcriptional regulation. Levels of
239
mRNA encoding 9 proteins were significantly changed during the differentiation process
9
240
(Figure 4). The mRNA level of Inositol-3-phosphate synthase 1 increased during
241
differentiation in response to 007MB treatment and reached an approximate 2.5-fold change
242
at 48 hours. In mature adipocytes (Day 8) the mRNA level of Inositol-3-phosphate synthase 1
243
was half that of preadipocytes (day 0). At 48 hours, the mRNA level of Inositol-3-phosphate
244
synthase 1 was significantly different from the reference treatment for both 007MB and 007
245
treatments. However, the fold change of 007MB treatment was 2 times larger than the fold
246
change of 007 treatment alone. The mRNA levels of three proteins Hydroxymethylglutaryl-
247
CoA
248
dioxygenase 1 increased during differentiation and appeared to be significantly different in
249
mature adipocytes (day 8 007MB treatment) compared to both day 0 and the reference
250
treatment. The mRNA level of Hydroxymethylglutaryl-CoA synthase and Phosphoglycerate
251
mutase were significantly up-regulated at the 48 hour time point in case of both 007 and
252
007MB treatments compared to the control treatment. Procollagen-lysine, 2-oxoglutarate 5-
253
dioxygenase 1 mRNA levels were significantly increased by every treatment together with
254
the reference at the 24 hour and 48 hour time point compared to day 0. However, no
255
significant differences at those two time points were observed between cAMP analogues and
256
the reference. A similar situation was observed in the case of Laminin subunit gamma-1,
257
except that no significant difference was observed between the treatments at day 8. The
258
mRNA level of three proteins, Lysosomal alpha-glucosidase, Beta-hexosaminidase subunit
259
alpha and Hippocalcin-like protein 1 (for proteomic data see figure 2) were down-regulated
260
during differentiation. In all three cases, the mRNA levels in cells treated with 007MB were
261
significantly down-regulated compared to cells receiving the reference treatment at the 48
262
hour and the day 8 time point. In mature adipocytes, the mRNA levels of all three proteins
263
were also significantly down-regulated compared to preadipocytes (not shown). The mRNA
synthase,
Phosphoglycerate
mutase
and
Procollagen-lysine,2-oxoglutarate
5-
10
264
level of Importin subunit beta-1 was significantly different only in cells treated with 007MB
265
compared to the other treatments at day 8.
266
267
Data analysis
268
Gene ontology and protein-protein interaction analysis
269
In order to obtain an overview of the functions of identified proteins, we performed gene
270
ontology (GO) as well as Reactome Main Pathway analyses (table S1). These analyses
271
indicated
272
(G6PD;GAA;HMGCS1;ISYNA1;PGAM1;PGAM2;TALDO1) as well as carbohydrate
273
catabolic processes (G6PD;GAA;PGAM1;PGAM2;TALDO1). Detailed involvement of
274
identified proteins described above in carbohydrate (particularly glucose) catabolic processes
275
is shown in the figure 5. Most of the proteins, G6PD, ISYNA1, PGAM1, PGAM2, TALDO1,
276
were regulated by the combined action of Epac and PKA (see table S1); whereas Lyag and
277
HMGCS1 were changed in response to Epac activation as well as simultaneous activation of
278
both Epac and PKA. Due to the small number of identified proteins, we performed gene
279
enrichment analyses together with protein-protein interaction network analyses (Figure 6 and
280
Table S2). Protein-protein interaction analyses were restricted to highly significant
281
interactions. Nineteen proteins were recognized and submitted to the network analyses.
282
Fourteen of them formed an interaction network. The entire network consists of 81 proteins.
283
Six proteins, identified by proteomics,
284
dienoyl-CoA isomerase, Cofilin, Beta-hexosaminidase subunit alpha, Endoplasmin and
285
Procollagen-lysine,2-oxoglutarate 5-dioxygenase 1 exhibit direct interactions, whereas
286
Endoplasmin and Ras-related protein Rap have a common interaction protein – von
287
Willebrand factor (VWF). GO analysis for gene enrichment analysis indicated a possible role
relationships
of
identified
proteins
with
alcohol
metabolic
processes
Importin subunit beta-1, Delta(3,5)-Delta(2,4)-
11
288
of the identified proteins in hexose catabolic processes, monosaccharide catabolic processes,
289
alcohol catabolic processes, small GTPase mediated signal transduction, carbohydrate
290
metabolic processes, NADP metabolic processes, cellular carbohydrate catabolic process and
291
glucose catabolic processes. The Reactom Main Pathway analysis (Figure 6 and Table S2)
292
indicated a possible connection of the identified proteins in relation to integrin cell surface
293
interactions, metabolism of carbohydrates and homeostasis.
294
295
Ingenuity pathway analysis
296
We used the Ingenuity pathways analysis™ software to show indirect relationships between
297
identified proteins as well as to illuminate key regulators of identified proteins. Relationships
298
are presented in Figure 7. The identified proteins created two independent networks. TGFB1
299
is a central part of first network and is involved in indirect interactions with 3 proteins the
300
exhibited changes in expression in response to a combined activation of Epac and PKA
301
(Glucose-6-phosphate 1-dehydrogenase, Importin subunit beta-1, Ras-related protein Rap)
302
and two proteins changed by Epac activation alone (Endoplasmin and Procollagen-lysine,2-
303
oxoglutarate 5-dioxygenase 1).
304
common binding partner - inhibitor of kappaB kinase epsilon (IKBKE). Calcium ions are
305
central regulators of the second network. It regulates directly or indirectly three proteins
306
down-regulated during the differentiation process. The proteomics analyses demonstrated that
307
the expression of two of them (beta-hexosaminidase subunit alpha and lysosomal alpha-
308
glucosidase) was altered in response to Epac activation (see also figure 5). Platelet-derived
309
growth factor (PDGF) has been reported to regulate Cofillin (changed by PKA activation in
310
our study) as well as Guanine deaminase and Hydroxymethylglutaryl-CoA synthase, which
311
were up-regulated by Epac activation.
Inositol-3-phosphate synthase 1 and Transaldolase have
12
312
Myo inositol and 3T3L1 cells differentiation
313
We further investigated the change in myo inositol synthase expression observed at the 48
314
hour time point of 3T3-L1 cells differentiation. Western blot analysis showed two species of
315
the protein present in the cells, clearly separated by SDS PAGE (see Figure 8). The
316
specificity of the antibody was confirmed by shRNA knockdown of the Ino1 mRNA (see
317
Figure 8). While the intensity of lower band did not change during first 48 hours of
318
differentiation, the intensity of the upper band increased markedly during the first 48 hours of
319
differentiation in response to the combined action of 007 and MB. Replacement of cAMP
320
analogues with IBMX gave a similar effect with an even more pronounce increase in the
321
intensity of the upper band at the 24 hour time point. Total Ino1 protein was more than 2
322
times down-regulated in mature adipocyte compared to preadipocytes.
323
Western blot analysis of different mouse fat depots showed that most of the Ino1 protein
324
migrated as the upper band form (Figure 8). High fat feeding had no effect on abundance of
325
Ino1 protein in epididymal and inguinal fat depots (Figure S2). Regular Dulbecco's Modified
326
Eagle Medium contains 40µM of inositol. Mammalian cells can utilize extracellular inositol
327
employing two classes of transporters both of which use sodium or protons to co-transport
328
inositol. To investigate the ability of 3T3-L1 cells to utilize inositol from the culture medium,
329
we measured mRNA expression of inositol transporters. The mRNA of solute carrier family 5
330
member 3 (SLC5a3) was detected in both preadipocytes and mature adipocytes. Other
331
inositol transporters were not detected. During the differentiation, the mRNA expression of
332
SLC5a3 was strongly down-regulated when IBMX was present in differentiation medium.
333
334
13
335
DISCUSION
336
Since its discovery, Epac has been implicated in numerous cellular processes (for review see
337
[10]). We have shown that activation of Epac is necessary for cAMP-induced acceleration of
338
adipocyte differentiation of 3T3-L1 preadipocytes, mouse embryo fibroblast [3] and human
339
hMADS cells [8]. Here we have examined protein changes induced by cAMP-mediated
340
signaling driven by Epac, PKA or both factors during differentiation of 3T3-L1
341
preadipocytes. .
342
It has been reported that cAMP elevation in preadipocytes dramatically increases glucose
343
uptake and intracellular glucose levels [11]. A number of identified proteins exhibiting
344
changes in abundance in response to the combined activation of Epac and PKA were
345
enzymes involved in glucose metabolism (Figure 5 and table S1 and S2). Regulation of three
346
distinct pathways could be observed. Glucose 6-phosphate dehydrogenase is the first enzyme
347
in the pentose phosphate pathway providing precursors for nucleic acids synthesis. Glucose
348
6-phosphate dehydrogenase also plays a role in production of NADPH, which is important
349
for fatty acids and cholesterol synthesis. Regulation of the pentose phosphate pathway during
350
the early stage of 3T3-L1 differentiation has previously been reported [12]. Inhibition of
351
glucose 6-phosphate dehydrogenase inhibits differentiation of 3T3-L1 preadipocytes and
352
lipid accumulation [12, 13]. Transaldolase, the enzyme that links the pentose phosphate
353
pathway and glycolysis was in the present study also found to be up-regulated at the protein
354
level.
355
expression of this enzyme during first 4 days of differentiation and an increase of its
356
expression in the later phase of differentiation [14]. Transaldolase directs carbohydrate
357
metabolites derived from glucose to the glycolysis pathway (see figure 5). An increase in the
358
activity of this enzyme may be needed for switch from nucleic acids synthesis (needed during
359
the clonal expansion phase) to glycolysis and subsequent to de novo fatty acids synthesis.
A microarray study conducted on 3T3-L1 cells showed a decrease in mRNA
14
360
However, in contrast to the microarray study, we did not observed changes in transaldolase
361
mRNA levels. Phosphoglycerate mutase, an enzyme catalyzing eight step of glycolysis was
362
also found to be up-regulated in the present study both at protein and at the mRNA level (see
363
figure 2 and 4), but has not so far been reported to play a role in adipogenesis.
364
Phosphoglycerate mutase was identified on 2D gels as two different spots displaying different
365
isoelectric point, suggesting that two different protein species may be involved in 3T3-L1
366
cells differentiation. Inositol-3-phosphate synthase 1 (Ino1) is another protein involved in
367
glucose metabolism. Increase in Ino1 expression was observed at both the protein and the
368
mRNA level (see figure 2 and 4) in response to combine action of Epac and PKA. Ino1
369
converts glucose 6-phosphate into myo inositol, a precursor of all phospho inositol species
370
covalently bound to phospholipids. An increase in Ino1 mRNA levels during adipocyte
371
differentiation was previously reported [14], however the function of this enzyme as well as
372
myo inositol during 3T3-L1 preadipocyte differentiation has not been studied. Western blot
373
analysis of Ino1 showed two species of the protein present in the cells (Figure 8A and B). The
374
protein species represented by upper band on the western blot seems less abundant in
375
preadipocytes than in mature adipocytes. Band of the same molecular mass constitutes the
376
majority of the Ino1 protein in mouse fat tissue (see figure 8D), suggesting the biological
377
importance of this protein species. The level of the lover band present in preadipocytes did
378
not changed during the differentiation process. The possibility that the upper band
379
represented a phosphorylated state of the Ino1 protein is unlikely, as phosphatase treatment
380
did not alter the intensity of the upper band (data not shown). Cell can obtain inositol in two
381
ways. One is de novo synthesis from glucose and the second is by transport across the plasma
382
membrane from the extracellular space. Mammalian cells express two types of inositol
383
transporters, Slc5a3 and Slc5a11, which both are sodium co-transporters, and Slc5a13 which
384
is a H+ co-transporter [15]. 3T3-L1 fibroblasts and mature adipocytes express only the
15
385
Slc5a3 transporter, the expression of which is strongly down-regulated during the first 48
386
hours of differentiation in response to IBMX treatment (figure 8E). Down-regulation of
387
inositol transport together with up-regulation of inositol producing enzymes seem to be
388
contradictory and suggest that further study on the role of inositol in adipocyte differentiation
389
need to be conducted.
390
Epac has not been observed to transcriptionally regulate a broad spectrum of genes,
391
suggesting that its action may be mostly related to the posttranscriptional regulations. In our
392
proteomic analysis, 7 different proteins were found to be regulated by Epac (see table 1). This
393
regulation did not required activation of PKA. However, only one protein, 3-hydroxy-3-
394
methylglutaryl-CoA synthase 1 (HMGCS1) was observed to be regulated by Epac at the level
395
of transcription level (see figure 4). Two proteins were observed to be down-regulated at the
396
protein level in response to Epac activation, Beta-hexosaminidase subunit alpha (Beta-
397
hexosaminidase) and Lysosomal alpha-glucosidase (Lyag), were observed to be down-
398
regulated at the mRNA level only when Epac signaling was supported by PKA signaling (see
399
figure 4), suggesting a different regulation at the transcriptional, translational and
400
posttranslational level by Epac alone or the combination with PKA.
401
Epac has been reported to regulate calcium signaling in cardiac muscle cells and in pancreatic
402
β-cell islets [16, 17]. Ingenuity pathway analyses revealed calcium ions as one of the central
403
regulators of the network created by the identified proteins (see figure 7) Two of the proteins
404
in the network, beta-hexosaminidase and Lyag, were observed to be regulated by Epac alone.
405
Lyag is one of the glycogen degrading enzymes, shown to play a role in insulin secretion by
406
β-cells and to be regulated by calcium ions [18, 19]. Hexosaminidase is an enzyme taking
407
part in degradation of glycoconjugates such as glycoproteins or glycolipids. Both enzymes
408
are located in lysosomes and there role in adipogenesis has never been tested. Adipocyte
409
differentiation has been shown to be inhibited by high intracellular level of calcium in several
16
410
cells models [20, 21] and beta-hexosaminidase and Lyag may be potential mediators of
411
calcium-mediated inhibition of adipocyte differentiation.
412
413
414
Acknowledgments
415
This work was supported by grants from the Danish Natural Research Council, The Novo
416
Nordisk Foundation, the Carlsberg Foundation and the Danish council for Strategic Research,
417
Grant No.: 09-065171.
418
419
420
REFERENCES
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
[1] Tang QQ, Lane MD. Adipogenesis: from stem cell to adipocyte. Annual review of biochemistry
2012;81:715-36.
[2] Rubin CS, Hirsch A, Fung C, Rosen OM. Development of hormone receptors and hormonal
responsiveness in vitro. Insulin receptors and insulin sensitivity in the preadipocyte and adipocyte
forms of 3T3-L1 cells. The Journal of biological chemistry 1978;253:7570-8.
[3] Petersen RK, Madsen L, Pedersen LM, Hallenborg P, Hagland H, Viste K, et al. Cyclic AMP (cAMP)mediated stimulation of adipocyte differentiation requires the synergistic action of Epac- and cAMPdependent protein kinase-dependent processes. Molecular and cellular biology 2008;28:3804-16.
[4] Madsen L, Kristiansen K. The importance of dietary modulation of cAMP and insulin signaling in
adipose tissue and the development of obesity. Annals of the New York Academy of Sciences
2010;1190:1-14.
[5] Furukawa N, Ongusaha P, Jahng WJ, Araki K, Choi CS, Kim HJ, et al. Role of Rho-kinase in
regulation of insulin action and glucose homeostasis. Cell metabolism 2005;2:119-29.
