Content uploaded by Fiona Grennan-jones
Author content
All content in this area was uploaded by Fiona Grennan-jones
Content may be subject to copyright.
Gsasignalling suppresses PPARg2 generation and inhibits 3T3L1
adipogenesis
Lei Zhang, Carol Paddon, Mark D Lewis, Fiona Grennan-Jones and Marian Ludgate
School of Medicine, Centre for Endocrine and Diabetes Sciences, Cardiff University, Heath Park, Cardiff CF14 4XN, UK
(Correspondence should be addressed to M Ludgate; Email: ludgate@cf.ac.uk)
Abstract
Since TSH receptor (TSHR) expression increases during
adipogenesis and signals via cAMP/phospho-cAMP-response
element binding protein (CREB), reported to be necessary
and sufficient for adipogenesis, we hypothesised that TSHR
activation would induce preadipocyte differentiation. Retro-
viral vectors introduced constitutively active TSHR
(TSHR*) into 3T3L1 preadipocytes; despite increased
cAMP (RIA) and phospho-CREB (western blot) there was
no spontaneous adipogenesis (assessed morphologically, using
oil red O and QPCR measurement of adipogenesis markers).
We speculated that Gbg signalling may be inhibitory but
failed to induce adipogenesis using activated Gsa(gsp*).
Inhibition of phosphodiesterases did not promote adipogen-
esis in TSHR* or gsp* populations. Furthermore, differen-
tiation induced by adipogenic medium with pioglitazone was
reduced in TSHR* and abolished in gsp* expressing 3T3L1
cells. TSHR* and gsp* did not inactivate PPARg(PPARG as
listed in the HUGO database) by phosphorylation but
expression of PPARg1 was reduced and PPARg2 undetect-
able in gsp*. FOXO1 phosphorylation (required to inactivate
this repressor of adipogenesis) was lowest in gsp* despite the
activation of AKT by phosphorylation. PROF is a mediator
that facilitates FOXO1 phosphorylation by phospho-Akt. Its
transcript levels remained constantly low in the gsp*
population. In most measurements, the TSHR* cells were
between the gsp* and control 3T3L1 preadipocytes. The
enhanced down-regulation of PREF1 (adipogenesis
inhibitor) permits retention of some adipogenic potential in
the TSHR* population. We conclude that Gsasignalling
impedes FOXO1 phosphorylation and thus inhibits PPARg
transcription and the alternative promoter usage required to
generate PPARg2, the fat-specific transcription factor
necessary for adipogenesis.
Journal of Endocrinology (2009) 202, 207–215
Introduction
Adipose tissue expands via two mechanisms, hypertrophy of
individual adipocytes and hyperplasia due to the proliferation
and differentiation of preadipocyte precursors (Drolet et al.
2008). In vitro protocols to induce preadipocyte differentiation
(adipogenesis) use agents to increase cAMP. Furthermore,
studies in the 3T3L1 murine preadipocytes, the cell line that
has been central in unravelling the complex mechanisms
driving adipogenesis, indicated that treatment with forskolin
or isobutylmethylxanthine (IBMX) alone suffices to trigger
the process (Boone et al. 1999) and it has been shown
subsequently that cAMP-response element binding protein
(CREB) activation is both necessary and sufficient to induce
adipogenesis (Reusch et al. 2000).
Many G protein-coupled receptors, including the thyro-
tropin receptor (TSHR), which is the main regulator of
thyroid function and growth, signal predominantly via the
cAMP/protein kinase A (PKA) pathway, ultimately leading
to phosphorylation of CREB (Vassart & Dumont 1992).
TSHR expression is also increased during adipogenesis
(Haraguchi et al. 1996). Does this suggest a role for TSHR
activation in adipocyte biology? The question is clinically
relevant since thyroid dysfunction is common, affecting up to
2% of the population, and the majority of these individuals
will have overactivation of the TSHR either by supraphy-
siological concentrations of TSH in hypothyroidism or from
thyroid stimulating antibodies in Graves’ hyperthyroidism
(Vanderpump et al. 1995).
On the basis of the known signalling pathways following
from TSHR activation, we hypothesise that it should trigger
adipogenesis spontaneously or at least enhance that induced
by known differentiating agents such as synthetic PPARg
(PPARG as listed in HUGO database) agonists. We have used
an in vitro model in which naturally occurring constitutively
active mutant forms of TSHR, found in e.g. toxic nodules
(Paschke & Ludgate 1997), are introduced using retroviral
vectors (Fuhrer et al. 2003). Our previous studies used the
model to investigate the effect of TSHR activation on human
orbital primary preadipocytes (of relevance to the eye disease
occurring in Graves’ patients). We demonstrated spontaneous
induction of early, but inhibition of the later stages of
adipogenesis, even in the presence of PPARgagonists (Zhang
et al. 2006). Orbital preadipocytes are a particular fat depot,
207
Journal of Endocrinology (2009) 202, 207–215 DOI: 10.1677/JOE-09-0099
0022–0795/09/0202–207 q2009 Society for Endocrinology Printed in Great Britain Online version via http://www.endocrinology-journals.org
This is an Open Access article distributed under the terms of the Society for Endocrinology’s Re-use Licence which permits unrestricted non-commercial use,
distribution, and reproduction in any medium, provided the original work is properly cited.
