Growth Factors, Part A

Visit the link below to download the full version of this book:

https://medidownload.com/product/growth-factors-part-a/

Click Download Now

 LIST OF CONTRIBUTORS

Sarah Spiegel Department of Biochemistry and

Molecular Biology
Georgetown University Medical Center
Washington, D.C.
Douglas K. Tadaki Naval Medical Research Institute
Immune Cell Biology Program
Bethesda, Maryland

Xiao-Fan Wang Department of Pharmacology

Duke University Medical Center
Durham, North Carolina

Bengt Westermark Department of Pathology

University Hospital
Uppsala, Sweden

Lewis T. Williams Chiron Corporation

Emeryville, California
PREFACE

Advances in molecular technology in recent years have catalyzed an explosive

growth of information about intercellular peptide messengers and their receptors.
For example, ten years ago the only neurotrophin characterized at the molecular
level was nerve growth factor (NGF) and the only recognized neurotrophin receptor
was the p75 NGF receptor. At present, the number of described neurotrophic
peptides approaches 30 and the number of receptors is increasing apace. Just six
years ago, the characterized interleukins numbered about three while now there are
at least 16. Because many of these new peptide ligands and receptors were identified
by "reverse genetic" techniques the understanding of their biological roles lags
behind the knowledge of their molecular structures. Over the past few years,
however, a new era of functional studies has begun because recombinant proteins
have become available for clinical studies. In addition, animal models have been
and are being developed using recombinant DNA techniques. Both the clinical
studies and studies of transgenic and target deleted mice will allow for further
physiologic elucidation of the biological roles of these messenger peptides and their
receptors.
This series on Growth Factors and Cytokines is divided into three main sections:
Growth Factors (Volume I), Cytokines (Volume II) and Systems (Volume III).
Although volumes I and II are separate the distinction between "growth factors"
and "cytokines" is probably more historical or pragmatic than indicative of differ­
ences in function. The term "growth factors" refers to a wide variety of locally or
systemically produced proteins with pleiotropic actions on tissue growth and

XI
xii PREFACE

differentiation. The term "cytokines" describes a group of proteins identified

primarily within the immune and hematopoietic systems, although it is likely that
such a narrow view of cytokines will not survive for long. For example it appears
that some interleukins and interleukin receptors are expressed by neuroepithelial
cells in vivo suggesting that these interleukins may have intrinsic roles within the
nervous system. Furthermore, tumor necrosis factor (TNF) has been identified as
a potential adipose tissue regulatory factor which is both produced and acts locally.
The third volume entitled Systems deals more directly with the role of these factors
in both normal physiology and the disease processes resulting from the deficiency
or excess of growth factors/cytokines and their receptors.
The first volume deals with peptide growth factors and their receptors. Here too
there is an arbitrary division of ligands and their receptors. In some instances (e.g.,
insulin-like growth factors) the proteins and their corresponding receptors are
discussed in the same chapter, whereas in other cases, for example, NGF and
platelet-derived growth factor they are discussed separately. While we have at­
tempted to be as comprehensive and inclusive as possible, there will always be
some regrettable omissions. At the publishing date we recognize that a few growth
factors and cytokines have not been included in this review. These new discoveries
will for certain be reviewed in similar pages in the future.

Derek Le Roith
Carolyn Bondy
INSULIN-LIKE GROWTH FACTORS

Derek LeRoith and Carolyn Bondy

Abstract 1
I. Introduction 2
II. Molecular and Cellular Aspects 2
A. IGFs 3
B. Receptors 5
C. IGF Binding Proteins 9
III. Physiological and Clinical Aspects 11
A. Embryonic Growth and Development 11
B. Postnatal Growth 12
C. IGF-I and Intermediary MetaboHsm 13
D. Clinical Uses of IGF-I 15
E. IGF-II 17
F. IGFs and Neoplasia 18
IV. Conclusion 18
References 19

ABSTRACT

The insulin-like growth factor family of peptides, binding proteins and receptors is
involved in normal growth and development. Later they are important in the differen­
tiated function of a number of tissues. Aberrations in this growth factor system are
associated with different diseases, rangingfromshort stature and diabetes to malignancy.

