| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
INSERM, U-515, Hôpital Saint-Antoine (M.H., R.Z., P.L., B.D., C.B., L.P., Y.L.B.), 75571 Paris, France; and INSERM, U-380, Faculté de Médecine Cochin-Port Royal (G.H.), 75014 Paris, France
Address all correspondence and requests for reprints to: Dr. Martin Holzenberger, INSERM, U-515, Hôpital Saint Antoine, 75571 Paris Cedex 12, France. E-mail: holzenberger{at}st-antoine.inserm.fr
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
50%) IGF-IR deficiency (XS mice) or moderate deficiency
(<50%, M mice). The growth of XS mice was significantly retarded from
3 wk after birth onward, with respect to M littermates. This
effect was twice as strong in males as in females. Growth deficits
persisted throughout adult life, and at 10–12 months, most organs and
tissues showed specific weight defects. Skin, bone and connective
tissue, muscle, spleen, heart, lung, and brain were the most severely
affected organs in the XS males. With the exception of muscle and
spleen, the same tissues were also significantly reduced in size in
females, although to a lesser extent. The most severe growth defect,
however, concerned adipose tissue. Fat pad size in XS males was
only 29% (females, 44%) of M mice. The estimated number of adipocytes
in XS male fat pads was only 21% that of M males (XS
female, 27%). Lipid content per cell was significantly higher in XS
adipocytes, whereas plasma glucose and insulin levels were low in
XS males. Thus, IGF type I receptor deficiency produced mice with
disproportionate postnatal organ growth, and these effects depended
strongly on sex. A marked reduction in IGF-IR levels resulted in a
major defect in adipose tissue. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Studies of the classical IGF-IR knockout (KO) suggested that during embryonic development, lung, muscle and skin depend more strongly on IGFs than other organs (12, 13). However, as classical IGF-IR KO mice invariably die at birth, it has been difficult to extend such studies to postnatal growth and development. The classical KO of IGF-I produces a less severe developmental retardation, and depending on genetic background, the complete, constitutive loss of IGF-I function is not necessarily lethal (14, 15, 16, 17). This model is of particular interest because some of the IGF signaling is maintained during embryonic development through the action of IGF-II, such that these mice become gradually IGF deficient toward the end of embryonic development and the beginning of postnatal life, when endogenous IGF-II is down-regulated. Interestingly, in juvenile IGF-I-/- mice, the growth of lung tissue is affected more than that of liver and kidney (17). From these and other findings, it was deduced that musculo-skeletal compartments may be major targets for postnatal action of IGF-I. However, it is a complex task to assess the tissue-specific role of IGFs, because these growth factors act via both paracrine and endocrine mechanisms. It was shown that tissue-specific overproduction of IGF-I in transgenic animals increases local growth (18, 19, 20). In addition, liver-specific IGF-I gene KO in mice (6) showed that most, if not all, postnatal and pubertal IGF-I-mediated growth can be achieved through tissue-derived (nonhepatic) IGF-I as the predominant postnatal source of IGF-I. The continuation of this and other work, in particular the use of tissue-specific gain and loss of function models, may show in the future which of the many cell type-specific effects of IGFs identified in in vitro studies are relevant in vivo.