[6] Begum N, Sandu OA, Duddy N. Negative regulation of rho signaling by insulin and its impact on
actin cytoskeleton organization in vascular smooth muscle cells: role of nitric oxide and cyclic
guanosine monophosphate signaling pathways. Diabetes 2002;51:2256-63.
[7] Roscioni SS, Elzinga CR, Schmidt M. Epac: effectors and biological functions. NaunynSchmiedeberg's archives of pharmacology 2008;377:345-57.
[8] Jia B, Madsen L, Petersen RK, Techer N, Kopperud R, Ma T, et al. Activation of protein kinase A
and exchange protein directly activated by cAMP promotes adipocyte differentiation of human
mesenchymal stem cells. PloS one 2012;7:e34114.
17
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
[9] Hansen JB, Petersen RK, Larsen BM, Bartkova J, Alsner J, Kristiansen K. Activation of peroxisome
proliferator-activated receptor gamma bypasses the function of the retinoblastoma protein in
adipocyte differentiation. The Journal of biological chemistry 1999;274:2386-93.
[10] Gloerich M, Bos JL. Epac: defining a new mechanism for cAMP action. Annual review of
pharmacology and toxicology 2010;50:355-75.
[11] Cornelius P, Marlowe M, Call K, Pekala PH. Regulation of Glucose-Transport as Well as Glucose
Transporter and Immediate Early Gene-Expression in 3t3-L1 Preadipocytes by 8-Bromo-Camp.
Journal of cellular physiology 1991;146:298-308.
[12] Park J, Rho HK, Kim KH, Choe SS, Lee YS, Kim JB. Overexpression of glucose-6-phosphate
dehydrogenase is associated with lipid dysregulation and insulin resistance in obesity. Molecular and
cellular biology 2005;25:5146-57.
[13] Shantz LM, Talalay P, Gordon GB. Mechanism of inhibition of growth of 3T3-L1 fibroblasts and
their differentiation to adipocytes by dehydroepiandrosterone and related steroids: role of glucose6-phosphate dehydrogenase. Proceedings of the National Academy of Sciences of the United States
of America 1989;86:3852-6.
[14] Hackl H, Burkard TR, Sturn A, Rubio R, Schleiffer A, Tian S, et al. Molecular processes during fat
cell development revealed by gene expression profiling and functional annotation. Genome biology
2005;6:R108.
[15] Fu H, Li B, Hertz L, Peng L. Contributions in astrocytes of SMIT1/2 and HMIT to myo-inositol
uptake at different concentrations and pH. Neurochemistry international 2012;61:187-94.
[16] Ruiz-Hurtado G, Morel E, Dominguez-Rodriguez A, Llach A, Lezoualc'h F, Benitah JP, et al. Epac in
cardiac calcium signaling. Journal of molecular and cellular cardiology 2012.
[17] Kang G, Joseph JW, Chepurny OG, Monaco M, Wheeler MB, Bos JL, et al. Epac-selective cAMP
analog 8-pCPT-2'-O-Me-cAMP as a stimulus for Ca2+-induced Ca2+ release and exocytosis in
pancreatic beta-cells. The Journal of biological chemistry 2003;278:8279-85.
[18] Salehi A, Mosen H, Lundquist I. Insulin release transduction mechanism through acid glucan 1,4alpha-glucosidase activation is Ca2+ regulated. Am J Physiol-Endoc M 1998;274:E459-E68.
[19] Salehi A, Henningsson R, Mosen H, Ostenson CG, Efendic S, Lundquist I. Dysfunction of the islet
lysosomal system conveys impairment of glucose-induced insulin release in the diabetic GK rat.
Endocrinology 1999;140:3045-53.
[20] Jensen B, Farach-Carson MC, Kenaley E, Akanbi KA. High extracellular calcium attenuates
adipogenesis in 3T3-L1 preadipocytes. Experimental cell research 2004;301:280-92.
[21] Ntambi JM, Takova T. Role of Ca2+ in the early stages of murine adipocyte differentiation as
evidenced by calcium mobilizing agents. Differentiation; research in biological diversity 1996;60:1518.
478
479
480
481
482
18
483
Figure legends
484
485
Figure 1. Experimental design and identification of proteins significantly changed at the 48
486
hour time point of 3T3-L1 adipocyte differentiation in response to cAMP signaling via the
487
Epac, the PKA branch or both. The protein identification map shows proteins changed by
488
activation of Epac or PKA and by the combined action of both factors.
489
490
Figure 2. Changes of spots identified as ISYNA1 - Inositol-3-phosphate synthase 1, PGMA1/2
491
-
492
Hydroxymethylglutaryl-CoA synthase, HPCL1 - Hippocalcin-like protein 1 in 48hours time
493
point of 3T3L1 cells differentiation. A – Representative fragments of the gels with the spots
494
of interest. B – Quantification of the intensity of the spots of interest. * - intensity
495
significantly different (p<0.05) from the reference treatment.
Phosphoglycerate
mutase
1/2,
GUAD
-
Guanine
deaminase,
HMCS1
-
496
497
Figure 3. Venn diagram of up- or down-regulated proteins in reposnse to activation of Epac,
498
PKA or both Epac and PKA at the 48 hour time point of 3T3-L1 adipocyte differentiation.
499
500
Figure 4. Quantitative PCR measurement of mRNA expression of nine genes identified as
501
changed by proteomics analysis. * - intensity significantly different (p<0.05) from the
502
reference
503
Hydroxymethylglutaryl-CoA synthase, Lyag - Lysosomal alpha-glucosidase, Pgam1 -
treatment.
Isyna1
-
Inositol-3-phosphate
synthase
1,
Hmgcs1
-
19
504
Phosphoglycerate mutase 1, Lamc1 - Laminin subunit gamma-1, Hpcl1 - Hippocalcin-like
505
protein 1, Kpnb1 - Importin subunit beta-1, Plod1 - Procollagen-lysine,2-oxoglutarate 5-
506
dioxygenase 1, Hexa - Beta-hexosaminidase subunit alpha.
507
508
Figure 5. The role of identified proteins in pathways of glucose metabolism, cholesterol
509
biosynthesis and calcium signaling pathway.
510
511
Figure 6. Protein - protein interaction network between proteins identified as up- or down-
512
regulated due to stimulation of Epac, PKA or both Epac and PKA at the 48 hour time point of
513
3T3L1 adipocyte differentiation. Blue nodes represent proteins identified as changed in one
514
or more of the treatments when compared to the reference treatment, at 48 hour time
515
point of 3T3-L1 adipocyte differentiation. Green nodes represent interacting proteins.
516
Connections represent physical interactions between proteins. Arrows indicate up- or down-
517
regulation. Labels below nodes represent treatment group where the difference has been
518
observed.
519
520
Figure 8. Impact of cAMP on the presence of Ino1 species during 3T3-L1 adipocyte
521
differentiation. Western blot analysis of Ino1 protein during 3T3-L1 differentiation with
522
different cAMP analogues (A) or with IBMX (D). B) Western blot analysis of 3T3L1 cells with
523
shRNA knockdown of Ino1. C) Western blot analysis of Ino1 protein in 3T3-L1 cells and
524
mouse epididymal white adipose tissue. E) Quantitative PCR analysis of Slc5a3 mRNA levels
20
525
during 3T3-L1 adipocyte differentiation with or without IBMX. DI- Dexamethasone, insulin.
526
DMI- Dexamethasone, insulin and IBMX.
527
528
Figure S1. Proteins level changes during 3T3L1 cells differentiation caused by stimulation of
529
Epac, PKA, or both Epac and PKA.
530
531
Figure S2. Ino1 protein level and modification state in subcutaneous and epidydymal fat
532
under different feeding conditions.
533
534
21
535
536
Figure 1
22
537
538
Figure 2
539
540
541
542
543
544
545
546
23
547
548
549
Figure 3
550
551
552
553
554
555
556
557
558
559
560
561
562
24
563
564
Table 1. Identification table for proteins identified as up or down due to stimulation of Epac, PKA or both Epak
and PKA in 48hours time point of 3T3L1 cells differentiation. Fold change is shown only for significant changes.
565
566
567
568
569
570
571
572
573
574
575
576
577
25
578
579
580
Figure 4
581
582
583
584
585
586
587
588
589
590
591
592
593
26
594
595
596
Figure 5
597
27
598
599
Figure 6
600
601
602
603
604
605
606
607
608
609
610
611
28
612
613
Figure 7
29
614
615
Figure 8
616
617
618
619
620
621
622
623
624
625
626
627
628
629
30
630
SUPPLEMENT
631
632
633
Figure S1
634
635
636
637
31
638
639
Figure S2
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
32
660
661
662
Table S1. Gene Ontology and Reactome Main Pathways analysis of proteins identified as up or down-regulated
due to stimulation of Epac, PKA or both Epak and PKA in 48hours time point of 3T3L1 cells differentiation.
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
33
679
680
681
682
Table S2. Gene enrichment analysis of proteins identified as up or down-regulated due to stimulation of Epac,
PKA or both Epak and PKA in 48hours time point of 3T3L1 cells differentiation. Identified proteins together
with their interaction partners have been submitted to Gene Ontology, KEGG pathways and Reactome Main
Pathways analysis.
683
684
685
686
687
688
34
689
1
2
3
4
5
6
7
8
9
10
11
12
Perturbation of first phase insulin secretion in mice
with pancreas-specific ablation of Ptpn1 attenuates
sucrose induced peripheral insulin resistance
Kamil Borkowski1, Ahmed Bettaieb2, Jesse Bakke2, Yannan Xi2, Lene Secher Myrmel1,3, Fawaz Haj2,
Lise Madsen1,3, Karsten Kristiansen1
1 - Department of Biology, University of Copenhagen, DK-2200 Copenhagen, Denmark
2 - Departments of Nutrition, University of California Davis, Davis, CA 95616, USA
3 - National Institute of Nutrition and Seafood Research (NIFES), N-5817 Bergen, Norway
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Corresponding authors:
Karsten Kristiansen, Department of Biology, Copenhagen University, Ole Maaløes Vej 5,
2200 København N, Copenhagen, Denmark. Phone: +45 353-24443, E-mail: [email protected]
Lise Madsen, National Institute of Nutrition and Seafood Research (NIFES), Nordnesboder 2,
5005 Bergen, Norway. E-mail [email protected]
35
ABSTRACT
36
37
A considerable body of evidence indicates that sucrose potentiates the obesogenic effect of a
38
high fat diet in a dose-dependent manner and induces insulin resistance. We have previously
39
suggested that the obesity-promoting effect of sucrose at least in part is mediated by insulin.
40
Pharmacological inhibition of insulin secretion diminishes the effect of elevated sucrose
41
content in high fat diets on body fat mass accumulation. Protein tyrosine phosphatase 1 B was
42
recently shown to positively regulate insulin secretion. In the current study we have
43
investigated the effect of pancreas specific Ptpn1 deletion on the ability of sucrose to increase
44
fat mass and accentuate insulin resistance. We show that pancreas specific deletion of Ptpn1
45
did not influence fed or fasted insulin levels however, mice with pancreas-specific deletion of
46
Ptpn1 exhibited decreased sucrose-induced first phase insulin secretion in the context of a
47
high fat diet. Interesting, pancreas specific deletion of Ptpn1 attenuated peripheral insulin
48
resistance in mice fed a high fat-high sucrose diet.
49
50
INTRODUCTION
51
52
Animals studies have demonstrated that the carbohydrate:protein ratio determines the
53
obesogenic effect of high fat diets. This is supported by research showing that weight gain is
54
increased when isocaloric high fat diets are rich in sucrose or high-glycemic index
55
carbohydrates rather than in protein [1-3]. Insulin signaling is crucially involved in fat cells
56
differentiation [4] as well as an important regulator of metabolism and energy expenditure [5,
57
6]. Defects in insulin signaling have been associated with the metabolic syndrome and
1
58
obesity [7-9]. A possible mechanism, by which high fat diets promote obesity when
59
combined with sucrose, but not proteins, may relate to the effect of sucrose on insulin
60
secretion and enhanced insulin signaling. This finding is underscored by the finding that
61
nifedipine, a dihydropyridine calcium channel blocker that inhibits insulin secretion, was able
62
to protect mice from diet-induced obesity elicited by a high fat high sucrose diet (HF/HS) [2].
63
Dietary proteins and sucrose differ in their ability to stimulate insulin secretion. Animals fed
64
a HF/HS diet have elevated fasting insulin levels compared to animals on a high fat/low
65
sucrose diet. Ins1+/-:Ins2-/- mice which display reduced insulin secretion and are protected
66
against chronic hyperinsulinemia, are also protected against high fat diet-induced obesity [9].
67
Moreover, elevated plasma insulin is correlated with insulin resistance in humans [10]. It is
68
worth noting that even though a high protein:sucrose ratio in the diets prevents high fat diet
69
induced obesity, the mice are still glucose intolerant [1]. However, as both fasting glucose
70
and fasting insulin levels are as low as in low fat fed control mice it is likely that insulin
71
sensitivity is increased. High protein diets may reduce the capacity of pancreatic β-cells to
72
secrete insulin in response to glucose. It was recently demonstrated that glucose-stimulated
73
insulin secretion is highly dependent on communication between βcells in the pancreas, and
74
this phenomenon is mediated by ephrins [11]. Both ephrin A5 and EphA5 can differentially
75
effect insulin secretion. Signaling from EphA5 attenuates insulin secretion, while signaling
76
by ephrin A5 enhances insulin secretion [11]. EphA5 is active when phosphorylated and
77
protein tyrosine phosphatase 1B (PTP1B) is responsible for its dephosphorylation
78
(unpublished data). Therefore, PTP1B is proposed to regulate insulin secretion in pancreatic
79
βcells. Whole body PTP1B deficiency Increases energy expenditure, decreases adiposity, and
80
enhances tissue-specific insulin sensitivity [6]. In the current study we used pancreas specific
81
Ptpn1 knockout mice to further investigate the role of insulin secretion in relation to intake of
82
high fat diets rich in sucrose.
2
83
84
MATERIALS AND METHODS
85
86
Ethics statement
87
All mouse studies were conducted according to federal guidelines and were approved by the
88
Institutional Animal Care and Use Committee at University of California Davis.
89
Animals and diets
90
Mice carrying the floxed alleles for Ptpn1, in which exons 6-8 (encoding the active site of the
91
enzyme) are flanked with loxP sites (Ptpn1 fl/fl), were generated [12] and used to generate
92
Ptpn1 fl/fl mice on a 129Sv/J x C57Bl/6J background. Pdx1-Cre mice on a C57Bl/6J
93
background were obtained from Dr. D. Melton (Harvard University, Boston, MA). All mice
94
were maintained on a 12-hour light-dark cycle in a temperature-controlled facility, with free
95
access to water and food. Genotyping for Ptpn1 floxed allele and the presence of Cre was
96
performed by PCR, using DNA extracted from tail tips. Approximately 8 week old male mice
97
with pancreas specific Ptpn1 knock out (KO) and wild type (WT) controls were fed ad
98
libitum a low fat, high fat high protein (HF/HP) or a high fat high sucrose (HF/HS) diet for
99
ten weeks (n=8-10 per diet). The composition of the diets is shown in table 1. Diets were
100
been obtained from Ssniff Spezialdiaten GmbH, Soest, Germany. In order to control feed-
101
intake and calculate feed-efficiency for each individual mouse, the mice were housed single
102
caged.
103
3
104
Glucose tolerance test (GTT), insulin tolerance test (ITT) and Glucose stimulated insulin
105
secretion test (GSIS).