Growth Factors and Cytokines in Health and Disease

Volume lA, pages 1—26.
Copyright © 1996 by JAI Press Inc.
All rights of reproduction in any form reserved.
ISBN: 0-7623-0091-4

1
 DEREK LEROITH and CAROLYN BONDY

With the advent of recombinant DNA technology, sufficient quantities of the ligands
(and binding proteins) have become available for clinical testing in the therapy of certain
diseases. These exciting new possibilities need to be assessed carefully for side-effects.

I. INTRODUCTION
The insulin-like growth factors (IGF-I and IGF-II) regulate growth and develop­
ment of multiple tissues during embryonic and fetal stages (reviewed in Daughaday
and Rotwein, 1989; Werner et al., 1994). During postnatal stages they continue to
affect growth and maintain the differentiated function in these numerous tissues
and in specific cell types. While the liver produces large amounts of both IGFs,
many extrahepatic tissues synthesize and secrete these factors as well (Lowe et al.,
1987; Hoyt et al., 1988). Circulating IGFs are of hepatic origin and act in a classical
endocrine mode, whereas extrahepatic IGFs act locally in a paracrine or autocrine
mode. The biological actions of the IGFs are mediated primarily by the type I IGF
receptor (IGF-I receptor) which is ubiquitously expressed (LeRoith et al., 1995).
The actions of the IGFs are also affected by a family of IGF-specific binding
proteins (IGFBPs) found in circulation and in extracellular fluids; these proteins
may enhance or inhibit the actions of the IGFs primarily by affecting their
availability to cell surface receptors (Baxter and Martin, 1989; Rechler, 1993; Jones
andClemmons, 1995).
In this review we will initially discuss the basic molecular and cellular aspects
of the IGFs, their binding proteins and receptors, and use examples from normal
physiology and pathology to highlight their importance. Then we summarize the
available data on the clinical studies of recombinant human IGF-I (rhIGF-I) and,
to a lesser extent, IGF-II which have recently become available for clinical research.

II. MOLECULAR AND CELLULAR ASPECTS

The IGFs are structurally similar demonstrating -65% amino acid similarity with
each other and -50% with insulin (Blundell et al., 1983; Daughaday and Rotwein,
1989; Sussenbach, 1989; Rechler and Nissley, 1990) (Figure 1). Circulating insulin
consists of an A- and B-chain, because the connecting(C) peptide is proteolytically
cleaved out during processing of the prohormone. Mature, circulating IGF-I and
IGF-II retain the smaller C-peptide and have a D-extension to the A-chain. The
E-peptide in the prohormone is cleaved off during processing (see below, Figure 2).
A. IGFs

The human IGF-I gene, on the long arm of chromosome 12 (Tricoli et al., 1984),
spans more than 90 kb of chromosomal DNA and contains at least six exons. Exons
Insulin-Like Growth Factors

B30

A21

INSULIN PROINSULIN

IGF (I) IGF (II)

Figure 1. Predicted tertiary structures of the insulin-iii<e growth factor (IGF) family of
peptides.

1 and 2 encode distinct, mutually exclusive 5'-untranslated regions (UTRs) as well

as distinct N-termini of the signal peptide (Figure 2) (Rotwein et al., 1986; Shimatsu
and Rotwein. 1987a). Exons 3 and 4 encode the mature peptide sequence, whereas
the E-peptide coding sequences are contained in exons 4, 5 and 6.
Transcriptional and posttranscriptional events are extremely complex. For exam­
ple, the exon 1 promoter lacks core promoter elements, such as TATA and CCAAT
boxes, and transcription of this exon is, therefore, initiated from at least four sites
dispersed over a ^350 bp region. Transcription from exon 2, which contains TATA-
and CCAAT-like motifs, is initiated over a smaller region (Jensen et al., 1991).
Exon 1-containing transcripts are expressed ubiquitously, and transcription is
regulated by multiple factors, generally, specific to each particular tissue. Exon 2-
 DEREK LEROITH and CAROLYN BONDY