By gene targeting we previously produced a hypomorphic IGF-I receptor allele (a so-called knockdown, with partially reduced expression of IGF-IR) and showed that reduced availability of the IGF-I receptor slows down postnatal growth (21). The pubertal growth spurt in these mice is smaller than that in normal mice and resulted in a durable growth defect despite IGF-I up-regulation and IGF-binding protein-4 (IGFBP-4) down-regulation. This hypomorphic gene mutation affected total body weight more than the weight of individual organs (e.g. thoracic organs, brain, or bone). We concluded that other body compartments not evaluated in that study, such as skin or muscle, would probably show more potent effects in response to IGF-I receptor deficiency. To identify these compartments, we here produced a mouse model with strongly and ubiquitously impaired IGF type I receptor function, employing a gene dosage strategy (22, 23). We also generated a Cre transgenic mouse that produces a useful pattern of mosaic-early embryonic-ubiquitous gene deletion (MeuCre). By crossing this line with mice harboring the floxed hypomorphic IGF-IR allele, we were able to generate mice with very low IGF-I receptor levels. Under these conditions, sexually dimorphic patterns of pubertal growth retardation were observed, and tissues and organs with significant growth deficits were identified.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), indicating that it occurs
largely before organogenesis. Later, this recombination pattern is
transmitted to all developing tissues. We found that recombination
efficiency differs considerably between animals; the average
Cre recombination efficiency measured in tail DNA (n =
24) is 39.9 ± 21.5% (±SD; range,
6.3–85%). This interindividual variability increases the chances for
producing a large spectrum of degrees of gene expression deficiency in
the chosen experimental setup. MeuCre mice were mated with
IGF-IRneolox mice (21) to produce Cre-lox
double transgenics, which were then mated with homozygous IGF-IRneolox
mice to produce the F2 generation (Fig. 1, A and B) (26).
|
|
Details of the anatomical dissection
The skin was removed, with the exception of the tail tip and
feet. The brain was separated from spinal cord and removed. Limb
muscles were sectioned at their proximal insertions, limbs were removed
at their proximal joints, and feet were removed (bones accounted for
about 5% of limb weight; the rest was muscle). The genital fat pad,
also termed the inguinal or abdominal fat pad, and the major abdominal
and thoracic organs were prepared. Blood vessels were sectioned at the
organ hilus (kidney and liver). The large blood vessels of the heart
were cut at the atrium, and coagulated blood was removed from the
ventricles. The weight of the rest (difference between body
weight before death and the sum of the weights of the organs and
tissues) corresponded essentially to the gastrointestinal tract,
bladder, female reproductive organs, and uncollected blood. Feet and
any connective tissue removed during the dissection were added to the
carcass, which comprised essentially the skeleton and connective tissue
from the body and the head.
DNA preparation
DNA was prepared using standard procedures (27).
Tail biopsy samples (5 mm) from 10-d-old animals or tissue samples (50
mg; fat tissue, 200 mg) from adult animals were digested overnight with
proteinase K (Eurobio, Les Ulis, France) and centrifuged, and the
supernatant was mixed with an equal volume of isopropanol. The
resulting precipitate was washed with 75% ethanol, dried, and
resuspended in 10 mmol Tris, pH 8.0. For quantitative purposes,
deoxyribonuclease-free ribonuclease (Roche Molecular Biochemicals, Basel, Switzerland) was added, and
phenol/chloroform extractions were performed before precipitation. The
amount of DNA extracted from weighed, representative tissue fractions
was determined by UV spectrophotometry (Genesys 5, Spectronic
Instruments, Rochester, NY), and expressed in microgram per µl. Two
DNA extractions were performed per animal, and ODs were determined
twice per extraction.
To calculate cell size from DNA content we estimated the mouse genome
to 3.5 Gbp, so that 1 µg double stranded purified DNA corresponded to
2.6 x 105 extracted cell nuclei. The
specific weight (
) of tissues was estimated to be 1 mg/µl,
with the exception of that for adipose tissue, which was estimated to
be 0.8 mg/µl. The average cell volume (V) was calculated using V
(fl) = W (mg) x ( (mg/µl) x DNA (µg) x
2.6 x 105
(µg-1))-1, with W being
the weight of the extracted tissue fraction. To quantitatively evaluate
the individual fat cell compartment, the number of adipocytes was
estimated dividing the adipose tissue volume by the average adipocyte
volume.