106
GTT was performed after 4 weeks of feeding. After one week of recovery from the GTT, the
107
ITT was performed. Both assays were repeated starting from week 8. Additionally, at week
108
10 GSIS was determined. For GTT and GSIS mice were fasted overnight before an
109
intraperitoneal injection of 2 g/kg glucose in saline. For ITT mice were fasted 4 hours before
110
an intraperitoneal injection of 0.5 Unit/kg human recombinant insulin in saline. Blood was
111
collected from the tip of the tail at indicated time points. Glucose concentrations were
112
measured with Bayer Contour glucometer. Insulin concentrations were measured using
113
Insulin (Mouse) ELISA kit (DRG Diagnostics).
114
115
Statistics and calculations
116
HOMA IR was calculated using fasting blood glucose and insulin values using the following
117
equation: plasma glucose (mmol/l)×plasma insulin (μU/ml)/ 22.5. Two tail distribution
118
Student t-test analysis for unequal variance for area under the curve in GTT and ITT, and
119
log10 transformed GSIS. Repeated measures ANOVA was used for body weight, and two
120
ways ANOVA for tissue masses. Data were considered statistical significant when P <0.05.
121
122
123
124
125
4
126
RESULTS
127
128
Pancreas specific Ptpn1 deletion attenuates the first phase insulin secretion in mice fed a
129
high fat-high sucrose diet
130
To investigate the effect of Ptpn1 deletion on first phase insulin secretion, a glucose-
131
stimulated insulin secretion test was performed after 10 weeks of feeding. As expected, the
132
first phase insulin secretion (measured 5 min after glucose injection) was higher in mice fed
133
the HF/HS diet compared with low fat fed mice (Figure 1). This increase was blunted in the
134
pancreas specific Ptpn1 knockout mice. Compared with low fat fed mice, insulin levels 5 min
135
after glucose injection were not increased in mice fed a HF/HP diet, and insulin levels were
136
similar in wild type and KO mice fed the HF/HP diet.
137
138
Pancreatic Ptpn1 deletion does not influence body mass in high fat fed mice irrespective
139
of the sucrose content in the diet.
140
As expected, wild type mice fed a HF/HS diet gained more weight compared with mice fed
141
the HF/HP diet (Figure 2A) even though, no significant difference in caloric intake was
142
observed (Figure 2B). The masses of retroperitoneal, gonadal, mesenteric and brown
143
interscapular fat depots as well as in total fat mass were significantly increased in the HF/HS
144
fed mice when compared to HF/HP fed mice (Figure 3) in agreement with our previous
145
findings. The same strong tendency was observe in the case of liver mass of the liver (p<
146
0.056) comparing HF/HS with HF/HP fed mice.
5
147
Total body weight was not influenced by pancreas specific Ptpn1 deletion regardless whether
148
or not sucrose was present in the diet. Similarly, no significant differences between KO and
149
WT mice were observed regarding fat depot masses as well as total fat mass (Figure 3).
150
151
Pancreatic Ptpn1 deletion does not affect blood glucose and insulin parameters.
152
Mice fed the HF/HS diet showed elevated fed plasma glucose level in comparison mice on
153
the HF/HP diet (Figure 4). On the other hand, HF/HS diet did not change fasting glucose as
154
well as fed or fasting insulin parameters compared with the HF/HP diet. None of the blood
155
parameters were changed in response to pancreas specific Ptpn1 deletion. The calculated
156
HOMA IR was influenced neither by sucrose content of high fat diet nor by the mice
157
genotype.
158
159
Pancreas specific Ptpn1 deletion prevents insulin resistance in mice fed the high fat-high
160
sucrose diet.
161
Mice fed HF/HS displayed a tendency towards lover glucose tolerance (p<0.075 in area under
162
the curve analysis of the GTT test) compare to mice fed the HF/HP diet (Figure 5). In the
163
same way, HF/HS fed mice exhibited peripheral insulin resistance (measured by IIT test) in
164
comparison to HF/HP fed mice (Figure 5). The effect of the diet was genotype dependent.
165
The effect of the HF/HS diet on glucose tolerance and peripheral insulin resistance was
166
diminished in mice with pancreas specific Ptpn1 KO. Ptpn1 deletion had no effect on glucose
167
tolerance or insulin resistance in HF/HP fed animals.
168
6
169
DISCUSION
170
171
We have previously reported that a HF/HS diet increased total body mass as well as adipose
172
tissue mass compared to a HF/HP diet [1, 2]. The effect of HF/HS diet on adipose tissue mass
173
was diminished by administration of nifedipine, an inhibitor of insulin secretion inhibitor.
174
The current study confirmed the effect of HF/HS diet on mice body weight (Figure 2) and
175
adipose tissue mass (Figure 3). However, pancreas specific Ptpn1 deletion did not
176
significantly change either body or fat depot masses.
177
Both high sucrose and high fat diets have been reported to induce insulin resistance in animal
178
models [13, 14]. Liver is the primary organ sustaining fasting glucose level through
179
degradation of glycogen and gluconeogenesis. Development of insulin resistance in liver is
180
associated with fasting hyperinsulinemia [15]. In our experiment, all mice fed the high fat
181
diet developed insulin resistance measured by HOMA IR (Figure 3). Replacement of proteins
182
with the sucrose in the high fat diet did not cause an increase in HOMA IR index or fasted
183
glucose or insulin levels (Figure 4). These findings may suggest that all high fat diet fed
184
animals developed hepatic insulin resistance and that the sucrose content did not aggravate
185
this process in our animal model.
186
High sucrose diets have previously been associated with peripheral insulin resistance [16]. In
187
our experiment mice fed with HF/HS diet displayed higher peripheral insulin resistance,
188
measured by insulin tolerance test in comparison with mice fed with the HF/HP diet (Figure
189
5). Interestingly, pancreas specific Ptpn1 deletion reduced insulin resistance caused by
190
HF/HS diet, bringing it to the level of that observed in mice fed the HF/HP diet. On the other
191
hand, pancreas specific Ptpn1 deletion had no effect on insulin resistance in HF/HP diet,
192
suggesting that the observed effect is sucrose dependent.
7
193
Sucrose, in contrast to fat potently induces insulin secretion. Hyperinsulinemia, the state of
194
elevated blood insulin level coexists with insulin resistance, and it has been demonstrated that
195
elevated insulin level induces insulin resistance in rodents [17]. Studies on ex vivo mouse and
196
human tissues have indicated that the combined action of insulin and glucose is required for
197
development of glucose-induced insulin resistance in peripheral tissues [18, 19]. This
198
suggests that glucose-stimulated insulin secretion may be involved in glucose-induced insulin
199
resistance. Here we provided evidence that the first phase insulin secretion may be involved
200
in sucrose induced peripheral insulin resistance. A HF/HS diet significantly increased the first
201
phase insulin secretion as compared to that observed in the low fat control group (Figure 1).
202
Pancreas specific Ptpn1 deletion blunted HF/HS induced elevation in the first phase insulin
203
secretion. Despite of reduction of the first phase insulin secretion, pancreas specific Ptpn1
204
deletion did not significantly changed fed or fasted blood insulin and glucose levels (Figure
205
4) suggesting, that the second phase insulin secretion was maintained. These findings support
206
the hypothesis that the elevated first phase insulin secretion is associated with sucrose-
207
induced peripheral insulin resistance.
208
Increased fat mass and body weight have been shown to correlate with insulin resistance [20]
209
and a reduction in the body weight has been demonstrated to reduce insulin resistance [21]. In
210
our experimental model, mice fed the HF/HS diet increased body weight compared to that of
211
mice fed the HF/HP diet (Figure 2) as shown previously [1]. The HF/HS diet also increased
212
the mass of retroperitenal, gonadal, mesenteric and brown interscapular fat depots (Figure 3).
213
Our data support the notion that PTP1B is involved in regulating first phase insulin secretion,
214
however, probably not in the regulation of general overall insulin secretion caused by
215
prolonged elevated plasma glucose levels as indicated by the lack of differences in fed
216
plasma insulin and glucose levels in WT and KO mice. WT mice fed the HF/HS diet
217
exhibited an increase in peripheral insulin resistance in comparison with mice fed the HF/HP
8
218
diet. However, it is important to note that insulin sensitivity was increased in Ptpn1 KO mice
219
despite that no significant changes in body weight and fat mass in the WT and Ptpn1 KO
220
mice were observed suggesting that development of HF/HS diet-induced insulin resistance
221
somehow is linked to first phase insulin secretion.
222
223
224
Acknowledgments
225
This work was supported by grants from the Danish Natural Research Council, The Novo
226
Nordisk Foundation, the Carlsberg Foundation and the Danish council for Strategic Research,
227
Grant No.: 09-065171.
228
229
230
REFERENCES
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
[1] Ma T, Liaset B, Hao Q, Petersen RK, Fjaere E, Ngo HT, et al. Sucrose Counteracts the AntiInflammatory Effect of Fish Oil in Adipose Tissue and Increases Obesity Development in
Mice. PloS one 2011;6.
[2] Hao Q, Lillefosse HH, Fjaere E, Myrmel LS, Midtbo LK, Jarlsby RH, et al. High-glycemic
index carbohydrates abrogate the antiobesity effect of fish oil in mice. Am J Physiol-Endoc M
2012;302:E1097-E112.
[3] Madsen L, Pedersen LM, Liaset B, Ma T, Petersen RK, van den Berg S, et al. cAMPdependent signaling regulates the adipogenic effect of n-6 polyunsaturated fatty acids.
Journal of Biological Chemistry 2008;283:7196-205.
[4] Madsen L, Kristiansen K. The importance of dietary modulation of cAMP and insulin
signaling in adipose tissue and the development of obesity. Annals of the New York
Academy of Sciences 2010;1190:1-14.
[5] Guller S, Corin RE, Mynarcik DC, London BM, Sonenberg M. Role of Insulin in Growth
Hormone-Stimulated 3t3-Cell Adipogenesis. Endocrinology 1988;122:2084-9.
9
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
[6] Klaman LD, Boss O, Peroni OD, Kim JK, Martino JL, Zabolotny JM, et al. Increased energy
expenditure, decreased adiposity, and tissue-specific insulin sensitivity in protein-tyrosine
phosphatase 1B-deficient mice. Molecular and cellular biology 2000;20:5479-89.
[7] Warram JH, Martin BC, Krolewski AS, Soeldner JS, Kahn CR. Slow glucose removal rate
and hyperinsulinemia precede the development of type II diabetes in the offspring of
diabetic parents. Ann Intern Med 1990;113:909-15.
[8] Kitabchi AE, Temprosa M, Knowler WC, Kahn SE, Fowler SE, Haffner SM, et al. Role of
insulin secretion and sensitivity in the evolution of type 2 diabetes in the diabetes
prevention program: effects of lifestyle intervention and metformin. Diabetes
2005;54:2404-14.
[9] Mehran AE, Templeman NM, Brigidi GS, Lim GE, Chu KY, Hu X, et al. Hyperinsulinemia
drives diet-induced obesity independently of brain insulin production. Cell metabolism
2012;16:723-37.
[10] Mayer-Davis EJ, Levin S, Bergman RN, D'Agostino RB, Karter AJ, Saad MF. Insulin
secretion, obesity, and potential behavioral influences: results from the Insulin Resistance
Atherosclerosis Study (IRAS). Diabetes Metab Res 2001;17:137-45.
[11] Konstantinova I, Nikolova G, Ohara-Imaizumi M, Meda P, Kucera T, Zarbalis K, et al.
EphA-Ephrin-A-mediated beta cell communication regulates insulin secretion from
pancreatic islets. Cell 2007;129:359-70.
[12] Bence KK, Delibegovic M, Xue B, Gorgun CZ, Hotamisligil GS, Neel BG, et al. Neuronal
PTP1B regulates body weight, adiposity and leptin action. Nature medicine 2006;12:917-24.
[13] Samuel VT, Shulman GI. Mechanisms for insulin resistance: common threads and
missing links. Cell 2012;148:852-71.
[14] Cao L, Liu X, Cao H, Lv Q, Tong N. Modified high-sucrose diet-induced abdominally
obese and normal-weight rats developed high plasma free fatty acid and insulin resistance.
Oxidative medicine and cellular longevity 2012;2012:374346.
[15] Shepherd PR, Kahn BB. Mechanisms of disease - Glucose transporters and insulin action
- Implications for insulin resistance and diabetes mellitus. New Engl J Med 1999;341:248-57.
[16] Grimditch GK, Barnard RJ, Hendricks L, Weitzman D. Peripheral Insulin Sensitivity as
Modified by Diet and Exercise Training. American Journal of Clinical Nutrition 1988;48:3843.
[17] Shanik MH, Xu Y, Skrha J, Dankner R, Zick Y, Roth J. Insulin resistance and
hyperinsulinemia: is hyperinsulinemia the cart or the horse? Diabetes care 2008;31 Suppl
2:S262-8.
[18] Richter EA, Hansen BF, Hansen SA. Glucose-Induced Insulin Resistance of SkeletalMuscle Glucose-Transport and Uptake. Biochemical Journal 1988;252:733-7.
[19] Renstrom F, Buren J, Svensson M, Eriksson JW. Insulin resistance induced by high
glucose and high insulin precedes insulin receptor substrate 1 protein depletion in human
adipocytes. Metabolism-Clinical and Experimental 2007;56:190-8.
[20] Schindler TH, Cardenas J, Prior JO, Facta AD, Kreissl MC, Zhang ML, et al. Relationship
between increasing body weight, insulin resistance, inflammation, adipocytokine leptin, and
coronary circulatory function. J Am Coll Cardiol 2006;47:1188-95.
[21] Vazquez LA, Pazos F, Berrazueta JR, Fernandez-Escalante C, Garcia-Unzueta MT,
Freijanes J, et al. Effects of changes in body weight and insulin resistance on inflammation
and endothelial function in morbid obesity after bariatric surgery. The Journal of clinical
endocrinology and metabolism 2005;90:316-22.
10
293
FIGURE LEGENDS
294
Figure 1. Insulin levels during a glucose-stimulated insulin secretion test (GSIS). LF – low fat
295
diet, HF/HP – high fat, high protein diet, HF/HS – high fat high sucrose diet, WT – wild type,
296
KO – pancreas specific Ptpn1 knockout. n=6 per group.
297
Figure 2. A - body weight of experimental animals with the p values from ANOVA repeated
298
measurements analysis. B - the average calories intake of experimental animals. LF – low fat
299
diet, HF/HP – high fat, high protein diet, HF/HS – high fat high sucrose diet, WT – wild type,
300
KO – pancreas specific Ptpn1 knockout. (n=7-10).
301
Figure 3. Tissues mass of experimental animals. LF – low fat diet, HF/HP – high fat, high
302
protein diet, HF/HS – high fat high sucrose diet, WT – wild type, KO – pancreas specific
303
Ptpn1 knockout (n=7-10).
304
Figure 4. Plasma glucose and insulin level in fed and 12 hours fasted stats as well as
305
calculated HOMA IR. LF – low fat diet, HF/HP – high fat, high protein diet, HF/HS – high fat
306
high sucrose diet, WT – wild type, KO – pancreas specific Ptpn1 knockout. n=10 to 7 per
307
group.
308
Figure 5. Glucose tolerance test and insulin tolerance test of experimental animals. Upper
309
panel shows unmodified data and the lover panel shows calculated area under the curve
310
outlined by the measurement values. LF – low fat diet, HF/HP – high fat, high protein diet,
311
HF/HS – high fat high sucrose diet, WT – wild type, KO – pancreas specific Ptpn1 knockout.