being derived from the neural crest, and this may explain the
differences between our results following activation of the
cAMP pathway and those reported using a murine
preadipocyte cell line. In the current study, we have applied
the model (activating mutant TSHR–L629F) to 3T3L1. To
our surprise, despite modestly increasing intracellular cAMP
and elevating phospho-CREB levels, TSHR activation did
not induce adipogenesis and this was not changed by adding a
phosphodiesterase inhibitor. Furthermore, adipogenesis
induced by PPARgagonists was significantly reduced in the
mutant TSHR expressing cells compared with the non-
modified population. Since TSHR activation liberates two
functional moieties of the Gs protein, a(which acts via PKA)
and bg (signals via PI3K), we suggested that the inhibitory
effects were due to the latter. Thus, we have introduced
constitutively active Gsa(gsp*), Q227L (Ludgate et al. 1999),
which yields only the asubunit. Again there was no
spontaneous adipogenesis and in vitro induced differentiation
was completely abolished. Further investigations revealed a
reduction in PPARg1 and the complete absence of PPARg2
proteins, the fat-specific isoform, in gsp* expressing 3T3L1.
Materials and Methods
Reagent source; cell culture and adipogenesis protocols
All chemicals were obtained from Sigma–Aldrich and tissue
culture media and serum from BioWhittaker-Lonza
(Verviers, Belgium) unless otherwise stated. The 3T3L1 cell
line was purchased from the ATCC (Atlanta, GA, USA).
3T3L1 murine preadipocytes were routinely cultured in
DMEM/F12 10% FCS (complete medium, CM). Adipogen-
esis was induced in confluent cells by replacing with
differentiation medium (DM) containing 5% FCS, biotin
(33 mM), panthothenate (17 mM), tri-iodothyronine (1 nM),
dexa-methasone (100 nM), thiazolidinedione (1 mM) and
insulin (500 nM), for 10–12 days, as previously described
(Zhang et al. 2006).
Activating mutant human TSHR, L629F (TSHR*) and rat
Gsa, Q227L (gsp*), were introduced using retroviral vectors,
previously produced in our laboratory (Ivan et al. 1997,
Fuhrer et al. 2003). All experiments were performed on the
mixed pools of cells that resulted from geneticin selection.
Furthermore, at least two independent mixed pools were
generated for each cell population, with results being
comparable. Initial characterisation of the G418 resistant
pools of 3T3L1 employed RT-PCR and direct sequencing to
demonstrate the expression of human TSHR and rat Gsain
the appropriate populations as previously described (Ivan et al.
1997,Fuhrer et al. 2003). In many of our experiments we also
included 3T3L1 populations transduced with the wild-type
(WT) and a second mutant form of TSHR, M453T. Results
obtained (basal levels of cAMP/pCREB, transcript measures
for PPARgand GPDH, etc.) indicated that there was no
significant difference in the behaviour of WT compared
with non-modified or M453T compared with L629F. In the
interests of brevity, the results are reported for the non-
modified, L629F and gsp* expressing populations.
Effect on cAMP levels
The non-modified 3T3L1 and populations expressing gsp*or
TSHR* were plated at 5!10
4
/well in 12-well plates. The
following day they were incubated for 4 h in medium
containing 10
K4
M IBMX with no further addition (basal
conditions), or plus 10 mU/ml bovine TSH. Forskolin (in
addition to IBMX) at 10
K5
M was included as a positive
control in all populations. cAMP was extracted in 0.1 M HCl
and measured using an in-house RIA capable of detecting
femtogram quantities of the second messenger, as previously
described (Fuhrer et al. 2003). Coulter counting of adjacent
wells provided an accurate cell number for results to be
expressed as picomoles cAMP/10
4
cells.
Western blotting
The three different 3T3L1 populations were propagated in
duplicate in 6-well plates in CM or DM. Proteins were
extracted, at various time points, in Laemmli buffer
containing 1 mM sodium orthovanadate and 1 mM phenyl-
methylsulphonyl fluoride. To prepare nuclear and cytosolic
fractions, cells were harvested in ice-cold PBS, centrifuged
and resuspended in HEPES buffer containing protease
inhibitors. The supernatant produced by centrifugation at
12 000 gprovided the cytosolic and high-salt extraction of the
pellet, the nuclear fractions respectively.
Samples (containing 20 mg protein) were separated by 10%
SDS-PAGE and then the gel electroblotted onto PVDF
membrane as previously described (Al-Khafaji et al. 2005).