•GF-1 rmM wym^^mv^^ in

Pi Pz P3 P4
'Gi^-^ J n n n nn I
1 2 3 4 5 6 7 8 9
Figure 2. Structure of mammalian IGF-I and IGF-II genes. Exons are numbered.
Known promoter sites in the IGF-II gene are labeled P1-P4.

containing transcripts, on the other hand, are particularly abundant in the liver and
are generally more responsive to growth hormone (GH). Thus, during development,
exon 2-containing transcripts appear after exon 1-containing transcripts in the liver,
but increase markedly at the onset of GH-dependent linear growth (Jensen et al.,
1991; Kim et al., 1991; Kikuchi et al, 1992).
At the posttranscriptional level there appears to be regulation of mRNA splicing,
in certain species of IGF-I mRNA (Shimatsu and Rotwein, 1987b), of a 186 bp
region of exon 1 that potentially influences translatability of that transcript. Fur­
thermore, two alternative E-peptides may be transcribed depending on exon 5 or
exon 6 usage; GH seems to favor exon 5 retention. At the level of mRNA stability,
the longer -'7.5 kb transcripts, derived from distal polyadenylation site usage in the
long exon 6, are more unstable than the shorter -'1 kb mRNAs derived by more
proximal polyadenylation site usage (Lowe et al., 1988; Lund et al., 1989; Heppler
et al., 1990; Steenbergh et al., 1991).

IGF-U

The IGF-II gene spans ^30 kb of chromosomal DNAon the distal end of the short
arm of human chromosome 11 (Tricoli et al., 1984), immediately 3' to the insulin
gene. Like IGF-I, the IGF-II gene is complex consisting of nine exons. The mature
peptide is encoded by exons 7,8 and 9. Transcription is controlled by four different
promoters (PI - P4) (Dull et al., 1984) (Figure 2). The promoters are activated in
a tissue- and development-specific manner; promoter PI is activated in adult liver,
whereas promoters P2, P3 and P4 are active in most fetal tissues and adult
nonhepatic tissues (dePagter-Holthuizen et al., 1987; dePagter-Holthuizen, 1988;
Holthuizen et al, 1990). PI is a TATA-less, GC-rich promoter with heterogeneous
transcription initiation. Liver-specific expression of PI is regulated by the
CAAT/enhancer binding protein (C/EBP) and the liver-enriched activator protein
Insulin-Like Growth Factors 5

(LAP) (Sussenbach et al, 1994). Promoters P3 and P4 contain a TATA box and P3
also has a CCAAT box. P3 and P4 exhibit transcription from specific sites and are
more highly regulated than PI. Human IGF-II promoter P3 is expressed in many
fetal and non-hepatic adult tissues and is regulated by the krox 20/egr2 and krox
24/egrl transcription factors (Sussenbach et al., 1994). Multiple IGF-II mRNA
transcripts are produced as a function of specific promoter usage and different
lengths of 3' UTRs resulting from use of multiple polyadenylation sites. Transcripts
from P2 and P4 have shorter 5'-UTRs and are preferentially translated (Irminger et
al., 1987; Rechler, 1991).
In human tissues, the IGF-II gene demonstrates promoter-specific genomic
imprinting (Vu and Hoffman, 1994). When promoter P1 is used, as with adult liver,
both maternal and paternal alleles are transcribed, but in the case of promoters
P2—P4 usage, only the paternal allele is expressed. Because promoters P2—P4 are
clustered in a '-5 kb DNA region, whereas PI is ~20 kb upstream (Figure 2), this
suggests that the imprinting signals lie between promoters PI and P2 (Vu and
Hoffman, 1994).
In murine species, the equivalent of promoter PI is absent and the remaining
three promoters are all imprinted except for the choroid plexus and leptomeninges
where biallelic expression occurs (Pedone et al., 1994; DeChiara et al., 1991a;
DeChiara et al., 1991b). Interestingly, in the mouse and rat, postnatal IGF-II
expression decreases dramatically leading to undetectable circulating IGF-II levels
(except in the CNS), whereas in humans, where the PI promoter is active in liver,
IGF-II expression persists throughout life. Because the PI promoter is biallelically
expressed, it has been postulated that this may allow more abundant synthesis of
the IGF-II peptide which is released into the circulation. Loss of genomic imprint­
ing has been implicated as a reason for the high level of expression of IGF-II in many
cancers (Steenman et al, 1994; Wemer and LeRoith, 1995), where the IGF-II peptide
may be important as an autocrine growth factor enhancing tumor growth (Zhan et al.,
1994).
B. Receptors