312
n=10 to 7 per group.
313
314
11
315
Table 1. Composition of the experimental diets. All values in g/kg of the diet.
Casein
L-cystine
Corn starch
316
Cellulose
Sucrose
Corn oil
Choline bitartrate
Vitamine mix AIN-93-VX
Mineral mix AIN 93
t-butylhydroquinone
Low fat
200
3
529.486
50
90
70
2.5
10
45
0.014
High fat/high
protein
500
3
9.486
50
130
250
2.5
10
45
0.014
High fat/high
carbohydrates
200
3
9.486
50
430
250
2.5
10
45
0.014
317
318
319
320
321
322
323
12
324
325
Figure 1
326
327
328
329
330
331
332
13
333
334
Figure 2
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
14
352
353
354
Figure 3
355
356
357
358
359
360
361
362
363
15
364
365
366
Figure 4
367
368
369
370
371
372
373
374
16
375
376
Figure 5
377
378
379
380
17
1
2
Weight cycling promotes circadian shift and fat gain in
C57BL/6J mice
3
4
5
6
SN Dankel1,2, EM Degerud1,2, K Borkowski3,4, E Fjære3,4, LK Midtbø3,4, C Haugen1,2,
MH Solsvik1,2, AM Lavigne2, K Kristiansen4, B Liaset3, JV Sagen1,2, G Mellgren1,2‡, L
Madsen3,4
7
8
1
Department of Clinical Science, University of Bergen, N-5021 Bergen, Norway
9
2
Hormone Laboratory, Haukeland University Hospital, N-5021 Bergen, Norway
10
11
12
3
National Institute of Nutrition and Seafood Research (NIFES), N-5817 Bergen, Norway
4
Department of Biology, University of Copenhagen, DK-2200 Copenhagen, Denmark
‡
Corresponding author: [email protected]
13
14
Conceived and designed the experiment: LM, BL, KK, SND, GM
15
Carried out experiments/collected samples: EMD, KB, EF, LKM, CH, MHS, AML
16
Analyzed clinical and biochemical data: SND, EMD, KB, EF, LKM, KK, BL, JVS, GM, LM
17
Analyzed gene expression data: SND, EMD
18
Wrote the manuscript: SND, EMD
19
Reviewed manuscript: All co-authors
20
21
22
23
Abbreviated title: Weight cycling and adipose tissue clock genes
24
25
Corresponding author and reprint requests:
26
Professor Gunnar Mellgren, MD, PhD
27
Hormone Laboratory, Haukeland University Hospital, N-5021, Bergen, Norway
28
Phone/Email: +47 55 97 50 00 / [email protected]
29
ABSTRACT
30
To lose or maintain weight, many individuals initiate periods with low energy intake (“yo-yo
31
dieting”), resulting in weight cycling which might ultimately result in further gain of body fat.
32
We tested this by a controlled experiment in C57BL/6J mice, and searched for genes with
33
differential expression in adipose tissue that could be involved in an adaptive mechanism to
34
increase fat storage efficiency. Eleven mice were subjected to four weight cycles, followed by
35
a final three-week overfeeding period. As controls, eleven mice were subjected to chronic
36
overfeeding and low-fat diet, respectively. The weight cycled mice gained significantly more
37
weight than chronically overfed mice (mean±SEM 12.4±0.8 compared to 10.0±0.4 grams,
38
p=0.018). Total energy intake was identical in the weight cycled and chronically overfed
39
groups. Weight cycled mice showed increased adipose tissue mass and leptin levels. While
40
chronic overfeeding primarily increased adipocyte size, weight cycled mice gained weight by
41
a normalization of adipocyte recruitment, and were protected from a worsening of circulating
42
lipid, glucose, and adiponectin levels. Global gene expression was analyzed by microarrays.
43
Weight cycled mice were characterized by a down-regulation of several circadian clock genes
44
(Dbp, Tef, Per1, Per2, Per3, Nr1d2) in adipose tissues, which was confirmed by qPCR. In
45
3T3-L1 cells, we found reduced expression of Dbp and Tef early in adipogenic differentiation,
46
which was mediated by cAMP. Our data suggest that circadian clock transcription factors
47
play a role in the regulation of adipocyte recruitment and fat storage efficiency in response to
48
cyclic feeding patterns.
49
50
51
52
1
53
54
KEYWORDS
55
Energy restriction, Adipose tissue, Adipocyte, Clock genes, Weight cycling
2
56
INTRODUCTION
57
Weight cycling results from intermittent overeating and dieting, representing a major
58
challenge for many individuals (22, 51, 56) which could ultimately promote further expansion
59
of adipose tissue (8, 27, 40, 46, 58). Evolutionary conserved molecular mechanisms allow
60
fine-tuning of metabolic processes to environmental cues. Fat storage is a key evolutionary
61
process conducive to survival in organisms ranging from the worm C. elegans to humans (34).
62
The adipocyte, being the primary fat storing cell, responds to changes in energy availability in
63
part by releasing the peptide hormones leptin and adiponectin. These potent hormones exert
64
local, central and peripheral effects, coordinating the systemic response to fasting/refeeding
65
cycles by modulating energy storage/expenditure, appetite, biological rhythms, and other
66
functions. In adipose tissue, signals such as leptin may modulate the capacity for lipogenesis
67
in response to weight fluctuations, predisposing to compensatory fat regain during consequent
68
energy surplus (23, 24, 52). Identification of novel genes associated with weight cycling and
69
weight regain may provide new molecular insight into the evolutionary adaptation to variable
70
energy availability.
71
Adipocytes are derived from multipotent mesenchymal stem cells, in a process involving
72
commitment to the adipocyte lineage followed by terminal differentiation of the committed
73
preadipocytes. Research using cell lines such as 3T3-L1 or mouse embryo fibroblasts (MEFs)
74
has provided information on terminal adipocyte differentiation. Treatment of these cells with
75
fetal bovine serum, glucocorticoids and high levels of insulin (or physiological concentrations
76
of insulin like growth factor-1 (IGF-1)) initiates differentiation. The fasting signal cyclic
77
adenosine monophosphate (cAMP) and cAMP-dependent processes are pivotal during the
78
early stages of adipocyte differentiation. Factors that increase cellular cAMP, such as
79
isobutylmethylxanthine (IBMX) or forskolin, strongly accelerate the initiation of the
3
80
differentiation program in vitro and lead to commitment the cells for terminal differentiation
81
and lipid-filling (7, 33).
82
The synergistic actions of insulin/IGF-1 and cAMP signaling during initiation of adipocyte
83
differentiation is remarkable and contrasts the normal interplay between insulin and cAMP
84
signaling in liver, muscle, and mature adipocytes. Aiming to recreate this concurrence of
85
fasting/refeeding adipogenic signals in vivo, we switched C57BL/6J mice to ten days periods
86
of overfeeding immediately after four days of energy restriction. We hypothesized that
87
elevated cAMP signaling during periods of energy restriction contributes to development of
88
new preadipocytes that expand into lipid-filled mature adipocytes when calories again are in
89
excess. Moreover, we wanted to identify novel genes in adipose tissue that could be involved
90
in mediating such an effect. To our knowledge, there are no reports on the transcriptomic
91
responses to weight cycling in adipose tissues. In rats, it has previously been shown that
92
weight cycling increases feed efficiency (the weight gained relative to the amount of energy
93
ingested) and resistance to weight loss after repeated weight cycling (5, 18, 45). There are
94
conflicting reports (13, 29), but these diverging results might be explained by different
95
protocols for weight cycling. We designed a weight cycling experiment for C57BL/6J mice
96
based on the protocol used for in vitro differentiation of mouse 3T3-L1 preadipocytes and
97
MEFs. Using this protocol we found greater increase in total body mass and visceral fat mass
98
after weight cycling compared to chronic overfeeding, which was associated with a
99
suppressed expression of circadian clock genes that may be novel regulators of metabolism in
100
adipose tissue.
101
102
MATERIALS AND METHODS
103
Weight cycling experiment
4
104
The study was approved by the Norwegian State Board of Biological Experiments with
105
Living Animals. Thirty-six male C5BL/6J BomTac mice were obtained from Taconic, Europe
106
(Ejby Denmark) . At 28 ± 2
107
stainless-steel wire cages with wooden chips and paper peel that were changed every two
108
weeks. At arrival, the mice were six to eight weeks old, weighed 26 ± 1.3 grams and were fed
109
low-fat diet for one week. All mice had free access to tap water during the entire experiment.
110
The mice were fed Monday, Wednesday and Friday morning and feed leftover was removed,
111
weighed and registered. Body mass was measured at start and at sacrifice, as well as every
112
Monday and Friday throughout the experiment.
113
After one week of acclimatization, the 36 mice were randomly divided into three groups
114
(n=11): low-fat diet fed, chronically overfed (CO), and weight cycled (WC). The low-fat diet
115
(D12450B) contained 20e% protein, 35e% starch, 35e% sucrose, and 10e% fat (Research
116
Diets, USA). The WC mice were overfed for ten days by unlimited access to an energy dense
117
high-fat high-sucrose diet (S8672-E056S), containing 15e% protein, 1e% starch, 37e%
118
sucrose and 47e% fat (ssniff Spezialdiäten GmbH, Germany). This energy-dense diet
119
provided 900 more kilocalories per kilogram of food compared to the low-fat diet, equivalent
120
to approximately 25% more energy. After each of four overfeeding periods, WC mice were
121
energy restricted for four days by limiting their access to the energy dense diet to 70% of the
122
energy ingested during the previous overfeeding period. After the fourth energy restriction
123
phase, the WC mice received the energy-dense diet for three weeks ad libitum before
124
sacrifice.
125
One week prior to sacrifice, after an overnight fast, a glucose tolerance test (GTT) was
126
performed as previously described (38). Briefly, mice were injected with glucose dissolved in
127
saline in the intraperitoneal cavity (2 gram glucose per kg body weight). Blood was drawn
128
from the tail at 0, 15, 30, 60 and 120 minutes after injection.
C with 50%
5
129
The mice were sacrificed in random order between 9 and 12 AM on two separate days. The
130
mice were anesthetized with Isofluran (Isoba-vet. Schering-Plog, Denmark) using the
131
Univentor 400 anesthesia Unit (Univentor Limited, Sweden), and euthanized by cardiac
132
puncture in the fed state. Tissues were quickly dissected out, weighed, placed in marked
133
plastic bags, freeze clamped and stored at -80ºC.
134
135
Overfeeding experiments
136
Thirteen male C57BL/6J BomTac and 14 male 129S6/SvEvTac were obtained from Taconic,
137
Europe (Ejby Denmark) at 8 weeks of age and acclimated as described above. After
138
acclimatization, the mice were housed individually and randomly assigned to the
139
experimental diets (n=6-7). The low-fat diet (S8672-E050) contained 20e% protein, 35e%
140
starch, 35e% sucrose, and 10e% fat, and the western diet (S8672-E400) contained 18e%
141
protein, 13e% starch, 31e% sucrose, and 38e% fat (ssniff Spezialdiäten GmbH, Germany).
142
The animals were fed the experimental diets described for 10 weeks ad libitum. Throughout
143
the experiment, all mice were weighed once a week and feed intake was assessed every
144
Monday, Wednesday and Friday. GTT, euthanization and tissue collection were performed as
145
described above.
146
147
Blood measurements
148
Blood lipids were measured in 80µl heparin-plasma by a MAXMAT PL instrument
149
(MAXMAT S.A., France). Plasma insulin was measured by the DRG® Mouse Insulin
150
ultrasensitive ELISA kit (EIA-3440, DRG International, Inc., USA), according to the
151
manufacturer’s protocol. Leptin and adiponectin were measured by commercial available
6
152
ELISA kits (Mouse and Rat Leptin, RD291001200R, BioVendor – Laboratorní medicína a.s.,
153
Mouse Adiponectin, EZMADP-60K, Millipore).
154
155
Adipose tissue morphology
156
Fat tissue was fixed overnight at 4oC by immersion in phosphate buffered 4%
157
paraformaldehyde. Samples were then dehydrated and embedded in paraffin. 5µm thick
158
sections were taken from two different locations of embedded tissue and were hematoxylin
159
and eosin (H&E) stained. Three to six microscopic images from five mice per group were
160
taken from random places of each section. The area of each cell on each image was measured
161
by drawing the circumference using the ImageJ open source software. Additionally, using the
162
same software the number of cells per surface was estimated by extrapolating the average
163
adipocyte volume (µm3). Total adipocyte number in each fat pad was estimated by dividing
164
total fat pad mass by average adipocyte volume.
165
166
In vitro adipocyte differentiation
167
3T3-L1 mouse cells were cultured in Dulbecco´s modified Eagle´s medium (DMEM) with
168
high glucose (4.5g/l) containing 10% calf serum (CS) and 1% penicillin and streptomycin
169
(PEST), and were allowed to grow to 100% confluence in 6-well plates. Two days after total
170
confluence (“day 0”), differentiation was induced by a two-day treatment with the synthetic
171
glucocorticoid dexamethasone (0.5mM), insulin (175nM), phosphodiesterase inhibitor
172
methylisobutylxanthine (IBMX) (0.5M), and 10% fetal bovine serum (FBS). On day 2, the
173
medium was changed to DMEM with 10% FBS and 175nM insulin. On day 4 the medium
7
174
was changed to DMEM supplemented only with 10% FBS, and the cells were then allowed to
175
develop until day 10.
176
To collect cell lysates for qPCR, each well in one 6-well plate was washed with 2ml room
177
tempered PBS before adding 350µL buffer RLT (Qiagen) or 500µl of TRI Reagent® (Sigma
178
Aldrich), followed by scraping and transfer to 1.5ml eppendorf tubes. The lysates were
179
vortexed for ten seconds, spun down and stored at -80ºC.
180
181
RNA purification
182
To homogenize the mouse inguinal and epididymal white adipose tissues, approximately
183
100mg of frozen tissue was transferred to 2ml micro-centrifuge tubes with round bottom. One
184
ml Qiazol (Qiagen) and a stainless steel bead 5mm in diameter were added, immediately
185
before shaking in a TissueLyser II (Qiagen) three times at 25Hz for two minutes. RNA was
186
extracted using the RNeasy Lipid Tissue Mini Kit together with the On-Column DNase
187
digestion with the RNase-Free DNase set (Qiagen), according to the protocol of the
188
manufacturer. RNA from the primary human adipose culture was purified in a QIAcube using
189
the RNeasy Lipid Tissue Mini Kit (Qiagen). The quantity and quality of RNA was analyzed
190
by the NanoDrop®ND-1000 spectrophotometer and the Agilent 2100 BioAnalyzer (Agilent
191
RNA 6000 Nano Kit).
192
193
Microarray analysis
194
Global gene expression was measured in iWAT, eWAT and iBAT of WC and CO mice by
195
Illumina microarray analysis. The microarray data are MIAME compliant and are available in
196
ArrayExpress (awaiting accession number). Each group had eleven iWAT and iBAT samples
8
197
and ten eWAT samples. The microarrays (MouseRef-8 v2.0 Expression BeadChip) contained
198
25,697 unique probes representing approximately 19,100 unique genes. 400ng of RNA from
199
each sample was reversely transcribed, amplified and labeled with biotin-16-UTP using the
200
Illumina®TotalPrep RNA Amplification Kit Ambion, using an Eppendorf Mastercycler
201
(Applied Biosystems/Ambion, USA). 750ng cRNA of each sample was hybridized to the
202
MouseRef-8 v2.0 Expression BeadChip according to manufacturer’s protocol. Signal
203
detection was performed using the Illumina iScan Reader. All RNA integrity numbers (RIN)
204
were above 7.5. Raw data were imported and analyzed in J-Express (6). Rank product
205
analysis (4) was used to compare global gene expression in the WC and CO groups.