To investigate various signalling pathways, the blots were
probed with the following rabbit antibodies (all from Cell
Signalling Technology unless otherwise stated): anti
phospho-CREB (Ser 133, 1:2000 overnight; 4 8C); anti
total-CREB (1:1000, room temperature, 1 h; Santa Cruz
Biotechnology, Santa Cruz, CA, USA); anti phospho-Akt
(Thr 308, 1:1000, 4 8C, overnight); anti total-Akt (1:1000,
room temperature, 1 h, Santa Cruz); and anti phospho-SYK
(Tyr 352, 1:5000, overnight, 4 8C, BD Biosciences, San Jose,
CA, USA). To investigate transcription factors, rabbit
antibodies to phospho-PPARg(Ser 82, 1:1000 overnight,
48C, Upstate), anti total-PPARg(1:1000, overnight, 4 8C),
anti total-FOXO1 (1:1000, overnight, 4 8C) and anti
phospho-FOXO1 (Ser 256, 1:1000, overnight, 4 8C) were
employed. In all cases proteins were detected using either an
anti-mouse IgG-HRP conjugate or an anti-rabbit IgG-HRP
conjugate (1:5000, room temperature for 1 h, GE Health-
care, Amersham, UK) and visualised by enhanced chemi-
luminescence (ECL Plus, GE Healthcare).
Films were analysed using the Alpha Imager 1200
digital imaging system (Alpha Innotech Corp., San Leandro,
CA, USA). The blots were initially probed with the
L ZHANG and others .Gsasignalling, PPARg2 and adipogenesis208
Journal of Endocrinology (2009) 202, 207–215 www.endocrinology-journals.org
phospho-specific antibodies; they were then stripped and
reprobed with antibodies that recognise total proteins.
Effect on spontaneous and PPARginduced adipogenesis
The various cell populations (in 12-well plates) were
examined in CM and DM. Microscopic examination
provided a means of determining whether morphological
changes, for e.g. rounding-up of cells and/or acquisition of
lipid-filled droplets (oil red O staining), had occurred.
Effect on mitotic clonal expansion
To investigate whether the expression of the various mutants
had any effect on the mitotic clonal expansion phase, reported
by some authors to be required for differentiation in 3T3L1
preadipocytes (Tang et al. 2003), the cells were counted 1, 2, 3
and 10 days after addition of DM, using a Coulter particle
counter; results are expressed (meanGS.E.M.) fold increase in
cell number.
QRT-PCR measurement of transcript copy number
The various cell populations were plated in 6-well plates in
CM or DM. Ten days later, RNA was extracted, reverse
transcribed and transcript copy numbers for PPARg,GPDH,
PREF1,ID2,PROF, and acidic ribosomal phosphoprotein
(ARP) were measured using Sybr green and a Stratagene
MX3000 light cycler. Primers used are listed in Ta b l e 1 .
Standard curves (the PCR amplicon subcloned into
pGEM-T at 10
6
to 10
2
copies) were included for each gene
and results are expressed relative to the housekeeping gene
ARP. This gene was selected on the basis of its high expression
level (C
t
value 16G0.29 in non-modified 3T3L1 in CM),
which was not modified in the L629F and gsp* populations
(16G0.31 and 16G0.32 respectively in CM) or during
adipocyte differentiation (16G0.28 in non-modified cells on
day 9 in DM; the C
t
values are the meanGS.E.M. obtained
from four separate experiments).
Statistical analysis
Means were compared using Student’s t-test for parametric
data and non-parametric data were analysed using the
Wilcoxon signed ranks test.
Results
TSHR* and gsp* increase unstimulated cAMP levels
In non-modified 3T3L1, the unstimulated level of cAMP was
397G18 pmol/10
4
cells (results are the meanGS.E.M. of three
experiments all performed in at least triplicate). Unstimulated
cAMP levels were significantly increased in cells expressing
TSHR* or gsp* reaching 129G3.9% (P!0.02) and
140G5.9% (P!0.01) respectively of the non-modified
population. TSH (1 mU/ml) elicited a cAMP response in
cells expressing mutant TSHR (650G218% of unstimulated)
but not in the non-modified or gsp* 3T3L1 cells, in keeping
with the very low level of endogenous TSHR in these cells
prior to differentiation (Haraguchi et al. 1996). There was no
significant difference in the response to forskolin, which
induced a robust O20-fold increase in cAMP in all the three
populations.
Phosphorylated CREB levels are increased in TSHR* and gsp*
expressing cells and are modulated during adipogenesis
Western blots to investigate activation of CREB by
phosphorylation are usually performed following acute
stimulation of the cells followed by analysis in the minutes
after exposure to the stimulant. By contrast, we have
examined the basal levels of phospho-CREB in the three
populations of 3T3L1 to assess the effects of chronic activation
via the various mutants. We have used a value of 1 for the
phosphorylated:total CREB ratio of the non-modified
preadipocytes. Based on at least seven experiments using
two independently generated populations of TSHR* or gsp*
expressing cells, there was a modest, but significant increase in
the ratio in L629F (1.5G0.58 P!0.02) and gsp*(1
.19G0.22
P!0.02), reflecting chronic activation of this pathway.
A representative western blot is shown in Fig. 1.