The biological effects of the IGFs are mediated by a family of specific membrane-
associated glycoprotein receptors including the insulin IGF-I and IGF-II receptors
and the insulin receptor-related-receptor (IRR) (Zhang and Roth, 1992; Reinhardt
et al., 1993; Kovacina and Roth, 1995; LeRoith et al., 1995). The insulin IRR, and
IGF-I receptors are closely related tyrosine kinase receptors, whereas the IGF-II
receptor is identical to the cation-independent mannose-6-phosphate (M-6-P) re­
ceptor (Figure 3). Because the biological effects of the IGFs on growth, develop­
ment, and differentiated functions are primarily via the IGF-I receptor, we will
concentrate on its structural and functional characteristics. Because the IRR has
limited tissue distribution and does not bind the IGFs, it does not apparently affect
IGF action.
 U 03 2 dJ O
Oj C tn
^^ <-
Ia
JD (U a; 4S re
CD U
— S" ^
I1^
a
Si o CD
uL ^V ^
U ?^
Q. I .
X o
UJ Q I o O (J
CD
2 E
^
CD
(O O Q. " " Q. § cr
? Q- I I I I I I I I I I I I I IX U C CD
u _c CD
UL S CD c a
CL .C
c|S
t/i 13 »-
CD • - ' -
o
-^ £ W)
Q. -O CD
O KYVyj
_ -c Q.
•E DC
3 2
0) C
JiYXYI "O
T.
CD
U
C 1}
— >» CQ.'g ^
I
' c h Q.
u_ = : 203 ybi Qj^
u 5j
-^ -^ S:
o B S ^ s
C CD Q.
2 y t CD
: Q.
^ -Q E
DC DC
2 ?^
•^ (n u
^5 CD c
§ "^ CD
- D ■;=; . E
-n 2 c
^ g OJ
13 5^ §"° -
i.E
o u. ^ ^ Q)
Q . *<D O DO <D
.KXXXl,
Js '^ *-
-5 -c a.
= "§
3 CD
03 CD
a V) ci; .9
1 K.AXXJ 1 to to
Q.
— CDO tn CD
CD
CE O ^ Q. U
1 NAAA^
E Q.
X
o CD
O u (i;
1 d
c
S..E <u
tn
to
O 2 o -D
CD __ u
Q.
(0 9) CO
III 8 2E U
E^ 'E
O GC O
<v y^ 'i—
u •£ E — 03 O CT3