Southern and dot blotting
DNA (8 µg) from each mouse was digested with
HincII, subjected to electrophoresis in a 1% agarose gel,
transferred to nylon membranes (Hybond+,
Amersham Pharmacia Biotech, Little Chalfont, UK) by
capillary action, and hybridized with a genomic probe (radiolabeled
using Rediprime, Amersham Pharmacia Biotech). The probe
recognized the 750-bp intronic region directly upstream from the
inserted neomycin resistance cassette. We detected a 2.4-kb wild-type
IGF-IR genomic fragment, a 4.5-kb targeted allele with floxed
neo-cassette and exon 3 (neolox (NL)), a 3.6-kb exon
3-excised (
ex3) allele, a 2.5-kb neo-cassette excised
(neo) allele, and a 1.7-kb total excision (KO) allele. The size of
the loxP insertions (40–50 bp) must be added to the
fragment sizes indicated in Fig. 1
. Membranes were placed against x-ray
film (Curix RP-2, Agfa-Gevaert, Mortsel, Belgium) for 1–4 d at -80 C
with amplifying screens. To quantify individual receptor deficiency
from genomic data, blots were exposed to phosphorimager screens and
analyzed using a STORM 850 PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA) and ImageQuant. For dot-blotting, DNA (10
µg) was denatured by heating, transferred to nylon membranes, and
probed with 1.1 kb Cre cDNA.
Receptor binding assays
Recombinant human IGF-I (rhIGF-I; GroPep Pty. Ltd.,
Adelaide, Australia) was labeled with 125I
(Amersham Pharmacia Biotech) using the chloramine-T method
(SA, 
100 µCi/µg). All chemicals were obtained from
Sigma (St. Louis, MO). Crude membranes were prepared from
whole brain as previously described (28). After
decapitation, brains were quickly removed and homogenized. The
homogenate was centrifuged twice at 4 C. The pellet was resuspended,
and the suspension was immediately used for binding assays. Proteins
were measured using the bicinchoninic acid protein assay from
Pierce Chemical Co. (Rockford, IL). Binding assays were
performed as previously described (21) in 0.5 ml 50
mM Tris-HCl buffer containing the membrane preparation (300
µg protein) and 15–20 pM
[125I]rhIGF-I. Nonspecific binding was
determined in the presence of 200 nM rhIGF-I.
IGF-I assay
Plasma samples (10–25 µl) were incubated in 0.01
M HCl for 30 min at room temperature and subjected to
ultrafiltration through Centricon 30 columns (Amicon, Millipore Corp., Bedford, MA). The ultrafiltrate was lyophilized,
resuspended in 0.1 M phosphate buffer and 1 mg/ml BSA, pH
7.4, and incubated for 3 d with
[125I]rhIGF-I (3000 cpm/tube) and a polyclonal
antihuman IGF-I antibody (final dilution, 1:120,000; a gift from J.
Closset, CHU, Liège, Belgium), that cross-reacts with
murine IGF-I (29, 30). Samples were tested at four
concentrations, each in duplicate (30). After incubation,
free and bound IGFs were separated using albumin-coated charcoal as
previously described (31). The detection threshold of the
assay was 2 ng/ml plasma. There was 5% intraassay variation and 10%
interassay variation.
Western ligand blotting of IGFBPs
Plasma samples (3 µl/animal) were subjected to 12.5%
nonreducing PAGE (32). Proteins were electrotransferred
onto nitrocellulose membranes and incubated with
[125I]IGF-I and
[125I]IGF-II (500,000 cpm each). Membranes were
washed and placed against x-ray film (Agfa-Gevaert) at -80 C. Pooled
normal samples of human and mouse origin were included on each gel as
standards. Western ligand blots were quantified using phosphorimager
technology and ImageQuant software (Molecular Dynamics, Inc.). Individual results were compared with standard mouse
samples on the same blot and normalized. Numerical results on
circulating levels of IGFBP-1 to -4 were statistically evaluated using
the Mann-Whitney test.
Histology and blood biochemistry
Genital fat pads from males and females were quickly frozen over
liquid nitrogen and embedded in Cryo-M-Bed (Bright, Huntingdon, UK).