206
207
cDNA synthesis and qPCR
208
cDNA was synthesized from 300ng of total RNA using the SuperScript® VILO™ cDNA
209
Synthesis Kit. Reaction and enzyme mixtures were added to the samples along with PCR-
210
grade water to adjust to a total volume of 20µl. The samples were then run in the thermal
211
block cycler GeneAmp® PCR System 9700 at 25°C for 10 minutes, 42°C for 60 minutes, and
212
finally at 85°C for five minutes. The samples were diluted 10-fold with PCR-grade water and
213
stored at -20°C.
214
qPCR was performed using the LightCycler® 480 Probes Master kit and a LightCycler® 480
215
(Roche). Primers and probes were designed using the Universal Probe Library (UPL) online
216
design center (www.roche-applied-science.com) (Table 1). Primers were obtained from
217
Sigma-Aldrich (www.sigmaaldrich.com/configurator/servlet/DesignCenter). Amplification
218
efficiency of target genes was assessed by running 1:5 or 1:10 dilution standard curves based
219
on concentrated cDNA of mouse or human origin. Mouse and human target gene
9
220
concentrations were calculated relative to the mRNA concentration of reference genes Rplp0
221
and IPO8, respectively.
222
223
Statistical analyses
224
Rank product analysis and Significance of Microarray Analysis (SAM) were performed using
225
J-Express 2009 (6). All other statistical analyses in this study were performed with PASW
226
Statistics 18 for Windows. Statistical hypothesis testing was performed using parametric 2-
227
tailed t-test (for normally distributed data, presented as mean) or non-parametric Mann-
228
Whitney U test (for non-normally distributed data, presented as median). Normality was
229
determined by the Shapiro-Wilk test (p>0.05).
230
231
RESULTS
232
Increased body weight after weight cycling
233
The weight cycled (WC) mice gained significantly more weight by the end of the study than
234
chronically overfed (CO) mice (measured as the difference in body mass between the last and
235
first day) (mean 12.4±0.8 compared to 10.0±0.4 grams, p=0.018) (Fig. 1A). Since total energy
236
intake was identical in the WC and CO mice at the end of the experiment (mean 1180±14 and
237
1172±23 kcals, p=0.746), the WC mice had a significantly higher overall feed efficiency
238
(total weight gain relative to total energy intake) compared to CO mice (mean 10.5±0.6 and
239
8.5±0.3 after 81 days, p=0.013) (Fig. 1B). Comparing CO to low-fat fed mice, the CO mice
240
did not gain significantly more total body mass (mean 10.0±0.4 and 9.4±0.5 grams, p=0.336).
241
However, energy intake was significantly higher in CO mice than in low-fat fed mice (mean
242
1172±23 vs. 1008±19 kcal, p=2.47E-5).
10
243
To determine where energy surplus was stored in response to the different diets, we
244
weighed adipose tissues, muscle and liver. The iWAT mass was on average 1.39-fold higher
245
after WC compared to CO, but this was not significant (mean 0.43±0.05 and 0.31±0.04
246
grams, p=0.083) (Fig. 1C). The eWAT of the WC group weighed 1.64-fold more than the CO
247
group (mean 1.20±0.14 compared to 0.73±0.11 grams, p=0.018) (Fig. 1C). The iBAT mass
248
was significantly higher after WC than CO (mean 0.094±0.008 and 0.075±0.003 grams,
249
p=0.032), but was not significantly different between WC and low-fat fed mice (mean
250
0.094±0.008 and 0.083±0.005, p=0.177). There was no significant difference between the WC
251
and CO mice in the weight of liver (mean 1.49±0.07 and 1.39±0.06 grams, p=0.280) or
252
muscle (mean 0.130±0.006 and 0.124±0.004 grams, p=0.466) (Fig. 1C).
253
The adaptation to fasting/refeeding could involve a change in signals that influence
254
preadipocyte recruitment and/or terminal differentiation of adipocytes. Therefore, we assessed
255
differences in number and size of adipocytes between the groups, based on slides from five
256
mice per group. In iWAT, we found a strong trend towards a reduced number of adipocytes in
257
CO compared to the low-fat group (p=0.058) (Fig. 2B). On the other hand, adipocyte number
258
in WC mice tended to be higher than in CO mice (2.3-fold, p=0.079) (Fig. 2B). Similarly, the
259
relative adipocyte number in eWAT tended towards a reduction in the CO group compared to
260
the low-fat group (p=0.160), but was on average ~2-fold albeit non-significantly higher in
261
WC compared to CO (p=0.240) (Fig. 2E). These consistent tendencies suggest that WC led to
262
a normalization of preadipocyte recruitment relative to chronic overfeeding, which could
263
account for increased fat mass after WC. Regarding size, adipocytes were on average
264
similarly enlarged in iWAT in both WC (mean 2135 µm2) and CO (mean 2347 µm2) mice
265
compared to low-fat fed mice (mean 1224 µm2) (p=0.031 and p=0.028, respectively) (Fig.
266
2C). In eWAT, average adipocyte size was not significantly higher after WC compared to CO
267
(p=0.253), though the observed ~1.6-fold higher adipocyte size in WC mice was non11
268
significant compared to the low-fat group (p=0.0505) (Fig 2F), although it should be noted
269
that statistical power was reduced in these analyses. However, the overall consistency in the
270
data support that weight cycled mice gained fat mass by efficient recruitment of new
271
preadipocytes, which was not evident after chronic overfeeding.
272
273
Circulating hormones, glucose and lipids
274
There were no significant differences in fed state plasma insulin or glucose between WC, CO
275
and low-fat fed mice (Fig. 3). Interestingly, despite the higher fat mass, WC mice were not
276
less glucose tolerant than CO mice relative to low-fat fed mice (mean area under the curve for
277
blood glucose 1387, 2249 and 2479 mmol/liter*120 min for low-fat fed, CO, and WC,
278
respectively (CO and low-fat fed p=1.67E-4; WC and CO p=0.345) (Fig. 3).
279
The increased fat mass in the WC mice may be reflected in altered circulating levels of lipids
280
and adipokines. As expected, the concentration of free fatty acids (FFA) was significantly
281
increased in CO compared to low-fat fed mice (mean 1.55-fold, p=0.016) (Fig. 3).
282
Interestingly, however, there was no further increase in FFA after WC, rather the FFA tended
283
to be decreased in WC compared to CO mice (mean 1.21-fold, p=0.218). Similarly, plasma
284
triacylglycerol (TG) levels were reduced or similar rather than increased in WC compared to
285
low-fat fed mice (mean 1.53-fold lower, p=0.008) and CO mice (mean 1.3-fold, p=0.260).
286
There were no significant differences in plasma glycerol between the groups (Fig.3). Total
287
plasma cholesterol levels were increased similarly in both CO and WC compared to low-fat
288
fed mice (mean 1.27 and 1.30-fold, p<0.0001) (Fig. 3).
289
As expected from the high-energy feeding and reduced glucose tolerance (37), plasma
290
adiponectin levels were significantly decreased after CO compared to the low-fat group (mean
291
1.46-fold, p=0.006) (Fig. 3). However, despite the increased fat mass after WC, WC mice
12
292
tended to have higher adiponectin levels rather than a further suppression relative to CO
293
(1.28-fold, p=0.073). Plasma leptin levels were significantly increased in the WC mice
294
relative to both low-fat diet (4.22-fold, p=0.006) and CO (2.53-fold, p=0.031) (Fig. 3). There
295
was a strong positive correlation between plasma leptin and feed efficiency within the WC
296
group (p=0.006), but not within the CO group (p=0.424) (Table 2). Taken together, the data
297
indicate that WC promotes increased fat storage efficiency although not worsening metabolic
298
health compared to CO.
299
300
Depot-specific transcriptomic responses
301
To obtain a global view of the effect of weight cycling on gene expression in adipose tissues,
302
we performed a microarray-based analysis on gene expression in iWAT, eWAT and iBAT
303
using Illumina microarrays. A rank product analysis (4) was performed to identify
304
differentially expressed genes between WC and CO in each fat pad (Table S1-S3). Because
305
eWAT contributed more to the increased fat mass after weight cycling, we searched for genes
306
responding more strongly to weight cycling in eWAT than in iWAT, as these genes may be
307
related to the eWAT mass development. Comparing WC and CO mice, five transcripts
308
(Fgf13, Ubd, Tph2, Sfrp5, Lipf) were more up-regulated and two transcripts (Mycl1, Inmt)
309
were more down-regulated after WC in eWAT than in iWAT (≥ 1.2-fold greater WC/CO ratio
310
in eWAT compared to iWAT) (Table 3). The up-regulated genes were positively correlated
311
with feed efficiency within the WC group but not within the CO group, and the correlation
312
coefficients were stronger for eWAT than iWAT (Table 4). Corresponding negative
313
correlations were observed for the down-regulated genes, with stronger correlation
314
coefficients for eWAT (Table 4). In iWAT, three genes were more up-regulated (Apoa1,
13
315
Apoa2, Ambp) while none were more down-regulated compared to eWAT after WC (data not
316
shown).
317
We moreover hypothesized that gene networks altered in eWAT might describe
318
molecular processes underlying the increased feed efficiency after WC. Gene ontology
319
analysis was performed on differentially expressed genes in eWAT that were defined to be
320
unaffected in iWAT (63 transcripts with higher expression (Table S4) and 46 transcripts with
321
lower expression (Table S5) after WC compared to CO, fold change > 1.2, eWAT q-value <
322
0.05 and iWAT q-value > 0.95). Based on PANTHER gene ontology analysis, there was an
323
over-representation of up-regulated eWAT related genes in the functional categories
324
development, angiogenesis, lipid transporter activity, and inflammation (Table 5). Of
325
particular note, genes encoding markers of pro-inflammatory macrophages were up-regulated
326
after WC in eWAT but not in iWAT, including Ccl2/Mcp-1, Cd68, and Lgals3/Mac-2 (Table
327
S4). For down-regulated eWAT related genes after WC, there was an over-representation of
328
genes involved in defense response to bacterium (Defb11, Defb20, Defb37, Defb38) and
329
oxygen and reactive oxygen species metabolism (Table S5). In comparison, in iWAT there
330
were only two up-regulated genes (Npm3, Eepd1) and 14 down-regulated genes that were
331
defined to be unaffected in eWAT (Table S6 and S7). Most of the down-regulated iWAT
332
related genes not affected in eWAT were related to muscle contraction, muscle organ
333
development, and mesoderm development. Of note, these genes were similarly down-
334
regulated in iBAT. In total, there were 81 up-regulated and 49 down-regulated genes in iBAT
335
after WC compared to CO (q-value < 0.05, fold difference > 1.2) (Table S8 and S9). Among
336
these, 34 up-regulated and six down-regulated genes were defined to be unaffected in eWAT
337
and iWAT (q-value > 0.95) (Table S10 and S11).
338
14
339
Altered clock gene expression in adipose tissues and liver
340
We next investigated whether WC induced systemic changes that could be detected via
341
common transcriptomic responses in adipose tissues. By rank product meta-analysis (4) of all
342
three fat pads (q-value < 0.05, and fold difference > 1.2 for each depot), we found six
343
transcripts with higher and ten with lower expression after WC compared to CO (Table 3,
344
Table S12). Interestingly, four of the ten common genes exhibiting reduced expression after
345
WC encoded circadian clock transcription factors (Dbp, Per2, Tef, Nr1d2/Rev-Erb beta)
346
(Table 3). These factors act as negative elements in the circadian feedback loop, whereas the
347
Arntl/Bmal1 and Clock represent positive elements that activate the negative elements (12,
348
39). By qPCR, we verified a significant down-regulation of Dbp, Tef (variant 2) and Nr1d2 in
349
eWAT after WC compared to CO (Fig. 4A). Although qPCR did not verify reduced Per2
350
expression in eWAT, the expression levels of Per1 and Per3 were significantly decreased. The
351
decreased positive loop clock gene expression in eWAT was generally specific to WC, and
352
not an effect of high-energy feeding per se, since expression in CO was similar (Dbp, Nr1d2)
353
or higher (Per1, Per3) compared to low-fat fed mice, rather than reduced (Fig. 4A). Moreover,
354
as expected, we observed significantly increased expression of the positive loop clock genes
355
Bmal1 and Clock in eWAT of both CO and WC mice compared to the low-fat group (Fig.
356
4A). In iWAT, there was overall less difference in clock gene expression between WC and
357
CO, where Dbp, Tef, Nr1d2, and Per1 were step-wise but non-significantly lower in WC and
358
CO compared to the low-fat group.
359
To verify that the reduced expression of negative loop clock genes in eWAT was a specific
360
characteristic of weight cycling, and not a normal response to high fat feeding or eWAT mass,
361
we analyzed independent high-fat feeding experiments in C57BL/6J (obesity-prone) and
362
SV129 (obesity-resistant) mice. We found no reduction in clock gene expression in these
363
independent experiments (Fig. 4A).
15
364
Altered clock gene expression has recently been implicated in metabolic regulation in the
365
liver (10, 15). Since we found altered clock gene expression in all fat pads, we examined if
366
these genes were also differentially expressed in liver as part of a systemic effect. There was a
367
tendency towards a similar decrease in the expression of clock genes also in liver, though the
368
data for each gene were not statistically significant (Fig. 4B). Thus, we cannot rule out that
369
weight cycling affected clock gene expression also in other peripheral metabolic tissues.
370
371
Clock genes in adipogenesis
372
Our primary hypothesis was that weight cycling would promote adipogenesis and
373
subsequently fat storage, and we designed our weight cycling protocol to mimic the
374
synergistic action of insulin/IGF-1 and cAMP signaling during initiation of adipocyte
375
differentiation. However, based on the in vivo data we could not determine if clock genes are
376
regulated by these adipogenic signals. We therefore differentiated 3T3-L1 cells in vitro and
377
compared the mRNA expression of Dbp and Tef at different time-points relative to
378
undifferentiated cells. Dbp and Tef are members of the PAR (proline and acidic amino acid-
379
rich) subfamily of basic leucine zipper (bZip) transcription factors together with the third
380
member Hlf (hepatic leukemia factor), which operate as heterodimers (9). These factors have
381
previously been implicated in lipid metabolism in the liver (10), but a function for these genes
382
in preadipocyte recruitment and/or terminal differentiation has not been described. We found
383
reduced expression of Dbp and Tef (variant 2) mRNA in cells treated with differentiation
384
stimuli compared to untreated cells, beginning at about 16 hours after induction (Fig. 5A, 5B)
385
(Tef variant 1 was expressed at a diminished level). These data suggest that a reduction in
386
Dbp and Tef expression could potentially facilitate early adipocyte development. On the other
387
hand, in the later stage of adipogenesis (from day 5-6), expression was higher in the
16
388
differentiating compared to the untreated cells. Thus, Dbp and Tef may be dynamically
389
involved in the adipogenic program.