to
CJ
"cu o -^
to
.2? tj
Insulin-Like Growth Factors 7

IGF-! Receptor

The human IGF-I receptor spans more than 100 kb of chromosomal DNA at the
distal end of the long arm of chromosome 15 and consists of 21 exons (Abbott et
al., 1992). The gene encodes contiguous a and p subunits which are cleaved during
processing and the mature functional receptor is a heterotetrameric glycoprotein in
an (aP)2 configuration where the subunits are joined by disulfide bridges. The a
subunits are entirely extracellular and bind the ligands, primarily, in the region of
the cysteine-rich domain. The P subunits are anchored in the membrane and contain
a cytoplasmic tyrosine kinase domain which has 84% homology with the equivalent
region in the insulin receptor and IRR (Ebina et al., 1985; Ullrich et al., 1985;
Ullrich etal, 1986).
The IGF-I receptor is widely expressed at high levels during embryogenesis,
suggesting an important role in tissue development (see below). Extensive studies
on characterizing the promoter region have revealed a number of features involved
in regulating IGF-I receptor gene expression (Werner et al., 1989; Werner et al.,
1990; Cooke et al., 1991; Mamula and Goldfme, 1992; Werner et al., 1992). Both
the unusually long (-^1 kb) 5' UTR and the proximal -^500 bp 5' flanking region are
very GC-rich with multiple Spl binding sites and no TATA or CCAAT boxes.
However, transcription initiation begins at a single site surrounded by an "initiator"
sequence. Spl binding to the promoter region may regulate transcription initiation
in the absence of TATA and CCAAT boxes (Smale and Baltimore, 1989).
Ligand binding to the IGF-I receptor initiates receptor autophosphorylation
which involves the phosphorylation of a cluster of three tyrosines (1131,1135 and
1136) in the kinase domain (Gronberg et al., 1993; Kato et al., 1993; Kato et al.,
1994). As with the insulin receptor, autophosphorylation activates the tyrosine
kinase activity of the receptor, leading to tyrosine phosphorylation of cellular
substrates. The substrate that has been best characterized to date is the insulin
substrate-1 (IRS-1) (Sun et al., 1991; Myers and White, 1993) (Figure 4). IRS-1
contains multiple tyrosine residues in a Tyr-Met-X-Met (YMXM) or related motifs.
Phosphorylation of tyrosines in these motifs mediate binding of other substrates,
such as the 85-kDa subunit of phosphatidyl-inositol 3' (PI3') kinase and Grb2
(Myers et al., 1992; Yamamoto et al., 1992; Baltenspergan et al., 1993; Giorgetti et
al., 1993; Myers et al., 1993; Skolnik et al., 1993). IGF-I stimulates PI3' kinase
activity and MAP kinase activity, the latter being the result of activation of the
ras/raf kinase pathway initiated by Grb2 binding to IRS-1. Other known pathways
involved in IGF-I receptor activation include protein tyrosine phosphatase-IB
(Kenner, 1993) and SHC and crk (Beitner-Johnson and LeRoith, 1995).

IGF'II/M-6'P Receptor

The IGF-II/M-6-P receptor is a bifunctional receptor with a large extracellular

region containing 15 contiguous repeats and a very short cytoplasmic tail (Lobel
et al., 1988; MacDonald et al., 1988; Morgan et al., 1987). Unlike the insulin and
 DEREK LEROITH and CAROLYN BONDY

MAP Kinases (ERKs)

Transcription Factors

Gene Expression

Figure 4. Schematic representation of intracellular signaling pathways of the IGF-I

receptor. Upon binding IGF-i, the IGF-I receptor undergoes autophosphorylation at
multiple tyrosine residues. The intrinsic kinase activity of the receptor also phospho-
rylates IRS-1 at multiple tyrosine residues. Various SH-domain-containing proteins,
including PI3-kinase, Grb2, Syp, Nek and crk associate with specific phosphotyros-
ine-containing motifs within IRS-1, as shown. Activation of IGF-I receptors also results
in tyrosine phosphorylation of She, which then complexes with Grb2. Grb2 is tightly
associated with the mammalian guanine nucleotide exchange factor Sos, which
activates Ras. IGF-I can apparently activate Ras via both the IRS-1-Grb2-Sos or the
Shc-Grb2-Sos pathways. This leads to the activation of a cascade of protein kineses
including Raf-1 and one or more related kineses, MAP kinase kineses (or MEKs), the
MAP kineses, and S6 kinase. These protein kineses, in turn, activate various other
elements, including nuclear transcription factors.

IGF-I receptors, the tail does not contain any tyrosine kinase activity. The major
function of this receptor is to target recently synthesized lysosomal enzymes from
the tmnS'Golgi network to lysosomes and to internalize lysosomal enzymes that
have escaped the cell. It also internalizes surface-bound IGF-II, targeting it to the
lysosomal compartment for degradation. Certain studies have suggested that, in a
limited number of cells, the IGF-II/M-6-P receptor may mediate Ca^"^ uptake, cell
growth, and generation of inositol phosphate (IP3) and diacylglycerol (Hari et al.,
1987; Matsmaga et al., 1988; Rogers and Hammerman, 1988). These effects were
mediated by a G-protein (Gi2) -related mechanism (Nishimoto et al., 1987; Nishi-
moto et al., 1989; Okamoto et al., 1990). However, a recent study has failed to
Insulin-Like Growth Factors 9