To investigate whether CREB phosphorylation is affected
by the adipogenic process, we have measured the phosphor-
ylated:total CREB ratios in samples obtained from the
different 3T3L1 populations at various time points following
addition of DM. In the non-modified cells there is a slight
reduction in the ratio in mid-differentiation, but a significant
increase in the later stages (P!0.05), to achieve levels
150–300% of those on day 0 (prior to addition of DM). Apart
from the higher basal levels of phospho-CREB, a similar
Table 1 Primer sequences (and amplicon sizes) employed for QPCR measurements
Forward primer Reverse primer
Gene
PPARg220 bp TTTTCAAGGGTGCCAGTTTC AATCCTTGGCCCTCTGAGAT
GPDH 124 bp ATGCTCGCCACAGAATCCACAC AACCGGCAGCCCTTGACTTG
PREF1 148 bp CGTGATCAATGGTTCTCCCT AGGGGTACAGCTGTTGGTTG
ARP 72 bp GAGGAATCAGATGAGGATATGGGA AAGCAGGCTGACTTGGTTGC
PROF 72 bp GATCACTCTGTCATCATGTGG CTTACTGTCCCACATTTGCTTG
ID2 141 bp GGACCACAGCTTGGGCAT CGTTCATGTTGTAGAGCAGACTCAT
Gsasignalling, PPARg2 and adipogenesis .L ZHANG and others 209
www.endocrinology-journals.org Journal of Endocrinology (2009) 202, 207–215
pattern of expression was apparent in the TSHR* and gsp*
expressing cells through adipogenesis (data not shown).
The increased phospho-CREB levels do not produce spontaneous
adipogenesis
In CM, even when the cells had been confluent for up to
10 days, there were no morphological changes consistent with
adipogenesis in cells expressing gsp*orTSHR*when
compared with the control 3T3L1. Occasional cells containing
small lipid vacuoles were observed in all populations.
Furthermore, we did not see a consistent change in
transcript levels of PREF1,PPARgor GPDH in cells
expressing TSHR* or gsp* when compared with the non-
modified population (data not shown).
Since chronic stimulation of the cAMP pathway might
induce up-regulation of e.g. phosphodiesterases, we included
a phosphodiesterase inhibitor (IBMX) in the culture medium.
Cells were allowed to reach confluence in CM and 0.5mM
IBMX was added for various time periods, but there was still
no morphological evidence for adipogenesis and the
transcript levels for markers of differentiation were not
significantly changed (data not shown).
TSHR* and gsp* inhibit induced adipogenesis by affecting
PPARgisoforms
PPARgagonist induced differentiation of the non-modified
cells produced the expected change in morphology and
similar signs of adipogenesis were also present in the L629F
expressing cells, but at a reduced level compared with the
control; by contrast, the gsp* population were completely
devoid of differentiating cells, as illustrated in Fig. 2, by the
absence of oil red O stained cells.
We compared transcr ipts for markers of adipogenesis in
non-modified, L629F and gsp* expressing cells; statistical
analyses of the results reported as fold changes are shown in
Table 2.Figure 3 is a representative experiment illustrating
that PPARgagonist induced differentiation of non-modified
3T3L1 resulted in sustained and significant increases in
PPARgand GPDH (P!0.05 and 0.02 respectively).
The L629F and gsp* populations displayed an attenuated
increase in PPARg(although not significantly different from
non-modified) and in GPDH (both significantly less than
non-modified) being more severe in the gsp* expressing cells
in keeping with their morphological appearance. Further-
more, expression of PREF1, an EGF-like transmembrane
protein that inhibits adipogenesis (Smas & Sul 1993), is
significantly down-regulated (P!0.04), by the differentiation
Figure 1 Representative (one out of seven performed) western blot of
the three 3T3L1 populations in complete medium (basal conditions)
showing phosphorylated CREB (upper panel) and total CREB (lower
panel) both having an apparent molecular weight of 43 kDa.
Figure 2 Oil red O staining in 3T3L1 cells following nine
days in differentiation medium containing pioglitazone.
A, non-modified; B, L629F; and C, gsp* expressing populations.
Magnification !200.
L ZHANG and others .Gsasignalling, PPARg2 and adipogenesis210
Journal of Endocrinology (2009) 202, 207–215 www.endocrinology-journals.org
protocol, to 54G0.12% of the day 0 value, in the non-
modified cells. The behaviour of PREF1 in the other
populations diverges. PREF1 expression in L629F cells after
exposure to DM was reduced to 30G0.06% of day 0 levels
(P!0.005), and this differentiation induced PREF1 down-
regulation is significantly greater than in the non-modified
and gsp* populations (P!0.04 and 0.02 respectively). By
contrast, the reduction to 89G0.07% of pre-treatment values
in the gsp* population is not significantly different from day 0
PREF1 transcripts in CM in these cells.
Coulter counting of the cells in the first 3 days after
addition of DM (nZ3) revealed that the non-modified
3T3L1 had a 2.4G0.9-fold increase in cell number on day 3
compared with day 0, indicating mitotic clonal expansion
(MCE). Both L629F and gsp* populations displayed
significantly increased proliferation compared with the non-
modified, 4.9G0.5(P!0.005) and 3.3G0.7(P!0.01) fold
respectively, suggesting no influence of either mutation on
MCE. In the following seven days, there was no further
increase in either the non-modified or L629F cell number,
but the gsp* population continued to proliferate, in keeping
with the absence of differentiation.
The transcriptional activity of PPARgis reduced when it is
phosphorylated (Hu et al. 1996). We were unable to detect
any phosphorylated PPARgeither in total cell lysates or
nuclear extracts. However, as shown in Fig. 4, the expression
of PPARg1 is greatly reduced (at all time points) in L629F
and gsp* and PPARg2 is completely absent from the latter, in
contrast to the non-modified 3T3L1.