390
Our morphological data suggested increased preadipocyte recruitment after weight cycling, in
391
which any of the early adipogenic signals could play a role. We therefore assessed which of
392
the individual adipogenic stimuli may have suppressed adipose tissue Dbp and Tef mRNA in
393
the weight cycled mice. After 24 hours treatment of 3T3-L1 cells, we found that IBMX
394
(promoting increased cAMP levels) was the most potent suppressor of Dbp and Tef
395
expression (Fig. 5C, 5D). Dexamethasone increased Tef expression (Fig. 5D); however, this
396
was completely abolished by simultaneous IBMX treatment (data not shown).
397
398
399
400
DISCUSSION
401
In the present study, we show that young C57BL/6J mice experience increased feed efficiency
402
(weight gained relative to energy consumed) after repeated cycles of high and reduced energy
403
intake. This obesogenic effect of weight cycling was related to a normalization of adipocyte
404
recruitment compared to chronic overfeeding. We moreover performed a comprehensive
405
analysis of gene expression in adipose tissues in response to weight cycling. The key finding
406
is that weight cycling associated with a consistent alteration in clock gene expression, which
407
was not observed with isocaloric chronic overfeeding. The clock gene expression was
408
measured three weeks after the last energy restriction phase, indicating that clock gene
409
expression in adipose tissue was persistently affected.
17
410
There is emerging evidence for a key role of clock genes in energy homeostasis (49).
411
Night-shift work has been found to increase the risk of obesity and metabolic syndrome (31).
412
Importantly, although central and peripheral circadian rhythms are synchronized, peripheral
413
mammalian clocks do not depend on central control (39, 53), which is also the case for
414
adipocytes (12). Feeding pattern is a primary factor influencing the autonomous peripheral
415
clock (57), referred to as the food-entrainable oscillator (FEO) (50). Disruption of the gene
416
encoding the clock transcription factor Arntl/Bmal1 in hepatocytes (28) and beta cells (35) of
417
mice alters glucose metabolism. Members of the period (Per) proteins have also been
418
implicated in lipid metabolism, at least in part via direct regulation of PPARγ (14, 16). More
419
recently, clock genes were for the first time causally implicated adipocyte development, as it
420
was demonstrated that adipocyte-specific knock-out of Bmal1 affects feeding pattern and
421
promotes obesity (44). This effect was explained at in least part by reduced energy
422
expenditure without a change in total energy intake (the mice ate less than normal during the
423
dark phase and more during the light phase). A previous study also showed that knock-down
424
of Bmal1 diminished lipid accumulation, whereas Bmal1 overexpression induced expression
425
of lipogenic genes in 3T3-L1 adipocytes (47). Another very recent study further demonstrated
426
that Bmal1 and Clock deficient mice have increased fat storage and are sensitive to fasting
427
(48). These data support a functional involvement of clock genes in the obesogenic response
428
to weight cycling.
429
We provide new evidence that the circadian regulators Dbp (albumin D box-binding protein)
430
and Tef (thyrotroph embryonic factor) may play a role in adipose tissue development and
431
function. Dbp and Tef are members of the PAR (proline and acidic amino acid-rich)
432
subfamily of basic leucine zipper (bZip) transcription factors together with the third member
433
Hlf (hepatic leukemia factor) (9). Previously, it has been shown that complete knock-out of all
434
three PAR members display differential expression of genes involved in lipid metabolism in
18
435
the liver (10), and that Tef knock-out mice lose more weight during energy restriction than
436
wild-type animals (20). We observed reduced expression of Dbp and Tef mRNA after weight
437
cycling compared to chronic overfeeding, which fits with the tendency of up-regulated
438
Arntl/Bmal1 as these genes act as negative and positive regulators of the circadian clock,
439
respectively (39). Furthermore, our in vitro data show that Dbp and Tef mRNA is negatively
440
regulated by cAMP, suggesting that a suppression of these genes may promote initiation of
441
adipocyte differentiation, with subsequent development of the newly recruited preadipocytes
442
during overfeeding.
443
It is interesting that we observed altered clock gene expression three weeks after the
444
last energy restriction cycle. Of note, an effect on clock genes was not observed in our
445
independent standard diet-induced obesity experiment in C57BL/6J mice. A previous study
446
showed that high-fat feeding altered the diurnal expression of clock genes in adipose tissue of
447
mice (25). However, although several of the clock genes were up-regulated after high-energy
448
diet in SV129 mice (who are resistant to diet-induced obesity) and after chronic overfeeding
449
relative to low-fat diet, the weight cycled mice were characterized by reduced expression of
450
negative loop clock genes. This particular pattern of clock gene expression could be a lasting
451
adaptive response to the cyclic feeding. Conceivably, the weight cycling may have evoked
452
epigenetic changes, which has been indicated to ensure specificity and plasticity in the
453
regulation of circadian rhythm in response to metabolic cues (1, 3, 36). Differential histone
454
H3K9 acetylation was recently demonstrated at the promoters of clock genes such as Dbp,
455
Per2 and Bmal1 in mouse adipose tissue (17). The altered clock gene expression, in turn, may
456
have promoted a lasting propensity for fat storage in adipocytes rather than being a response
457
to the increased fat mass, as recently evidenced by Bmal1/Arnt knock-out mice (44).
458
Leptin is an adipocyte-derived peptide hormone which plays a fundamental role in the plastic
459
regulation of energy balance (59). We found a close correlation between plasma leptin levels
19
460
and feed efficiency within the weight cycling group, supporting increases in adipocyte size.
461
By our study design we could not determine whether temporal leptin suppression during the
462
energy restriction phases suppressed energy expenditure, thereby explaining the increased
463
feed efficiency in the subsequent periods. However, it is interesting to note that leptin
464
deficient ob/ob mice show blunted expression profiles of clock genes in liver and adipose
465
tissue but not in brain, which can be reversed by leptin treatment (2). Thus, it was proposed
466
that suppressed clock gene expression is a result of leptin deficiency rather than being a
467
secondary effect of metabolic abnormalities (2). Together with our study, this link with leptin
468
supports that peripheral clock genes participate in the adaptive strategy to limit future energy
469
deficiency after weight cycling. Interestingly, a weight cycling study in rats showed
470
suppressed leptin levels even 12 weeks after the last weight cycle, and this was associated
471
with increased adipose tissue expression of lipogenic enzymes (23). Additionally, leptin
472
administration in ob/ob mice has been shown to suppress the protein expression of lipogenic
473
enzymes such as fatty acid synthase (Fas) and ATP citrate lyase (Acl) (19). These data
474
suggest that changes in both energy expenditure and fat storage may have contributed to the
475
increased feed efficiency after weight cycling, in part via leptin-mediated effects on clock
476
genes. Of note, at the end of our study we found increased rather than decreased leptin levels,
477
and no significant correlation with the mRNA expression of clock genes in adipose tissue
478
(data not shown). These observations indicate leptin resistance, which may have caused
479
insufficient suppression of lipogenic enzymes after weight cycling, and further support that
480
the clock gene expression was unrelated to adipocyte size per se.
481
Despite greater fat mass gain after weight cycling compared to chronic overfeeding, the
482
weight cycled mice had lower levels of plasma triglycerides and fatty acids, and higher
483
adiponectin levels. This phenotype could be due to increased fat storage efficiency in
484
adipocytes. Of particular note, overexpression of adiponectin in leptin deficient mice leads to
20
485
extreme obesity while positively affecting lipid levels and glucose homeostasis (21).
486
Moreover, mice with time-restricted feeding (access to a high-fat diet only during 8 hours of
487
the dark phase), compared to isocaloric ad libitum feeding, showed positive effects on energy
488
expenditure and protection against obesity, hyperinsulinemia, and hepatic steatosis (15).
489
Intriguingly, these mice also showed altered expression of clock genes in the liver. Thus,
490
although weight cycling increased fat storage, particularly in eWAT which also showed
491
increased expression of pro-inflammatory markers, the fat gain after weight cycling
492
associated with relatively salutary effects on the whole-body metabolic profile. This is
493
supported by previous studies showing beneficial effects of time-restricted feeding on
494
metabolic disease (15) and of weight cycling on longevity compared to a steady state of
495
obesity (32).
496
In addition to clock genes, we identified several genes that may promote feed efficiency and
497
fat gain due to weight cycling. Of particular note, a study of genetically identical C57BL/6J
498
mice indicated Sfrp5 as a causal factor in fat mass expansion, since a higher expression in
499
seven-week-old mice before an obesogenic diet strongly correlated with the level of adiposity
500
after eight weeks of the obesogenic diet (26). Moreover, Sfrp5 knock-out mice are protected
501
from diet-induced obesity, an effect which was related to effects on adipocyte size rather than
502
number. Functional analyses further showed that a lack of Sfrp5 increased mitochondrial
503
activity (41). Thus, the increased expression of Sfrp5 we observed after weight cycling
504
particularly in eWAT may have directly contributed to increase adipocyte size and adiposity,
505
at least in part by suppressing energy expenditure. Additionally, Sfrp5 exerts anti-
506
inflammatory effects on inflammatory cells in adipose tissue thereby protecting against
507
glucose intolerance and hepatic steatosis (43), which is in line with the relatively healthy
508
phenotype of our obese weight cycled mice. The metabolic effects of other genes with a
21
509
similar expression pattern as Sfrp5, e.g. Tph2, Fgf13, Ubd, and Lipf, should be further
510
investigated.
511
The human relevance of our findings is of clinical interest. Accumulating evidence indicates
512
that adipose tissue expandability during energy surplus may determine the pathological
513
outcome, by preventing elevated circulating lipids and protecting from ectopic lipid storage in
514
vital organs (55). Experience from clinical practice has shown that avoiding weight regain
515
after fat loss is a major challenge; rather, individuals with a history of weight cycling may
516
regain weight more easily (8). It is therefore debated whether the focus should be on weight
517
maintenance rather than loss, though the long-term effect of weight cycling on disease risk is
518
unclear (8, 30, 54). Our findings highlight the intriguing regulation of feed efficiency, adding
519
to previous evidence that feeding pattern may ameliorate metabolic syndrome without
520
reducing energy intake (15). It remains to be determined to what extent clock genes
521
coordinate metabolic responses in human tissues in the context of weight cycling. While one
522
study of human adipose tissue found no effect of obesity and type 2 diabetes on diurnal clock
523
gene expression (42), another study showed that clock genes in human adipose tissue show a
524
diurnal expression that relates to adiponectin, leptin, and glucocorticoid-related genes (11).
525
It should be noted that the low-fat group in our experiment gained more weight than
526
anticipated, resulting in no difference between low-fat diet and chronic overfeeding in regards
527
to total body mass. We believe this could be due to the relatively high sucrose content of this
528
low-fat diet. Also, we kept the mice at 28°C (i.e. close to thermoneutrality), which might have
529
facilitated fat gain also in the low-fat group. Nonetheless, the mice on low-fat diet ate
530
considerably less than the chronically overfed and weight cycled mice, and showed a strong
531
tendency towards expected values for several metabolic parameters such as glucose tolerance,
532
hormone levels, and lipid levels. In the additional feeding experiments we used low-fat diet
533
with lower sucrose content, and found the expected increase in body mass compared to high22
534
fat diet. Regardless of the higher than expected fat gain in the low-fat group of the weight
535
cycling experiment, the down-regulation of negative loop clock genes was a specific feature
536
of weight cycling compared to isocaloric chronic overfeeding.
537
In conclusion, weight cycling increased weight gain per calorie in C57BL/6J mice compared
538
to chronic overfeeding, which appeared to occur by a normalization of preadipocyte
539
recruitment without worsening metabolic health. This was further associated with a shift in
540
the expression of circadian clock transcription factors in adipose tissues. Our data suggest that
541
weight cycling may lead to a persistent suppression of peripheral clock gene expression,
542
possibly as an adaptive response to elevated cAMP signaling during repeated
543
fasting/refeeding. These molecular changes in adipose tissue may promote development of
544
new preadipocytes that are able to expand during overfeeding.
545
546
Acknowledgment
547
We thank Jan-Inge Bjune, Vivian Veum and Rita Holdhus for expert technical assistance.
548
549
Grants
550
Funding was provided by The Western Norway Regional Health Authority, NIFES (National
551
Institute of Nutrition and Seafood Research), The Carlsberg Foundation, the Novo Nordisk
552
Foundation, Meltzerfondet and KG Jebsen Center for Diabetes Research, University of
553
Bergen.
554
555
23
556
Disclosures
557
We declare no conflicts of interest.
558
24
559
REFERENCES
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Team EE: WHO launches the World Health Statistics 2012. Eurosurveillance 2012, 17(20):1515.
Obesity epidemic puts millions at risk from related diseases. Indian journal of medical
sciences 1997, 51(12):482.
Giovannucci E, Michaud D: The role of obesity and related metabolic disturbances in
cancers of the colon, prostate, and pancreas. Gastroenterology 2007, 132(6):2208-2225.
Hausman DB, DiGirolamo M, Bartness TJ, Hausman GJ, Martin RJ: The biology of white
adipocyte proliferation. Obes Rev 2001, 2(4):239-254.
Spiegelman BM, Flier JS: Adipogenesis and obesity: rounding out the big picture. Cell 1996,
87(3):377-389.
Hirsch J, Batchelor B: Adipose tissue cellularity in human obesity. Clin Endocrinol Metab
1976, 5(2):299-311.
Salans LB, Cushman SW, Weismann RE: Studies of human adipose tissue. Adipose cell size
and number in nonobese and obese patients. The Journal of clinical investigation 1973,
52(4):929-941.
Spalding KL, Arner E, Westermark PO, Bernard S, Buchholz BA, Bergmann O, Blomqvist L,
Hoffstedt J, Naslund E, Britton T et al: Dynamics of fat cell turnover in humans. Nature 2008,
453(7196):783-787.
Knittle JL, Hirsch J: Effect of early nutrition on the development of rat epididymal fat pads:
cellularity and metabolism. The Journal of clinical investigation 1968, 47(9):2091-2098.
Brook CGD, Lloyd JK, Wolf OH: Relation between Age of Onset of Obesity and Size and
Number of Adipose Cells. Brit Med J 1972, 2(5804):25-&.
Guller S, Corin RE, Mynarcik DC, London BM, Sonenberg M: Role of insulin in growth
hormone-stimulated 3T3 cell adipogenesis. Endocrinology 1988, 122(5):2084-2089.
Klaman LD, Boss O, Peroni OD, Kim JK, Martino JL, Zabolotny JM, Moghal N, Lubkin M, Kim
YB, Sharpe AH et al: Increased energy expenditure, decreased adiposity, and tissue-specific
insulin sensitivity in protein-tyrosine phosphatase 1B-deficient mice. Molecular and cellular
biology 2000, 20(15):5479-5489.
Ingle DJ, Prestrud MC, Nezamis JE: The effect of insulin upon the level of blood amino acids
in the eviscerated rat as related to the level of blood glucose. The American journal of
physiology 1947, 150(4):682-685.
Reynolds TH, Supiano MA, Dengel DR: Regional differences in glucose clearance: effects of
insulin and resistance training on arm and leg glucose clearance in older hypertensive
individuals. Journal of Applied Physiology 2007, 102(3):985-991.
Karnieli E, Armoni M: Transcriptional regulation of the insulin-responsive glucose
transporter GLUT4 gene: from physiology to pathology. American journal of physiology
Endocrinology and metabolism 2008, 295(1):E38-45.
Shepherd PR, Kahn BB: Mechanisms of disease - Glucose transporters and insulin action Implications for insulin resistance and diabetes mellitus. New Engl J Med 1999, 341(4):248257.