reproduce these results, and the general conclusion is that the IGF-n/M-6-P receptor
does not contain any signaling function (Komer et al., 1995).
The role of the IGFs and their receptors in embryonic and postnatal growth is
strongly supported by studies in which targeted disruption of various components
has been obtained in mice by homologous recombination (DeChiara et al., 1991b;
Baker et al., 1993; Liu et al., 1993; Powell-Braxton et al., 1993). Most mice
homozygous for a null mutation of the IGF-I gene die perinatally, though a small
number survive to adulthood. These mice demonstrate a 40% reduction in body
weight at birth and a 70% reduction at eight weeks of age. They are infertile and
have delayed ossification, underdeveloped muscles, and poorly organized lungs,
suggesting an important role for IGF-I in tissue differentiation and development.
On the other hand, targeted disruption of the IGF-II gene results in mice that weigh
60% of their normal littermate weight but survive, grow postnatally, albeit at a lower
weight, and are fertile. Thus, IGF-II may be important for determining overall body
size, but not essential for organogenesis. The IGF-I receptor plays a very important
role in organ development and probably mediates the effects of both IGF-I and
IGF-II. Mice lacking the IGF-I receptor gene are severely growth retarded by birth,
weighing only 45% of controls and die due to impaired muscular development
leading to respiratory failure. Many other organs are also severely retarded in their
development. Both the naturally occurring Tme mutation and the IGF-II/M-6-P
receptor-lacking mice show the same phenotype, that is, increased body size and
polydactylic and embryonic/neonatal lethality. Because this phenotype is rescued
by breeding mice lacking both the IGF-II/M-6-P receptor and IGF-II, it suggests
that the phenotype of the Tme mutation is due to excess IGF-II acting through the
IGF-I receptor, which is normally expressed in these mice. The lack of the
IGF-II/M-6-P receptor in the Tme mutants probably results in reduced clearance of
IGF-II resulting in overstimulation of the IGF-I receptor, again, strongly supporting
a role for the IGF-I receptor in signal transduction and for the IGF-II/M-6-P receptor
in IGF-II internalization and degradation (Lau et al., 1994; Wang et al., 1994).
C. IGF Binding Proteins

In addition to the IGFs and their receptors, the IGFBPs play a critical role in IGF
action. These are a family of six proteins, IGFBP-1 to 6, that specifically bind the
IGFs with high affinity and modulate their bioavailability and biological actions
(Table 1) (Baxter and Martin, 1989; Rechler, 1993; Jones and Clemmons, 1995).

Molecular

The IGFBPs are a family of closely related proteins, with N- and C-termini that
are highly conserved at the amino acid level, including 18 cysteine residues, 12 in
the NH2-terminal region and six in the COOH-terminal region. Most of the
cysteines are disulfide bonded, giving rise to the tertiary structure (Wood et al.,
1988; Rechler, 1993).
10 DEREK LEROITH and CAROLYN BONDY

Table 1, Insulin-Like Growth Factor Binding Proteins

Protein
Mass RGD Relative Binding
IGFBPs (kDa) Glycosylation Phosphorylation Sequence Proteolysis Affinity
IGFBP-1 25 - + + - IGF-I = IGF-II
IGFBP-2 31.3 - - + + IGF-II > IGF-I
IGFBPS 28.7 N-linked + - + IGF-I = IGF-II
IGFBP-4 25.9 N-linked - - + IGF-I = IGF-II
IGFBPS 28.5 0-linked + - + IGF-I = IGF-II
IGFBPS 22.8 0-linked 7 - ? IGF-II»IGF-I