Cryosections (-30 C, 15 µm) were prepared on SuperFrost Plus slides
(Menzel-Gläser, Braunschweig, Germany), fixed in 4%
paraformaldehyde for 20 min at 4 C, and stained with hematoxylin for 10
min. Blood biochemistry analysis was performed on a
Hitachi 911 using kits from Roche
(Indianapolis, IN; creatinine, cholesterol, triglycerides, and urea)
and Randox (Crumlin, UK; glucose, proteins) and protocols adapted to
the small volumes of plasma available from mice. Plasma insulin was
determined using the INSIK5 kit (DiaSorin, Inc.,
Stillwater, MN).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). In the early embryo, this line
constitutively produced mosaic Cre-mediated recombination
patterns very useful for a gene dosage approach (details in
Materials and Methods). We crossed this
MeuCre line with mice carrying the hypomorphic IGF-IRneolox
allele (Fig. 1A) (21) following the strategy shown in Fig. 1B. As MeuCre also generated mosaicism among germline cells
of the F1 double transgenics (with the same
efficiency as in other tissues), F2 mice
inherited various combinations of IGF-IRneolox, wild-type (WT), and
complete KO alleles with or without the MeuCre transgene.
This produced a wide spectrum of mutant genotypes in the
F2 generation (Fig. 1C). To make it possible to
deduce the degree of receptor deficiency directly from these genotypes,
we measured IGF type I receptor levels in groups of mice with various
allelic combinations by receptor-ligand binding assay (Fig. 2). The IGF-IRneolox allele produced 40%
less receptor than the WT. No receptors were produced from the complete
KO allele, and heterozygous KO mice had about half the number of
receptors present in normal mice. Various combinations of alleles
confirmed that these mutations were essentially not compensated by the
sister allele, and that their effects were therefore additive. Thus,
the combination of the IGF-IRneolox and IGF-IR KO alleles resulted in
70% receptor deficiency. If such mice had also inherited a
MeuCre-transgene, then more than 80% IGF-IR deficiency was
possible through additional somatic Cre recombination. As
MeuCre is expressed during the early embryonic stages, these
acquired receptor deficiencies were considered constitutive for the
individual. The degree of IGF receptor deficiency was then determined
for all 126 F2 mice directly from their
individual allele distributions using phosphorimager data from genomic
Southern blots and an appropriate algorithm. As expected, the combined
effects of hypomorphic IGF-IRneolox alleles, complete KO alleles, and
somatic Cre recombination produced an
F2 generation with a wide range of receptor
deficiency (Fig. 3A). Two large groups of
mice, one that displayed moderate levels of deficiency (M) and one with
strong deficiency (XS), could be distinguished. Most of the M mice had
10–30% receptor deficiency; most of the XS mice had 70–80%
deficiency. XS mice had, on the average, three times fewer receptors
per cell than M mice. In this study XS mice were compared with M mice;
the latter served as the control group. Although some of the XS animals
had more than 80% receptor deficiency, none showed signs of the
classical IGF receptor KO phenotype (respiratory insufficiency and
neonatal death). Two of the 126 F2 mice died
before adulthood, but there was no indication that this was related to
their individual degree of receptor deficiency (21% and 45%,
respectively).
|
|
). Relative weight differences were stable
throughout growth and into adulthood. At 10–12 months, XS males were
78% the weight of M males, and XS females were 86% the weight of
their M littermates. The effect of IGF-IR gene dosage on male mice was
thus 1.8 times stronger than that on females. The mean plasma IGF-I
concentration in the M group (491 ± 28 ng/ml) was normal and
similar to that in the WT (490 ± 34 ng/ml); XS mice tended to
have lower plasma IGF-I levels (-8%). Plasma levels of IGFBP-1, -2,
-3, and -4 were lower than those of M mice; this difference was
statistically significant for IGFBP-2 (-21%) and IGFBP-4 (-22%;
Table 2). One has to bear in mind,
however, that Western ligand blotting is a semiquantitative method.
|
Tissue-specific growth deficits
We dissected 92 of the M and XS mice and found that organs and
tissues were affected differently by decreased IGF-IR levels (Table 3). We studied all major body
compartments and found that bone together with connective tissue (the
carcass) and skin were among the most strongly affected tissues. Lung,
heart, spleen, and brain were also much reduced in size. The muscle
compartment was affected in males, but was less affected in females.