Furthermore, the expression of PPARg1 and PPARg2
proteins in non-modified cells is increased in the first 24 h
following induction. PPARg1 expression continues to
increase throughout adipogenesis, but PPARg2 expression
is at the limit of detection during the MCE stage, but then
resumes in the terminal stages of differentiation.
Reduced FOXO1 phosphorylation may explain the lack of
PPARg2
The transcription factor FOXO1 represses the promoters for
PPARg1 and PPARg2(Armoni et al. 2006); it is inactivated
by phosphorylation when it translocates from the nucleus
to the cytoplasm (Nakae et al. 2003). As can be seen in
Table 2 Fold changes in transcript levels of adipogenesis markers in
the three populations of 3T3L1 cells following exposure to
differentiation medium. The three populations were plated in
12-well plates; once confluent, the cells were cultured for 9 days in
differentiation medium containing pioglitazone. mRNA was
extracted on days 0 and 9 and QPCR measurement of the
adipogenesis markers was performed. Results (meanGS.E.M. of the
three experiments were all performed in at least duplicate) are the
fold changes in transcripts for each gene (relative to the ARP
housekeeper) comparing day 9 and day 0
Non-modified
cells L629F population gsp* population
Marker
PPARg6.5G2.52
.7G0.72
.2G0.5
GPDH 159G43 19.5G8* 2.2G0.43
†
PREF1 0.54G0.12 0.3G0.06
‡
0.89G0.07
§
*P!0.03 compared with non-modified;
†
P!0.02 compared with non-
modified;
‡
P!0.04 compared with non-modified;
§
P!0.02 compared
with L629F.
Figure 3 QPCR measurement of adipogenesis markers – A, PREF1;
B, PPARg; and C, GPDH on day 0 (cells in complete medium) and
day 9 (cells in differentiation medium containing pioglitazone) in
non-modified, L629F and gsp* 3T3L1 cells. Results are the
meansGS.E.M. of triplicates, expressed as absolute transcript copy
numbers (transcripts) of the gene of interest per 1000 copies (100 for
PREF-1) of acidic ribosomal phosphoprotein (ARP). Representative
experiment, one out of three performed.
Gsasignalling, PPARg2 and adipogenesis .L ZHANG and others 211
www.endocrinology-journals.org Journal of Endocrinology (2009) 202, 207–215
Fig. 5, in the non-modified cells total FOXO1 protein
expression increases in the first 24 h after addition of DM in
the nucleus, where it is highly phosphorylated. It is essentially
absent from the cytoplasm during the MCE phase. In the gsp*
population, total FOXO1 protein expression is higher in basal
conditions, its expression is increased in DM and we observe
its translocation from the cytoplasm to the nucleus between
days 1 and 3. The main difference is the considerably
decreased level of phosphorylation from day 3 onwards
compared with the non-modified 3T3L1. The TSHR* cells’
behaviour is midway between that of the other two
populations.
What accounts for the behaviour of FOXO1?
Phosphorylation-dependent inactivation of FOXO1 depends
chiefly on the PI3K and phospho-Akt pathway (Sakaue et al.
1998). We investigated and observed an early increase in the
proportion of phospho-Akt in all populations of 3T3L1 cells
after addition of DM (Fig. 6), indicating the need for an
alternative explanation for the reduced FOXO1 phosphoryl-
ation in gsp*.
Two novel modulators of adipogenesis have recently been
described. PROF is a mediator between AKT and FOXO1
whose expression is transiently up-regulated during adipo-
genesis (Fritzius & Moelling 2008). ID2 is a small molecule,
whose transcription also increases during adipogenesis and
interacts with as yet to be identified factors to stimulate
PPARgexpression (Park et al. 2008). We investigated their
expression and found a transient increase in ID2 TCN in all
three populations in the first 4–8 h following addition of
DM, although transcript levels of ID2 were lowest in gsp*
cells at all time points (data not shown). By contrast, as shown
in Fig. 7,PROF transcription was increased in the non-
modified and, to a lesser extent, in the L629F populations, but
remained at a constant low level in the gsp* cells. Thus,
despite the abundance of phospho-Akt in the gsp* population,
the paucity of PROF prevents inactivation of FOXO1 by
phosphorylation, PPARg2 is not expressed and adipogenesis
does not occur.
Discussion
Our initial aim was to investigate the effects of TSHR
activation on 3T3L1 preadipocytes, following our previous
studies demonstrating increased adipogenesis in ex vivo fat
samples from Graves’ patients, in whom thyroid stimulating
antibodies (as opposed to gain-of-function mutation) produce
chronic activation of the TSHR (Starkey et al. 2003). From
reports of a pro-adipogenic effect of agents capable of
elevating cAMP, we hypothesised that spontaneous adipogen-
esis would arise in preadipocytes expressing naturally
occurring (e.g. in toxic thyroid adenoma) gain-of-function
Figure 4 Western blot analysis of PPARgprotein expression on day 0 (cells in complete medium) and at various time points
following addition of differentiation medium containing pioglitazone in non-modified, L629F and gsp* 3T3L1 cells.
Representative experiment, one out of three performed.