Jagadish MN, Fernandez CS, Hewish DR, Macaulay SL, Gough KH, Grusovin J, Verkuylen A,
Cosgrove L, Alafaci A, Frenkel MJ et al: Insulin-responsive tissues contain the core complex
protein SNAP-25 (synaptosomal-associated protein 25) A and B isoforms in addition to
syntaxin 4 and synaptobrevins 1 and 2. The Biochemical journal 1996, 317 ( Pt 3):945-954.
Kusari AB, Byon J, Bandyopadhyay D, Kenner KA, Kusari J: Insulin-induced mitogen-activated
protein (MAP) kinase phosphatase-1 (MKP-1) attenuates insulin-stimulated MAP kinase
activity: a mechanism for the feedback inhibition of insulin signaling. Mol Endocrinol 1997,
11(10):1532-1543.
25
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
O'Brien RM, Streeper RS, Ayala JE, Stadelmaier BT, Hornbuckle LA: Insulin-regulated gene
expression. Biochemical Society transactions 2001, 29(Pt 4):552-558.
Ducluzeau PH, Perretti N, Laville M, Andreelli F, Vega N, Riou JP, Vidal H: Regulation by
insulin of gene expression in human skeletal muscle and adipose tissue. Evidence for
specific defects in type 2 diabetes. Diabetes 2001, 50(5):1134-1142.
Plum L, Belgardt BF, Bruning JC: Central insulin action in energy and glucose homeostasis.
The Journal of clinical investigation 2006, 116(7):1761-1766.
Hao Q, Lillefosse HH, Fjaere E, Myrmel LS, Midtbo LK, Jarlsby RH, Ma T, Jia B, Petersen RK,
Sonne SB et al: High-glycemic index carbohydrates abrogate the antiobesity effect of fish oil
in mice. American journal of physiology Endocrinology and metabolism 2012, 302(9):E10971112.
Mehran AE, Templeman NM, Brigidi GS, Lim GE, Chu KY, Hu X, Botezelli JD, Asadi A, Hoffman
BG, Kieffer TJ et al: Hyperinsulinemia drives diet-induced obesity independently of brain
insulin production. Cell metabolism 2012, 16(6):723-737.
Mayer-Davis EJ, Levin S, Bergman RN, D'Agostino RB, Karter AJ, Saad MF: Insulin secretion,
obesity, and potential behavioral influences: results from the Insulin Resistance
Atherosclerosis Study (IRAS). Diabetes Metab Res 2001, 17(2):137-145.
Accili D, Nakae J, Kim JJ, Park BC, Rother KI: Targeted gene mutations define the roles of
insulin and IGF-I receptors in mouse embryonic development. Journal of pediatric
endocrinology & metabolism : JPEM 1999, 12(4):475-485.
Kido Y, Nakae J, Accili D: Clinical review 125: The insulin receptor and its cellular targets. The
Journal of clinical endocrinology and metabolism 2001, 86(3):972-979.
Siddle K: Signalling by insulin and IGF receptors: supporting acts and new players. Journal of
molecular endocrinology 2011, 47(1):R1-10.
de la Monte SM, Wands JR: Review of insulin and insulin-like growth factor expression,
signaling, and malfunction in the central nervous system: relevance to Alzheimer's disease.
Journal of Alzheimer's disease : JAD 2005, 7(1):45-61.
Leibiger IB, Leibiger B, Berggren PO: Insulin signaling in the pancreatic beta-cell. Annual
review of nutrition 2008, 28:233-251.
Escribano O, Arribas M, Valverde AM, Benito M: IRS-3 mediates insulin-induced glucose
uptake in differentiated IRS-2(-/-) brown adipocytes. Molecular and cellular endocrinology
2007, 268(1-2):1-9.
Boura-Halfon S, Zick Y: Phosphorylation of IRS proteins, insulin action, and insulin
resistance. American journal of physiology Endocrinology and metabolism 2009, 296(4):E581591.
Friedrichsen M, Poulsen P, Richter EA, Hansen BF, Birk JB, Ribel-Madsen R, Stender-Petersen
K, Nilsson E, Beck-Nielsen H, Vaag A et al: Differential aetiology and impact of
phosphoinositide 3-kinase (PI3K) and Akt signalling in skeletal muscle on in vivo insulin
action. Diabetologia 2010, 53(9):1998-2007.
Bjornholm M, Zierath JR: Insulin signal transduction in human skeletal muscle: identifying
the defects in Type II diabetes. Biochemical Society transactions 2005, 33(Pt 2):354-357.
Saltiel AR, Kahn CR: Insulin signalling and the regulation of glucose and lipid metabolism.
Nature 2001, 414(6865):799-806.
Warram JH, Martin BC, Krolewski AS, Soeldner JS, Kahn CR: Slow glucose removal rate and
hyperinsulinemia precede the development of type II diabetes in the offspring of diabetic
parents. Ann Intern Med 1990, 113(12):909-915.
Kitabchi AE, Temprosa M, Knowler WC, Kahn SE, Fowler SE, Haffner SM, Andres R, Saudek C,
Edelstein SL, Arakaki R et al: Role of insulin secretion and sensitivity in the evolution of type
2 diabetes in the diabetes prevention program: effects of lifestyle intervention and
metformin. Diabetes 2005, 54(8):2404-2414.
Morino K, Petersen KF, Shulman GI: Molecular mechanisms of insulin resistance in humans
and their potential links with mitochondrial dysfunction. Diabetes 2006, 55:S9-S15.
26
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
Shulman GI, Rothman DL, Jue T, Stein P, Defronzo RA, Shulman RG: Quantitation of Muscle
Glycogen-Synthesis in Normal Subjects and Subjects with Non-Insulin-Dependent Diabetes
by C-13 Nuclear Magnetic-Resonance Spectroscopy. New Engl J Med 1990, 322(4):223-228.
Samuel VT, Shulman GI: Mechanisms for insulin resistance: common threads and missing
links. Cell 2012, 148(5):852-871.
Rossetti L, Smith D, Shulman GI, Papachristou D, Defronzo RA: Correction of Hyperglycemia
with Phlorizin Normalizes Tissue Sensitivity to Insulin in Diabetic Rats. Journal of Clinical
Investigation 1987, 79(5):1510-1515.
Richter EA, Hansen BF, Hansen SA: Glucose-Induced Insulin Resistance of Skeletal-Muscle
Glucose-Transport and Uptake. Biochemical Journal 1988, 252(3):733-737.
Renstrom F, Buren J, Svensson M, Eriksson JW: Insulin resistance induced by high glucose
and high insulin precedes insulin receptor substrate 1 protein depletion in human
adipocytes. Metabolism-Clinical and Experimental 2007, 56(2):190-198.
McLarty JL, Marsh SA, Chatham JC: Post-translational protein modification by O-linked Nacetyl-glucosamine: Its role in mediating the adverse effects of diabetes on the heart. Life
sciences 2013, 92(11):621-627.
Marshall S, Bacote V, Traxinger RR: Discovery of a Metabolic Pathway Mediating GlucoseInduced Desensitization of the Glucose-Transport System - Role of Hexosamine
Biosynthesis in the Induction of Insulin Resistance. Journal of Biological Chemistry 1991,
266(8):4706-4712.
Kawanaka K, Han DH, Gao J, Nolte LA, Holloszy JO: Development of glucose-induced insulin
resistance in muscle requires protein synthesis. The Journal of biological chemistry 2001,
276(23):20101-20107.
Copeland RJ, Bullen JW, Hart GW: Cross-talk between GlcNAcylation and phosphorylation:
roles in insulin resistance and glucose toxicity. Am J Physiol-Endoc M 2008, 295(1):E17-E28.
Hanoune J, Defer N: Regulation and role of adenylyl cyclase isoforms. Annual review of
pharmacology and toxicology 2001, 41:145-174.
Bender AT, Beavo JA: Cyclic nucleotide phosphodiesterases: Molecular regulation to clinical
use. Pharmacol Rev 2006, 58(3):488-520.
Kass DA, Takimoto E, Nagayama T, Champion HC: Phosphodiesterase regulation of nitric
oxide signaling. Cardiovascular research 2007, 75(2):303-314.
Torii N: [The role of Ca ions and cAMP in ACTH lipolysis]. Nihon Naibunpi Gakkai zasshi
1985, 61(9):859-871.
Zhang C, Wang G, Zheng Z, Maddipati KR, Zhang X, Dyson G, Williams P, Duncan SA, Kaufman
RJ, Zhang K: Endoplasmic reticulum-tethered transcription factor cAMP responsive
element-binding protein, hepatocyte specific, regulates hepatic lipogenesis, fatty acid
oxidation, and lipolysis upon metabolic stress in mice. Hepatology 2012, 55(4):1070-1082.
de Rooij J, Zwartkruis FJT, Verheijen MHG, Cool RH, Nijman SMB, Wittinghofer A, Bos JL: Epac
is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature
1998, 396(6710):474-477.
Misra UK, Pizzo SV: Upregulation of mTORC2 activation by the selective agonist of EPAC, 8CPT-2Me-cAMP, in prostate cancer cells: assembly of a multiprotein signaling complex.
Journal of cellular biochemistry 2012, 113(5):1488-1500.
Mayati A, Levoin N, Paris H, N'Diaye M, Courtois A, Uriac P, Lagadic-Gossmann D, Fardel O, Le
Ferrec E: Induction of intracellular calcium concentration by environmental benzo(a)pyrene
involves a beta2-adrenergic receptor/adenylyl cyclase/Epac-1/inositol 1,4,5-trisphosphate
pathway in endothelial cells. The Journal of biological chemistry 2012, 287(6):4041-4052.
Laurent AC, Breckler M, Berthouze M, Lezoualc'h F: Role of Epac in brain and heart.
Biochemical Society transactions 2012, 40(1):51-57.
Bos JL: Epac: a new cAMP target and new avenues in cAMP research. Nat Rev Mol Cell Bio
2003, 4(9):733-738.
27
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
Yan J, Mei FC, Cheng H, Lao DH, Hu Y, Wei J, Patrikeev I, Hao D, Stutz SJ, Dineley KT et al:
Enhanced leptin sensitivity, reduced adiposity and improved glucose homeostasis in mice
lacking of Exchange protein directly activated by cAMP isoform 1. Molecular and cellular
biology 2012.
Ruiz-Hurtado G, Morel E, Dominguez-Rodriguez A, Llach A, Lezoualc'h F, Benitah JP, Gomez
AM: Epac in cardiac calcium signaling. Journal of molecular and cellular cardiology 2012.
Almahariq M, Tsalkova T, Mei FC, Chen H, Zhou J, Sastry SK, Schwede F, Cheng X: A Novel
EPAC-Specific Inhibitor Suppresses Pancreatic Cancer Cell Migration and Invasion.
Molecular pharmacology 2013, 83(1):122-128.
Yang Y, Shu X, Liu D, Shang Y, Wu Y, Pei L, Xu X, Tian Q, Zhang J, Qian K et al: EPAC null
mutation impairs learning and social interactions via aberrant regulation of miR-124 and
Zif268 translation. Neuron 2012, 73(4):774-788.
van Hooren KW, van Agtmaal EL, Fernandez-Borja M, van Mourik JA, Voorberg J, Bierings R:
The Epac-Rap1 signaling pathway controls cAMP-mediated exocytosis of Weibel-Palade
bodies in endothelial cells. The Journal of biological chemistry 2012, 287(29):24713-24720.
Insel PA, Murray F, Yokoyama U, Romano S, Yun H, Brown L, Snead A, Lu D, Aroonsakool N:
cAMP and Epac in the regulation of tissue fibrosis. British journal of pharmacology 2012,
166(2):447-456.
Roscioni SS, Maarsingh H, Elzinga CR, Schuur J, Menzen M, Halayko AJ, Meurs H, Schmidt M:
Epac as a novel effector of airway smooth muscle relaxation. Journal of cellular and
molecular medicine 2011, 15(7):1551-1563.
Suzuki S, Yokoyama U, Abe T, Kiyonari H, Yamashita N, Kato Y, Kurotani R, Sato M, Okumura
S, Ishikawa Y: Differential roles of Epac in regulating cell death in neuronal and myocardial
cells. The Journal of biological chemistry 2010, 285(31):24248-24259.
Fain JN, Garciasainz JA: Adrenergic Regulation of Adipocyte Metabolism. Journal of lipid
research 1983, 24(8):945-966.
Carmen GY, Victor SM: Signalling mechanisms regulating lipolysis. Cellular signalling 2006,
18(4):401-408.
Omar B, Zmuda-Trzebiatowska E, Manganiello V, Goransson O, Degerman E: Regulation of
AMP-activated protein kinase by cAMP in adipocytes: Roles for phosphodiesterases,
protein kinase B, protein kinase A, Epac and lipolysis. Cellular signalling 2009, 21(5):760766.
He YH, Zhang HQ, Teng JH, Huang LN, Li Y, Sun CH: Involvement of calcium-sensing receptor
in inhibition of lipolysis through intracellular cAMP and calcium pathways in human
adipocytes. Biochemical and biophysical research communications 2011, 404(1):393-399.
Hallenborg P, Feddersen S, Madsen L, Kristiansen K: The tumor suppressors pRB and p53 as
regulators of adipocyte differentiation and function. Expert opinion on therapeutic targets
2009, 13(2):235-246.
Rosen ED, Spiegelman BM: Molecular regulation of adipogenesis. Annu Rev Cell Dev Bi 2000,
16:145-171.
Petersen RK, Madsen L, Pedersen LM, Hallenborg P, Hagland H, Viste K, Doskeland SO,
Kristiansen K: Cyclic AMP (cAMP)-mediated stimulation of adipocyte differentiation
requires the synergistic action of Epac- and cAMP-dependent protein kinase-dependent
processes. Molecular and cellular biology 2008, 28(11):3804-3816.
Zilberfarb V, Siquier K, Strosberg AD, Issad T: Effect of dexamethasone on adipocyte
differentiation markers and tumour necrosis factor-alpha expression in human PAZ6 cells.
Diabetologia 2001, 44(3):377-386.
Hernandez R, Teruel T, Lorenzo M: Insulin and dexamethasone induce GLUT4 gene
expression in foetal brown adipocytes: synergistic effect through CCAAT/enhancer-binding
protein alpha. Biochemical Journal 2003, 372:617-624.
28
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
Shin SM, Kim K, Kim JK, Yoon SR, Choi I, Yang Y: Dexamethasone reverses TGF-betamediated inhibition of primary rat preadipocyte differentiation. FEBS letters 2003, 543(13):25-30.
Cristancho AG, Lazar MA: Forming functional fat: a growing understanding of adipocyte
differentiation. Nature reviews Molecular cell biology 2011, 12(11):722-734.
Madsen L, Kristiansen K: The importance of dietary modulation of cAMP and insulin
signaling in adipose tissue and the development of obesity. Annals of the New York
Academy of Sciences 2010, 1190:1-14.
Neal JW, Clipstone NA: Calcineurin mediates the calcium-dependent inhibition of adipocyte
differentiation in 3T3-L1 cells. Journal of Biological Chemistry 2002, 277(51):49776-49781.
Zhang P, O'Loughlin L, Brindley DN, Reue K: Regulation of lipin-1 gene expression by
glucocorticoids during adipogenesis. Journal of lipid research 2008, 49(7):1519-1528.
Shimizu-Albergine M, Tsai LC, Patrucco E, Beavo JA: cAMP-specific phosphodiesterases 8A
and 8B, essential regulators of Leydig cell steroidogenesis. Molecular pharmacology 2012,
81(4):556-566.