Thymus, kidney, and liver were less affected in both sexes (Fig. 4). For all of these organs and tissues,
the weight reduction in XS mice was concordant in both sexes, but more
pronounced in males than in females. The correlation between male and
female tissue-specific weight deficits was r2 =
0.30 (P < 0.05) without fat tissue and
r2 = 0.91 (P < 0.0001) with fat
tissue. We are sure that this result was not due to potential
tissue-specific effects of the MeuCre transgene, because we
obtained very similar results in F2 mice when all
35 MeuCre-positive individuals were excluded from the 92
dissected animals (results not shown). When we analyzed organ weight
relative to body weight, liver and kidney weights increased
significantly relative to body weight, whereas the relative weight of
fat tissue showed a highly significant decrease in XS mice.
|
|
) showed that in
M mosaics, many 50% receptor-deficient cells were present in adult
tissues together with cells with 20% deficiency. In the adult XS
mosaics, however, total KO cells were scarce, suggesting that these
cells contributed less to the growth of the tissues and/or that they
were less efficiently maintained during growth or afterwards, in the
adult tissues. We think that the relative underrepresentation of XS
mice with 80–90% reduction in receptor binding in Fig. 3A could also
be explained by this in vivo cell selection effect.
|
Lack of white adipose tissue in IGF-IR insufficiency
Little fat tissue was present in XS mice. The genital fat pad
weighed less than 30% in XS males and less than 45% in XS females
compared with that in M mice. The data for individual mice (Fig. 6A) showed that the mass of the fat pad
was distributed very differently in the M and XS populations. The fat
pads of M mice showed a large range of sizes (0.1–6 g), with most
between 0.1–2 g; the XS mice generally had fat pads weighing less than
1 g, and large fad pads were absent. To estimate cell volume and
density in genital fat pads, we extracted adipocyte DNA from M and XS
mice. In both sexes, there was significantly less DNA per unit wet
weight of adipose tissue from XS mice than from M mice (Table 4), indicating that cell volume was
larger in XS than in M fat pads. Similar results were obtained even if
all individuals with more than 1.3 g adipose tissue were removed
from the analysis, so that fad pads of approximately the same size but
with different degrees of receptor deficiency were compared.
|
|
). Although the volume of
fat cells was significantly greater in XS mice, these mice appeared to
be unable to build up larger fat depots. We estimated the number of
adipocytes in fat pads of XS males to be 56 x
109 (± 13 x 109)
cells, which corresponds to only 21.3% of the mean value for M mice
(27.1% for XS females). These figures are probably still an
overestimate, because DNA extraction did not differentiate between
adipocytes and other cell types present in fat tissue (e.g.
blood vessels or connective tissue). Histology confirmed that XS fat
tissue was composed of large adipocytes (Fig. 7, A–F) and contained very few cell
nuclei and large vacuole-like spaces. M adipocytes varied in size,
whereas XS fat tissue seemed to be mostly composed of uniformly large
adipocytes. We compared tail and fat tissue genotypes in
MeuCre-positive F2 mice. Mosaic
patterns induced by MeuCre in fat tissue did not differ
significantly from that in the tail (an example is shown in Fig. 7G),
suggesting that fat tissue-specific MeuCre activity was not
different from MeuCre activity in other tissues. Fat tissue
also showed the same relative lack of complete KO cells as did other
tissues.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Adult XS mice with about one quarter the number of receptors present in
normal mice were viable and fertile. However, consistent with previous
findings (21), this level of receptor deficiency
significantly impaired the postnatal and, in particular, pubertal
growth of male and female mice. We conclude that two functional
receptor alleles are necessary for normal postnatal growth. The growth
retardation did accumulate during prepuberty and puberty. No
significant catch-up growth was observed in XS mice after puberty, as
it has been reported for the double IGF-IR/M6P-R(IGF-IIR)-KO mutant
(34). Male XS mice in this study reached 70% of WT
weight, and their growth retardation was therefore less severe than
that of classical IGF-I KO mice (30% of adult WT weight)
(12). This suggests that in XS mice, the remaining
20–30% receptors mediate more than half of the IGF-dependent
postnatal weight gain.