Figure 5 Western blot analysis of FOXO1 protein expression on day 0 (cells in complete medium) and at various time points
following addition of differentiation medium containing pioglitazone in non-modified, L629F and gsp* 3T3L1 cells. A, Nuclear
extracts probed with anti-phospho-FOXO1 (upper panel) and anti-total FOXO1 (lower panel). B, cytosolic extracts (panels as in
A). Representative experiment, one out of three performed.
L ZHANG and others .Gsasignalling, PPARg2 and adipogenesis212
Journal of Endocrinology (2009) 202, 207–215 www.endocrinology-journals.org
TSHR mutations (which like thyroid stimulating antibodies,
activate adenylate cyclase/PKA), but this was not the case.
It should be noted that the increase in cAMP in preadipocytes
using retroviral vectors to express the mutant receptor is
considerably lower than that obtained following overexpres-
sion by transfection of COS cells. The model more closely
resembles the in vivo situation, in which modest elevation of
cAMP is sufficient to generate a thyroid toxic nodule, and
concurs with our previous findings in thyroid cells (Ludgate
et al. 1999,Fuhrer et al. 2003). The constitutively active
mutant TSHR responded further when stimulated with TSH,
in common with other such mutants, although its TSH
responsiveness was less than the WT TSHR (data not shown)
as reported by others http://innere.uniklinikum-leipzig.de/
tsh/frame.html.
We have then used a gain-of-function mutation in a second
component of the PKA pathway, Gsasubunit, to achieve
activation of CREB. Neither the receptor, nor its down-
stream G protein, induced spontaneous adipogenesis (even in
the presence of IBMX) as would be expected from previous
reports indicating that CREB activation is ‘necessary and
sufficient for adipogenesis in 3T3L1’ (Reusch et al. 2000).
Experiments were then conducted to investigate the effects
of TSHR* and gsp* on in vitro induced adipogenesis.
The reduced differentiation in 3T3L1 expressing TSHR*
mirror the results we obtained in human primary orbital
preadipocytes (Zhang et al. 2006). Our demonstration of an
inhibitory role for Gsain adipogenesis confirms and extends
the findings of Wa n g et al. (1992) who demonstrated
enhanced differentiation in 3T3L1 cells treated with
oligomers antisense to the protein. Subsequently, they
(Wang & Malbon 1999) reported that Gsarepression of
adipogenesis is mediated by the tyrosine kinase Syk, but how
this may impact adipogenesis is not clear and we are not aware
of further studies to address this question. Our investigations
of the pathway confirmed their findings, i.e. an early and
transient increase in phospho-SYK in the first day following
induced adipogenesis, but only in the non-modified
population. The L629F and gsp* expressing cells showed a
decrease in phospho-SYK levels (data not shown).
Our experiments illustrate that the failure to differentiate in
gsp* is due to reduced expression of PPARg1 and interference
in the alternate promoter usage required to generate
PPARg2, the isoform specific to adipose tissue (Tontonoz
et al. 1995).
There is some controversy concerning how PPARg2
expression changes following addition of DM. Some authors
report that it is induced, others that it is up-regulated. In our
differentiation protocol; PPARg2 protein expression
increases in the first 24 h, but is then repressed until day 5
post-induction. Its reappearance seems to coincide with the
start of the terminal differentiation phase and was a consistent
finding in all experiments, but contrasts with the results
obtained by others reporting that PPARg2 expression steadily
increases through all stages of differentiation (Saladin et al.
1999).
FOXO1 represses transcription from both the PPARg1
and PPARg2 promoters (Armoni et al. 2006) and so we
investigated its participation in our different 3T3L1 popu-
lations. We observed a rapid increase in FOXO1 protein
expression and its subsequent translocation to the nucleus
during the first 3 days post-induction in non-modified
3T3L1. At later time points, the nuclear FOXO1 sustained a
high degree of phosphorylation. The antibody we have used
detects phosphorylation of FOXO1 at serine 256, located in
the forkhead domain and indicative of inactivation of the
transcription factor; there are two additional phosphorylation
sites. Nakae et al. (2003) transduced cells with a tagged
Figure 6 (A) Western blot analysis of phospho-Akt (upper panel) and actin (lower panel) in nuclear extracts on day 0 (cells in
complete medium) and at various time points following addition of differentiation medium containing pioglitazone in
non-modified, L629F and gsp* 3T3L1 cells.
Figure 7 QPCR measurement of ProF expression on day 0 (cells in
complete medium) and at various time points following the addition
of differentiation medium containing pioglitazone in non-modified,
L629F and gsp* 3T3L1 cells. Results are the meansGS.D.of
triplicates, expressed as absolute transcript copy numbers
(transcripts) per 1000 copies of acidic ribosomal phosphoprotein
(ARP). Representative experiment one out of two performed.
Gsasignalling, PPARg2 and adipogenesis .L ZHANG and others 213
www.endocrinology-journals.org Journal of Endocrinology (2009) 202, 207–215
FOXO1 and were able to demonstrate a cytoplasmic location
for phosphorylated FOXO1. In our experiments investigating
endogenous FOXO1, it is predominantly cytoplasmic, then
nuclear and then present in both fractions, in agreement with
this work. Any phosphorylated endogenous FOXO1 in our
nuclear extracts is probably partially phosphorylated (on
serine 256, the preferred target for the kinase) and thus
inactive. Our experiment also revealed that phosphorylation
of FOXO1 is impeded in the gsp* population from day 3 post-
induction onwards and so provides some explanation for their
lack of adipogenesis.