The IGFBPs bind the IGFs with high affinity (-10 to ~11 M), often greater than
the affinity demonstrated by the IGF-I receptor. IGFBP-2 and 6 have greater affinity
for IGF-II than IGF-I. The IGFBPs are ubiquitously expressed and are found in
most body fluids including plasma, cerebrospinal fluid, amniotic fluid, lymph,
milk, etc. In some, IGFBP-3 and 4, N-glycosylation adds to the heterogeneity seen,
whereas IGFBP-6 is 0-linked glycosylated.
In plasma, the maj or portion of the IGFs circulate bound to a '^ 150-kDa molecular
weight ternary complex consisting of the ligand (IGF-I or IGF-II), IGFBP-3 and
an acid-labile subunit (ALS) (Baxter and Martin, 1989; Leong et al., 1992). Because
ALS usually circulates in excess, IGFBP-3 may play a more significant role in
determining the total concentration of circulating IGF. The IGFs are also found, to
a lesser degree, in a 40-50 molecular weight complex that includes IGFBP-2,4 or
6 and, to the least extent, IGFBP-1. Very little free IGF is measurable because of
the short half-life of the free hormone. While the 150-kDa molecular weight
complex controls the bioavailability of IGFs, the smaller molecular weight com­
plexes may be important in transporting IGFs out of circulation to the target tissue
(Baretal., 1990).

Modulation of IGF Action

Circulating IGFs are bound to IGFBPs which prolong their half-lives and deliver
them to their target tissues. At the cellular level, IGFBPs control IGF action either
by restricting access to cell surface receptors and, thereby, inhibiting growth and
related anabolic frmctions or by augmenting this interaction and thereby potentiat­
ing the cellular response (Elgin et al., 1987; Bar et al., 1989; Blum et al., 1989).
Most of the studies have demonstrated that, with the exception of IGFBP-4 which
seems to be uniformly inhibitory, all of the IGFBPs have the capacity to potentiate
and inhibit IGFs biological responses.
The affinity of released IGFBPs (i.e., IGFBP-3 and 5) for IGFs is significantly
higher than cell-surface associated forms of IGFBP. Therefore, if the majority of
the IGFBPs are in the interstitial fluids or culture medium, they will sequestrate
Insu I in-Like Growth Factors 11

IGFs and inhibit their interaction with the receptor. On the other hand, when
cell-surface bound, the IGFBPs have reduced affinity for IGFs, and the IGFs are
then able to interact with their receptors. How this effect results in potentiated
function is not clear. Some studies suggest that cell-surface-associated IGFBP-3
inhibits the rate of receptor internalization, thereby, prolonging its activity at the
cell surface. Another study using IGF analogues suggested that potentiation re­
quired binding of the IGF to both IGFBP-1 and the IGF-I receptor (Conover and
Powell, 1991; Camacho-Hubner et al., 1992).
The mechanism(s) of interaction of the IGFBPs with the cell surface membrane
is also not clear. Some IGFBPs contain Arg-Gly-Asp (RGD) sequences that allow
interaction with integrin receptors; others such as IGFBP-3, 4, and 5 are
glycosylated (Table 1) (Jones et al., 1993; Arai et al., 1994a; Arai et al, 1994b).
Whether these features are important for the functions described above is yet to be
clearly defined. IGFBP-1 and 3 contain multiple serine residues that are phospho-
rylated intracellularly. Phosphorylated IGFBP-1 inhibits IGF action on certain
tissues, whereas the dephosphorylated form has manyfold lower affinity for IGF-I
(Frost and Tseng, 1991).
Finally, multiple serine and metallothionein proteases have been isolated that
proteolytically cleave the IGFBPs and aid in the release of the IGFs; they, too, may
be significant in modulating IGF biological activity at the tissue level (Fowlkes and
Freemark, 1992; Nam et al., 1994)

III. PHYSIOLOGICAL AND CLINICAL ASPECTS

There is a very large literature on the effects of IGF-I and II in cell culture systems
demonstrating pleiotropic actions largely dependent upon the cell type, culture
conditions, and assay system employed. The IGF-I receptor, which is thought to
mediate most, if not all, of the biological effects of IGF-I and II, is very widely
expressed in vivo and in cell lines, and, thus, IGF actions have been demonstrated
in a very wide array of cell types. These in vitro studies have been comprehensively
reviewed (Lowe, 1991; Jones and Clemmons, 1995). This survey will focus on in
vivo or clinical studies, which have been proliferating recently with the advent of
targeted deletion of IGF and IGF receptor genes in mice and the availability of
recombinant IGF-I for humans.