The sexual dimorphism of pubertal growth depends on the
IGF-IR
In XS mice with low functional IGF-IR levels, much of the normal
sex-related postnatal growth dimorphism had disappeared. Androgens and
estrogens are the most likely candidates for regulating sexually
dimorphic growth in puberty and postpuberty. However, direct evidence
implicating these hormones in body growth is scarce. GH secretion
patterns, on the other hand, are sexually dimorphic, showing solid
regular peaks and low intermediate levels in males and more frequent
peaks with higher intermediate levels in females (35).
However, it is not known in detail how these dimorphic secretion
patterns generate divergent growth in males and females. GH-dependent,
hepatic IGF-I production does not seem to be involved, because in a
number of mammalian species, circulating IGF-I levels are clearly not
correlated with sex-linked differences in somatic development
(36). As the IGF receptor seems to mediate at least part
of the additional growth of males (Ref. 21 and results
reported here), we believe that the causes of sexually dimorphic growth
differences are probably situated at the level of the target cell.
Local IGF-I production, which is known to have determining effects on
body growth (6), may be regulated by androgens and/or
estrogens, or androgens may cooperate with IGF signaling in target
tissues. This would result in higher levels of IGF signaling, regulated
on a paracrine/autocrine basis, with no change in the total circulating
IGF concentration.
In the IGF-IR knockdown model, the growth of males was more retarded than that of females, whereas IGF-I was up-regulated, and IGFBP-4 was down-regulated. As greater receptor deficiency was observed in the present model than in the knockdown model, we expected the IGF-I concentration to be high. Circulating IGF ligand levels were, however, similar to or even lower than normal levels. As the IGF-I concentration is high in fat tissue (37, 38), it is possible that the significant lack of adipocytes in XS mice may have counteracted, to some degree, any underlying physiological up-regulation of IGF-I. In addition, IGFBP levels tended to be lower in XS mice, which also might lower the circulating IGF-I concentration in the adult. Plasma concentrations of IGFBP-2, IGFBP-3, and IGFBP-4 were significantly correlated with IGF-I levels. Finally, the down-regulation of IGFBP-4, here accompanied by a significant decrease in IGFBP-2, seems to be a consistent trait in IGF type I receptor deficiency. As IGF-I and IGFBP-3 levels are known to be normal in some human growth retardations, it may be interesting to study IGFBP-4 in these conditions.
Studies using transgenic mice suggested that IGFBP-4 is a functional antagonist of IGF-I in vivo (39). The low IGFBP-4 (and IGFBP-2) levels observed in our study could thus be a regulatory response by which IGF-IR-deficient XS males try to increase growth. Transgenic overexpression of IGF-I, although increasing IGFBP-5 production, did not alter IGFBP-4 levels (40). This indicated that regulation of IGFBP-4 may not be secondary to changes in circulating IGF-I levels (in the present study IGF-I levels of XS males were not significantly altered), but could be more directly related to reduced levels of IGF-IR (21). Clearly, further experiments are needed to clarify some of these important points concerning IGF system regulation.
Organ-specific effects of ubiquitous IGF-IR deficiency
We show here that receptor deficiency created a sexdimorphic
growth deficit that affected various tissues to different degrees. Body
compartments, such as bone and cartilage, muscle, and skin, which are
often not directly evaluated, showed the strongest effects. In
contrast, other organs, such as the heart, liver, and kidney, showed
relatively small effects. In XS males, there was only about one quarter
the normal amount of fat tissue, indicating a high sensitivity to IGF
receptor deficiency.