At no time point is signalling via phospho-Akt compro-
mised and yet FOXO1 phosphorylation had remained low in
the gsp* cells. Investigation of two recently described
modulators of adipogenesis, ID2 and PROF suggested they
may have a role in explaining this anomaly. PROF, also known
as WDFY2, is a member of the WD-repeat propeller-FYVE
protein family, which is a binding partner for phospho-Akt
and facilitates FOXO1 phosphorylation (Fritzius & Moelling
2008). In silico analysis of its proximal promoter reveals the
presence of a potentially functional half-site cAMP response
element; however, it lacks a ‘tata’ box, generally thought to be
required for robust transcriptional induction by cAMP
(Zhang et al. 2005). Our investigations demonstrate that its
expression level, although higher in basal conditions,
remained fairly constant at all time points in the gsp* cells
following addition of DM, in contrast to the marked increase
in the non-modified 3T3L1.
We also considered the possibility that the excess Gsain the
gsp* expressing cells might sequester free Gbg subunits
generated during adipogenesis and thus impede their down-
stream effects. However, the mutation in arenders it
permanently in the GTP-bound form and thus less likely to
bind bg (Ford et al. 1998).
In the majority of our experiments, the behaviour of the
TSHR* population is mid-way between that of the gsp*
expressing and parent cell lines, indicating that the inhibitory
effects of Gsaare abrogated, presumably by the b/gsubunits,
which in the thyroid have been reported to signal by PI3K
(Zaballos et al.2008). PREF1 inhibits adipogenesis by
preventing induction of PPARg2(Kim et al. 2007). The
significantly enhanced down-regulation of PREF1 in the
TSHR* expressing cells may contr ibute to the rescue
mechanism. The drawing depicted in Fig. 8 summarises the
mechanism we propose to explain the difference in
adipogenic potential of the TSHR* and gsp* expressing
3T3L1 populations.
In conclusion, our results identify an important role for
Gsasignalling in inhibiting the alternative promoter usage
necessary to produce PPARg2transcripts.
Declaration of interest
There is no conflict of interest that could be perceived as prejudicing the
impartiality of the research reported.
Funding
Supported by The Wellcome Trust (grant number WT076003MA).
Acknowledgements
We would like to thank Prof. Maurice Scanlon for his continued support.
References
Al-Khafaji F, Wiltshire M, Fuhrer D, Mazziotti G, Lewis MD, Smith PJ
& Ludgate M 2005 Biological activity of activating TSHR mutants:
modulation by iodide. Journal of Molecular Endocrinology 34 209–220.
Armoni M, Harel C, Karni S, Chen H, Bar-Yoseph F, Ver MR,
Quon MJ & Karnieli E 2006 FOXO1 represses peroxisome
proliferator activator receptor gamma 1 and gamma 2 gene promoters
in primary adipocytes – a novel paradigm to increase insulin
sensitivity. JournalofBiologicalChemistry28 19881–19891.
Boone C, Gregoire F & Remacle C 1999 Var ious stimulators of the cyclic
AMP pathway fail to promote adipose conversion of porcine preadipocytes
in primary culture. Differentiation 64 255–262.
Drolet R, Richard C, Sniderman AD, Mailloux J, Fortier M, Huot C,
Rheaume C & Tchernov A 2008 Hypertrophy and hyperplasia of
abdominal adipose tissues in women. International Journal of Obesity
32 283–291.
Ford CE, Skiba NP, Bae H, Daaka Y, Reuveny E, Shekter LR, Rosal R,
Weng G, Yang C-S, Iyengar R et al. 1998 Molecular basis for interactions of
G protein bg subunits with effectors. Science 280 1271–1274.
Fritzius T & Moelling K 2008 Akt- and Foxo1-interacting WD-repeat-FYVE
protein promotes adipogenesis. EMBO Journal 27 1399–1410.
Fuhrer D, Lewis MD, Al-Khafaji F, Starkey K, Paschke R, Wynford-Thomas
D, Eggo M & Ludgate M 2003 Biological activity of activating TSHR
mutants depends on the cellular context. Endocrinology 144 4018–4030.
Haraguchi K, Shimura H, Lin L, Saito T, Endo T & Onaya T 1996 Functional
expression of thyrotropin receptor in differentiated 3T3-L1 cells: a possible
model cell line of extrathyroidal expression of thyrotropin receptor.
Biochemical and Biophysical Research Communications 223 193–198.
Hu E, Kim JB, Sarraf P & Spiegelman BM 1996 Inhibition of adipogenesis
through MAP kinase-mediated phosphorylation of PPAR. Science 274
2100–2103.
Ivan M, Ludgate M, Gire V, Bond J & Wynford-Thomas D 1997 An
amphotropic retroviral vector expressing a mutant Gsp oncogene: effects on
human thyroid cells in vitro.Journal of Clinical Endocrinology and Metabolism
82 2702–2709.