Li H, Li H, Guo H, Liu F: Cholesterol suppresses adipocytic differentiation of mouse adiposederived stromal cells via PPARgamma2 signaling. Steroids 2013, 78(5):454-461.
Ma T, Liaset B, Hao Q, Petersen RK, Fjaere E, Ngo HT, Lillefosse HH, Ringholm S, Sonne SB,
Treebak JT et al: Sucrose Counteracts the Anti-Inflammatory Effect of Fish Oil in Adipose
Tissue and Increases Obesity Development in Mice. PloS one 2011, 6(6).
Hao Q, Lillefosse HH, Fjaere E, Myrmel LS, Midtbo LK, Jarlsby RH, Ma T, Jia BB, Petersen RK,
Sonne SB et al: High-glycemic index carbohydrates abrogate the antiobesity effect of fish oil
in mice. Am J Physiol-Endoc M 2012, 302(9):E1097-E1112.
Madsen L, Pedersen LM, Liaset B, Ma T, Petersen RK, van den Berg S, Pan J, Muller-Decker K,
Dulsner ED, Kleemann R et al: cAMP-dependent signaling regulates the adipogenic effect of
n-6 polyunsaturated fatty acids. Journal of Biological Chemistry 2008, 283(11):7196-7205.
Konstantinova I, Nikolova G, Ohara-Imaizumi M, Meda P, Kucera T, Zarbalis K, Wurst W,
Nagamatsu S, Lammert E: EphA-Ephrin-A-mediated beta cell communication regulates
insulin secretion from pancreatic islets. Cell 2007, 129(2):359-370.
Grimditch GK, Barnard RJ, Hendricks L, Weitzman D: Peripheral Insulin Sensitivity as
Modified by Diet and Exercise Training. American Journal of Clinical Nutrition 1988,
48(1):38-43.
Chicco A, D'Alessandro ME, Karabatas L, Pastorale C, Basabe JC, Lombardo YB: Muscle lipid
metabolism and insulin secretion are altered in insulin-resistant rats fed a high sucrose
diet. Journal of Nutrition 2003, 133(1):127-133.
Shanik MH, Xu Y, Skrha J, Dankner R, Zick Y, Roth J: Insulin resistance and hyperinsulinemia:
is hyperinsulinemia the cart or the horse? Diabetes care 2008, 31 Suppl 2:S262-268.
Ferreira Mdel R, Lombardo YB, Chicco A: beta-Cell adaptation/dysfunction in an animal
model of dyslipidemia and insulin resistance induced by the chronic administration of a
sucrose-rich diet. Islets 2010, 2(6):367-373.
Renstrom F, Buren J, Svensson M, Eriksson JW: Insulin resistance induced by high glucose
and high insulin precedes insulin receptor substrate 1 protein depletion in human
adipocytes. Metabolism: clinical and experimental 2007, 56(2):190-198.
Silva CM, Sato S, Margolis RN: No time to lose: workshop on circadian rhythms and
metabolic disease. Genes & development 2010, 24(14):1456-1464.
Lin YC, Hsiao TJ, Chen PC: Persistent Rotating Shift-Work Exposure Accelerates
Development of Metabolic Syndrome among Middle-Aged Female Employees: A Five-Year
Follow-Up. Chronobiology international 2009, 26(4):740-755.
809
810
29
811
812
813
Figure captions
814
815
816
Fig. 1 Weight cycling led to increased feed efficiency, hyperglycemia, and fat mass. Young male
817
C5BL/6J mice were fed low-fat or high fat/high sucrose with or without intermittent periods of
818
energy restriction, for a total of 11 weeks. A. Body weight development throughout the experiment.
819
B. Feed efficiency (weight gained relative to the amount of energy ingested) for each 10 days period
820
of weight regain after energy restriction. C. Tissues were dissected and weighed on the final day of
821
the experiment. All data are presented as mean ± SEM (n=11 per group). CO, chronic overfeeding;
822
WC, weight cycling.
823
*, p < 0.05 vs. CO.
824
825
Fig. 2 Increased adipocyte number and size after weight cycling. 5µm thick sections of paraffin-
826
embedded fat tissue from two different locations were hematoxylin and eosin (H&E) stained. The
827
photographs are representative of three to six microscopic images that were taken from random
828
places of each section. The average area of each cell on each image was measured by drawing the
829
circumference of each cell (µm2), and based on this the average cell volume (µm3) was estimated.
830
Cell size distribution was assessed by segregating the cell sizes into 20 brackets (given as percent).
831
A/D. Adipose tissue morphology. B/E. Relative adipocyte number in each fat pad between the groups
832
was estimated by dividing total fat pad mass by average adipocyte volume. C/F. Average adipocyte
833
size and cell size distribution. Data are presented as mean ± SEM (n=5 per group).
834
§, p < 0.05 vs. low-fat.
30
835
836
837
Fig. 3. Weight cycling did not worsen plasma parameters compared to chronic overfeeding despite
838
fat gain. Glucose tolerance test (GTT) was performed one week prior to sacrifice by injecting glucose
839
dissolved in saline into the intraperitoneal cavity (two grams per kg body weight). Blood was drawn
840
from the tail at 0, 15, 30, 60 and 120 minutes after injection. All other parameters were measured in
841
plasma collected at sacrifice. Data are presented as mean ± SEM (n=11 per group). LF, low-fat diet;
842
CO, chronic overfeeding; WC, weight cycling.
843
§, p<0.05 vs. low-fat; §§, p<0.01 vs. low-fat; *, p<0.05 vs. CO; **, p<0.01 vs. CO.
844
845
Fig. 4 Clock genes are down-regulated after WC but not after chronic overfeeding. Tissues from all
846
mice in each group were dissected and weighed on the final day of the experiment, and mRNA levels
847
were measured by qPCR relative to Tbp as a reference gene. A. Clock gene expression in iWAT and
848
eWAT (n=11 per group). As an additional control of a WC specific effect, gene expression was also
849
measured in adipose tissue from another overfeeding experiment in C57BL/6J mice (n=5-7 per
850
group) and an overfeeding experiment of obesity-resistant SV129 mice (n=7 per group). B. Clock
851
gene mRNA expression in liver relative to Tbp mRNA (n=11 per group). Data are presented as median
852
± SEM. LF, low-fat diet; CO, chronic overfeeding; WC, weight cycling.
853
§, p<0.05 vs. low-fat; §§, p<0.01 vs. low-fat; *, p<0.05 vs. CO; **, p<0.01 vs. CO.
854
855
Fig. 5 Altered expression of Dbp and Tef during adipogenesis. A/B. 3T3-L1 cells were seeded in 6-
856
well plates and induced to differentiate into adipocytes two days post-confluence, by changing from
857
10% calf serum to 10% fetal calf serum and adding dexamethasone (0.5mM), insulin (175nM) and
31
858
IBMX (0.5M). Lysates were collected at different time-points during differentiation. C/D. 3T3-L1 cells
859
were seeded in 24-well plates, and were treated two days post-confluence with 10% fetal calf serum,
860
dexamethasone (0.5mM), insulin (175nM) or IBMX (0.5M) and lysed after 24 hours. Dbp and Tef
861
(transcript variant 2) mRNA was measured by qPCR. Data are presented as mean ± SEM of triplicate
862
measurements and representative of two independent experiments. FCS, fetal calf serum; DEX,
863
dexamethasone; IBMX, phosphodiesterase inhibitor methylisobutylxanthine.
864
865
866
867
868
869
870
871
872
873
Tables
874
875
876
Table 1. Primers and probes used for qPCR.
Forward (left) primer
Reverse (right) primer
Dbp
5´-aggaagcagctggcacac-3´
5´-tcttcggttctgaaaccatactt-3´
30
Tef (variant 1)
5'-ggtcctgaagaagctgatgg-3´
5'-ggtactggctgctgctgact-3´
78
Tef (variant 2)
5'-ctggagcattctttgccttg-3´
5'-ggtactggctgctgctgact-3´
78
Per1
5'-gcttcgtggacttgacacct-3´
5'-tgctttagatcggcagtggt-3´
71
Per2
5'-ccctgatgatgccttcaga-3´
5'-gctttggcagactgctcact-3´
64
Target gene
UPL Probe
32
Per3
5'-gtgaagccagtggcagaga-3´
5'-tgagaggaagaaaagtccttctg-3´
9
Nr1d2
5'-gagaacaagaatagttacctgtgcaa-3´
5'-ggagacttactcattggacaaacc-3´
65
Bmal1
5'-tcagatgacgaactgaaacacc-3´
5'-cggtcacatcctacgacaaa-3´
74
Clock
5'-agccaccacagcagttcttac-3´
5'-cgaaggagggaaagtgctc-3´
74
Rplp0
5'-actggtctaggacccgagaag-3´
5'-tcccaccttgtctccagtct-3´
9
877
878
UPL, Universal Probe Library (Roche)
879
880
881
882
Table 2. Correlation between feed efficiency and circulating leptin.
883
WC
Leptin
Feed efficiency
rho
pvalue
.764
.006
CO
Leptin
Feed efficiency
rho
pvalue
.269
.424
884
885
WC, weight cycling; Co, chronic overfeeding.
886
33
887
Table 3. Differentially expressed genes after weight cycling with similar effects in iWAT, eWAT, and iBAT (q-value < 0.05, fold change > 1.2).
888
iWAT
Symbol
Gene ID
WC
eWAT
CO
WC/CO
iBAT
WC
CO
WC/CO
WC
CO
WC/CO
More up-regulated after WC vs. CO in eWAT than in iWAT
Fgf13
14168
2 072
677
2.38
625
501
1.29
123
124
1.01
Ubd
24108
1 022
361
1.90
783
696
1.21
137
144
-1.06
Tph2
216343
587
161
2.64
238
159
1.69
116
119
-1.02
Sfrp5
54612
1 038
610
1.64
283
243
1.24
142
136
1.01
Lipf
67717
265
170
1.74
162
136
1.33
112
117
-1.06
More down-regulated after WC vs. CO in eWAT than in iWAT
Mycl1
16918
253
557
-1.86
419
493
-1.29
176
182
-1.02
Inmt
21743
2 801
4 179
-1.63
1 293
1 831
-1.21
218
255
-1.13
Higher expression after WC vs. CO in all fat pads
Mup3
17842
677
285
1.82
1 452
604
1.79
210
135
1.52
Apoa2
11807
341
161
2.03
1 026
434
1.55
169
164
1.35
Ttr
22139
296
169
1.76
921
357
1.50
163
126
1.40
34
Serpina1b
20701
583
332
1.59
716
336
1.49
133
133
1.31
Serpina1d
20703
427
308
1.53
609
294
1.53
141
125
1.22
Serpina1b
20701
551
364
1.51
629
337
1.45
146
135
1.27
Mup1
17840
1 292
1 292
1.20
1 146
757
1.47
169
155
1.28
Lower expression after WC vs. CO in all fat pads
Cyp2e1
13106
9 536
14 708
-1.44
22 746
29 085
-1.32
452
935
-1.79
Dbp
13170
757
1 361
-1.62
1 039
1 657
-1.56
444
667
-1.33
Per2
18627
616
962
-1.47
893
1 111
-1.31
507
1 067
-1.65
Per2
18627
609
744
-1.33
646
899
-1.36
368
786
-1.60
Fseg
213393
592
793
-1.29
736
1 306
-1.50
645
1 057
-1.54
Tef
21685
589
720
-1.31
616
825
-1.39
649
710
-1.23
Gpx3
14778
5 753
7 565
-1.26
7 096
10 426
-1.47
16 067
18 251
-1.20
Nr1d2
353187
218
294
-1.23
242
368
-1.32
178
261
-1.25
Prodh
19125
336
414
-1.27
311
445
-1.35
256
313
-1.21
Usp2
53376
321
319
-1.23
520
719
-1.26
464
688
-1.28
889
890
891
A rank product meta-analysis was performed on rank product microarray data for iWAT (n=11), eWAT (n=10) and iBAT (n=11) comparing WC and CO. Genes with log2 fold
change > 1.2 q-value < 0.05 and their median signal intensities (quantile normalized Illumina microarray data) are shown. WC, weight cycling; CO, chronic overfeeding.
35
892
Table 4. Correlation between feed efficiency and gene expression in eWAT and iWAT.
WC
893
894
Feed efficiency
Leptin
eWAT (n=10)
iWAT (n=11)
eWAT (n=10)
iWAT (n=11)
Gene
Rho
p-value
Rho
p-value
rho
p-value
rho
p-value
Fgf13
.867
.0012
.764
.0062
.903
.0003
.900
.0002
Tph2
.818
.0038
.764
.0062
.915
.0002
.900
.0002
Sfrp5
.685
.0289
.536
.0890
.539
.1076
.773
.0053
Ubd
.879
.0008
.345
.2981
.903
.0003
.182
.5926
Lipf
.903
.0003
.809
.0026
.794
.0061
.936
.0000
Mycl1
-.624
.0537
-.445
.1697
-.685
.0289
-.482
.1334
Inmt
-.527
.1173
-.564
.0710
-.891
.0005
-.436
.1797
CO
Feed efficiency
Leptin
eWAT (n=10)
iWAT (n=11)
eWAT (n=10)
iWAT (n=11)
Gene
Rho
p-value
Rho
p-value
rho
p-value
rho
p-value
Fgf13
.418
.2291
.387
.2393
.770
.0092
.800
.0031
Tph2
.539
.1076
.292
.3843
.576
.0816
.882
.0003
Sfrp5
.273
.4458
.460
.1544
.939
.0001
.900
.0002
Ubd
.236
.5109
-.278
.4080
-.139
.7009
.145
.6696
Lipf
.164
.6515
.260
.4406
.745
.0133
.909
.0001
Mycl1
-.382
.2763
-.515
.1051
-.503
.1383
-.136
.6893
Inmt
-.139
.7009
.132
.6986
-.564
.0897
.518
.1025
WC, weight cycling; CO, chronic overfeeding; rho, Spearman’s rho.
36
895
896
897
Table 5. Over-represented PANTHER gene ontology categories for differentially expressed genes in
eWAT or iWAT.
Biological Process
REFLIST (26185)
# genes
# expected
P-value
Unclassified
11454
11
29.74
0.0002
Developmental process
2262
19
5.87
0.0006
mesoderm development
819
10
2.13
0.0085
system development
1311
10
3.4
0.3420
188
6
0.49
0.0018
Cellular process
6077
30
15.78
0.0195
Inflammation mediated by chemokine and cytokine
signaling pathway
287
6
0.75
0.0187
Oxygen and reactive oxygen species metabolic process
56
4
0.1
0.0007
Defense response to bacterium
53
3
0.1
0.0244
1851
8
0.99
0.0002
muscle contraction
337
7
0.18
0.0000
Developmental process
2262
7
1.21
0.0122
muscle organ development
180
5
0.1
0.0000
mesoderm development
819
5
0.44
0.0081
system development
1311
5
0.7
0.0735
11454
0
6.12
0.0544
Up-regulated WC vs. CO, eWAT not iWAT (68)
Angiogenesis
Up-regulated WC vs. CO, iWAT not eWAT (2)
N/A (only 2 genes)
Down-regulated WC vs. CO, eWAT not iWAT (49)
Down-regulated WC vs. CO, iWAT not eWAT (14)
System process
Unclassified
898
899
900
WC, weight cycling; CO, chronic overfeeding.
37
901
902
38
1
2
3
4
5