We found a significant correlation between male and female tissue-specific growth deficits. In both sexes, tissues showed either a strong growth deficit (adipose tissue, skin, heart, lung, bone, cartilage and connective tissue, spleen, and brain) or a weak deficit (kidney, thymus, and liver). Some tissues or organs, however, showed sex-specific growth deficits (bone, cartilage and connective tissues, spleen, muscle, and liver). These results were consistent with findings for the classical IGF receptor KO, in which the principal neonatal defects concerned skin, brain, skeleton, lung, and muscle (12). Our observations are also consistent with results from other studies (reviewed in Ref. 2), in that the degree of organ and/or tissue growth retardation in IGF-IR-deficient mice corresponds well to the magnitude of overgrowth in IGF-I overproducing mice, indicating the tissues whose growth is especially dependent on IGF signaling, such as heart, brain, skin, muscle, cartilage, and bone. A particular advantage of the present model over tissue-specific ligand overproduction, however, is that the genetically fixed receptor deficiency allows comparison of the IGF sensitivities of multiple tissues in a single experimental animal.
Adipose tissue
We observed a strong growth deficit in fat tissue. Interestingly,
fat tissue produces large amounts of IGF-I (even when compared with
synthesis in the liver). In addition, this IGF-I production is GH
dependent, and preadipocytes are sensitive to IGF-I signaling
(37). Moreover, preadipocytes produce significant amounts
of IGF-IR, whereas mature adipocytes do not (41). Our
findings are compatible with the idea that the IGF-I produced by fat
tissue may have major paracrine and autocrine functions during the
ontogenesis of fat tissue. Cell culture studies on the proliferation of
adipocyte precursor cells and their differentiation into mature
adipocytes have identified IGFs as important regulators of adipocyte
development (42, 43). Experiments using primary cultures
of adipocyte precursors have shown that IGFs increase the proliferation
of these precursors and inhibit terminal adipocyte differentiation
(44). A similar mechanism could explain the reduction in
size of the adipocyte compartment in XS mice, through limitation of
adipocyte precursor proliferation in vivo and stimulation of
differentiation into mature adipocytes.
Our findings are consistent with recent in vivo studies by Rajkumar et al. (45), who investigated adipogenesis in IGFBP-1-overproducing transgenic mice. No significant difference in fat pad development was observed between IGFBP-1-overproducing mice and WT mice fed a normal diet. However, if the mice were fed a diet rich in glucose, IGFBP-1-overproducing mice were less able than wild-type mice to increase their fat tissue. In contrast to our findings, however, adipocytes in IGFBP-1-overproducing mice were smaller than cells in wild-type mice. These researchers found evidence that IGFBP-1-overproducing fat tissue is less able to generate adipocyte colonies in primary culture. This is consistent with there being a defect in preadipocyte proliferation and adipocyte differentiation due to impaired IGF signaling. Complete insulin receptor deficiency has also been found to affect the formation of adipose tissue in mice (46), suggesting that insulin and IGFs may finally both be involved in the control of fat tissue development.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Fellow of Novo Nordisk (France). 
Abbreviations: IGFBP, IGF-binding protein; IGF-IR, IGF type I
receptor; KO, knockout; M, moderate IGF type I receptor deficiency;
MeuCre, mosaic-early
embryonic-ubiquitous Cre transgene; NL,
neolox; rhIGF, recombinant human IGF-I; WT, wild-type; XS, strong
(50%) IGF type I receptor deficiency.
Received May 10, 2001.
Accepted for publication July 3, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-actin-IGFBP-4 fusion gene induces smooth muscle hypoplasia.