Figure 8 A drawing summarising the effects of signalling via Gsa
and Gbg subunits in regulating adipogenesis via expression of
PPARg2.
L ZHANG and others .Gsasignalling, PPARg2 and adipogenesis214
Journal of Endocrinology (2009) 202, 207–215 www.endocrinology-journals.org
Kim KA, Kim JH, Wang Y & Sul HS 2007 Pref-1 (preadipocyte factor 1)
activates the MEK/extracellular signal-regulated kinase pathway to inhibit
adipocyte differentiation. Molecular and Cellular Biology 27 2294–2308.
Ludgate M, Gire V, Crisp M, Ajjan R, Weetman A, Ivan M & Wynford-
Thomas D 1999 Contrasting effects of Gsp and activating mutations of the
thyrotropin receptor on the function and proliferation of thyroid follicular
cells. Oncogene 18 4798–4807.
Nakae J, Kitamura T, Kitamura Y, Biggs WH III, Arden KC & Accili D 2003
The forkhead transcription factor Foxo1 regulates adipocyte differentiation.
Developmental Cell 4119–129.
Park KW, Waki H, Villanueva CJ, Monticelli LA, Hong C, Kang S,
MacDougald OA, Goldrath AW & Tontonoz P 2008 Inhibitor of DNA
binding 2 is a small molecule-inducible modulator of peroxisome
proliferator-activated receptor-gexpression and adipocyte differentiation.
Molecular Endocrinology 22 2038–2048.
Paschke R & Ludgate M 1997 The thyrotropin receptor and thyroid disease.
New England Journal of Medicine 337 1675–1681.
Reusch JEB, Colton LA & Klemm DJ 2000 CREB activation induces
adipogenesis in 3T3-L1 cells. Molecular and Cellular Biology 20 1008–1020.
Sakaue H, Ogawa W, Matsumoto M, Kuroda S, Takata M, Sug imoto T,
Spiegelman BM & Kasuga M 1998 Posttranscriptional control of adipocyte
differentiation through activation of phosphoinositide 3-kinase. Journal of
Biological Chemistry 273 28945–28952.
Saladin R, Fajas L, Dana S, Halvorsen YD, Auwerx J & Briggs M 1999
Differential regulation of peroxisome proliferator activator receptor gamma
1 (PPAR gamma 1) and PPAR gamma 2 messenger RNA expression in the
early stages of adipogenesis. Cell Growth & Differentiation 10 43–48.
Smas CM & Sul HS 1993 Pref-1, a protein containing EGF-like repeats,
inhibits adipocyte differentiation. Cell 73 725–734.
Starkey KJ, Janezic A, Jones G, Jordan N, Baker G & Ludgate M 2003 Adipose
thyrotropin receptor expression is elevated in Graves’ and thyroid eye
diseases ex vivo and indicates adipogenesis in progress in vivo.Journal of
Molecular Endocrinology 30 369–380.
Tang QQ, Otto TC & Lane MD 2003 Mitotic clonal expansion:
a synchronous process required for adipogenesis. PNAS 100 44–49.
Tontonoz P, Hu E, Devine J, Beale EG & Spiegelman BM 1995 PPAR2
regulates adipose expression of the phosphoenolpyruvate carboxykinase
gene. Molecular and Cellular Biology 15 351–357.
Vanderpump MPJ, Tunbridge WMG, French JM, Appleton D, Bates D,
Clark F, Gr imley J Evans, Hasan M, Rodgers H, Tunbridge F et al. 1995
The incidence of thyroid disorders in the community: a twenty-year
follow-up of the Whickham Survey. Clinical Endocrinology 43 55–68.
Vassart G & Dumont JE 1992 The thyrotropin receptor and the regulation of
thyrocyte function and growth. Endocrine Reviews 13 596–611.
Wang HY & Malbon CC 1999 Gs alpha inhibition of adipogenesis is via SYK.
Journal of Biological Chemistry 274 32159–32166.
Wang HY, Watkins DC & Malbon CC 1992 Antisense oligodeoxynucleotides
to G
s
proteina-subunit sequence accelerate differentiation of fibroblasts to
adipocytes. Nature 358 334–337.
Zaballos MA, Garcia B & Santisteban P 2008 Gbg dimers released in response
to TSH activate phosphoinositide 3 kinase and regulate gene expression in
thyroid cells. Molecular Endocrinology 16 342–352.
Zhang X, Odom DT, Koo S-H, Conkright MD, Canettieri G, Best J,
Chen H, Jenner R, Herbolsheimer E, Jacobsen E et al. 2005
Genome-wide analysis of cAMP-response element binding protein
occupancy, phosphorylation, and target gene activation in human tissues.
PNAS 102 4459–4464.
Zhang L, Baker G, Janus D, Paddon C, Fuhrer D & Ludgate M 2006
Biological effects of thyrotropin receptor activation on h uman orbital
preadipocytes. Investigative Ophthalmology & Visual Science 47
5197–5203.
Received in final form 29 April 2009
Accepted 21 May 2009
Made available online as an Accepted Preprint
21 May 2009
Gsasignalling, PPARg2 and adipogenesis .L ZHANG and others 215
www.endocrinology-journals.org Journal of Endocrinology (2009) 202, 207–215