Endocrinology 139:2605–2614This article has been cited by other articles:
 |
|
 |
|
  B. W. Sutherland, S. E. Knoblaugh, P. J. Kaplan-Lefko, F. Wang, M. Holzenberger, and N. M. Greenberg Conditional Deletion of Insulin-like Growth Factor-I Receptor in Prostate Epithelium Cancer Res., May 1, 2008; 68(9): 3495 - 3504. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Duffield, T. Vuocolo, R. Tellam, B. S. Yuen, B. S. Muhlhausler, and I. C. McMillen Placental restriction of fetal growth decreases IGF1 and leptin mRNA expression in the perirenal adipose tissue of late gestation fetal sheep Am J Physiol Regulatory Integrative Comp Physiol, May 1, 2008; 294(5): R1413 - R1419. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. B. Hurov, M. Huang, L. S. White, J. Lennerz, C. S. Choi, Y.-R. Cho, H.-J. Kim, J. L. Prior, D. Piwnica-Worms, L. C. Cantley, et al. Loss of the Par-1b/MARK2 polarity kinase leads to increased metabolic rate, decreased adiposity, and insulin hypersensitivity in vivo PNAS, March 27, 2007; 104(13): 5680 - 5685. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J Svensson, M Diez, J Engel, C Wass, A Tivesten, J-O Jansson, O Isaksson, T Archer, T Hokfelt, and C Ohlsson Endocrine, liver-derived IGF-I is of importance for spatial learning and memory in old mice. J. Endocrinol., June 1, 2006; 189(3): 617 - 627. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Raile, J. Klammt, A. Schneider, A. Keller, S. Laue, R. Smith, R. Pfaffle, J. Kratzsch, E. Keller, and W. Kiess Clinical and Functional Characteristics of the Human Arg59Ter Insulin-Like Growth Factor I Receptor (IGF1R) Mutation: Implications for a Gene Dosage Effect of the Human IGF1R J. Clin. Endocrinol. Metab., June 1, 2006; 91(6): 2264 - 2271. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sadagurski, S. Yakar, G. Weingarten, M. Holzenberger, C. J. Rhodes, D. Breitkreutz, D. LeRoith, and E. Wertheimer Insulin-like growth factor 1 receptor signaling regulates skin development and inhibits skin keratinocyte differentiation. Mol. Cell. Biol., April 1, 2006; 26(7): 2675 - 2687. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Svensson, B. Soderpalm, K. Sjogren, J. Engel, and C. Ohlsson Liver-derived IGF-I regulates exploratory activity in old mice Am J Physiol Endocrinol Metab, September 1, 2005; 289(3): E466 - E473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X.-J. Du Gender modulates cardiac phenotype development in genetically modified mice Cardiovasc Res, August 15, 2004; 63(3): 510 - 519. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Unger, S. D. Chernausek, W. Kiess, and R. J. Smith IGF-I Receptor Mutations and Intrauterine and Postnatal Growth Retardation N. Engl. J. Med., March 25, 2004; 350(13): 1362 - 1363. [Full Text] |
, loxP). A
neomycin resistance cassette (neo), equipped with a third
loxP site was cointroduced into intron 2. This created
the hypomorphic IGF-IRneolox allele. B, Breeding scheme for gene
dosage: MeuCre transgenic mice mated with IGF-IRneolox
mice produced double transgenic MeuCre/IGF-IRneolox
(F1). These were mated with homozygous IGF-IRneolox,
producing the F2 mice. These F2 mice comprised
six different genotypes that produced a large range of receptor
deficiency (indicated in percentage of WT). Using MeuCre and
IGF-IRneolox on C57BL/6 genetic background in F0 matings,
and homozygous IGF-IRneolox mice on 129/Sv background in F1
matings, we obtained C57BL/6–129/Sv first generation hybrids in
F2. C, Representative F2 genomic Southern blots
for each genotype. In a HincII tail DNA digest, the WT
allele and the four possible forms of the targeted allele were detected
as discrete bands (see Materials and Methods for
details). MeuCre-mediated excision occurs at early
embryonic stages in all tissues, including germline cells. Thus, some
F2 mice received from their
MeuCre/IGF-IRneolox parent (B6) an IGF-IR KO allele. The
other allele was always inherited as IGF-IRneolox (NL) allele.
Different alleles produced different amounts of receptor compared to
WT: IGF-IRneolox, 60% (
, 
, Males; •, 
, females, M mice are
shown in black; XS mice are shown in
white. Shaded areas enclose animals of
the same sex and group (in terms of receptor deficiency). Note that
there is no overlap between XS and M males and little overlap between
XS and M females.