A. Embryonic Growth and Development

IGF-I and II have important effects in murine embryonic development as dem­
onstrated by the dwarfism of mice bom with targeted deletion of either of these
genes or the IGF-I receptor (Baker et al., 1993; Liu et al., 1993). Deletion of the
active IGF-II allele results in proportionate dwarfism which is apparent from
midgestation (Baker et al., 1993), that is after placentation but not before, suggest­
ing that the dwarfism may result from limited placental development. These mice
grow and reproduce normally, although they maintain their diminutive size
12 DEREK LEROITH and CAROLYN BONDY

throughout life. IGF-I deletants are also bom small but their growth retardation
appears late in gestation and is not associated with reduction in placental size. In
contrast to the IGF-II null mice, most of the IGF-I null mice die in the first days or
weeks after birth (Liu et al., 1993; Powell-Braxton et al., 1994). The cause of death
in these mice is not clear. The few that survive demonstrate extremely retarded
postnatal growth, so that, by two months of age, they are approximately 30% the
size of wild-type littermates and both sexes are infertile. Combined IGF-I/IGF-II
deletants and IGF-I receptor null mice are bom very small and uniformly die within
hours of birth, apparently due to hypoplasia of respiratory muscles (Liu et al., 1993;
Powell—Braxton et al., 1994). IGF-I and II are present in human gestation (Han et
al., 1987; Hill, 1990) and, presumably, have important roles in fetal growth and
development, although as yet there are few fetal syndromes attributed to IGF-I or
-II deficiency. Dismption of the distal portion of chromosome 15q, which includes
the locus of the IGF-I receptor gene, is associated with severe fetal and postnatal
growth retardation (Roback et al., 1991), and a child with intrauterine growth
retardation was reported to have a 50% decrease in IGF-I binding to cultured
fibroblasts (Bierich et al., 1984). Overproduction of IGF-II during gestation has
been implicated in fetal overgrowth in the Beckwith—Wiedeman syndrome
(Weksberg et al, 1994).
B. Postnatal Growth

The most well-recognized clinical role of IGF-I is the promotion of statural

growth. Growth hormone (GH) secreted by the anterior pituitary acts on the liver
to stimulate hepatocyte synthesis and secretion of IGF-I into circulation. Circulat­
ing IGF-I levels normally reflect GH activity, peaking around the time of the
pubertal growth spurt and gradually declining during the aging process (Juul et al.,
1994). Deficient GH secretion or action results in abnormally low circulating IGF-I
levels and short stature, known as pituitary dwarfism. In addition to short stature,
affected individuals demonstrate poor muscular development, suggesting that
IGF-I promotes muscle and long bone growth. GH treatment results in normaliza­
tion of the growth process in most cases. However, some individuals are resistant
to GH due to genetic defects in the GH receptor or to antibodies to the hormone,
which may develop after GH treatment in patients with absent endogenous GH
(Laron, 1993). These GH-resistant patients grow in stature in response to treatment
with recombinant human IGF-I (Laron et al., 1992; Walker et al., 1992), supporting
the view that GH's growth-promoting effects are largely mediated by IGF-I (the
somatomedin hypothesis). Several more years will be necessary to determine if
fully normal growth pattems are produced in these children by IGF-I in the absence
of GH action.
IGF-I appears to enhance linear growth by acting directly on growth plate
chondrocytes during the time of their proliferation and hypertrophy. An increased
number of cell divisions and an increased dimension of individual chondrocytes
result in progressive linear expansion of the cartilage template upon which bone
growth depends. Elevation of IGF-I levels due to GH-producing tumors during
Insulin-Like Growth Factors 13

childhood results in gigantism, as epiphyseal chondrocytes undergo increased

cycles of division or increased hypertrophy, or perhaps both, augmenting long bone
growth. Elevation of IGF-I levels due to GH-producing tumors later in life causes
overgrowth of articular cartilage and soft tissues, but further growth in stature is
not seen due to the terminal differentiation of growth plate chondrocytes and fusion
of the epiphyses which occur in late puberty. These patients also have increased
bone mineral density, while GH and IGF-I-deficient adults often have reduced bone
density (Halse et al., 1981; Bouillon, 1991). These observations, together with the
fact that IGF-I promotes osteoblast activity in vitro (reviewed in Delaney et al.,
1994), suggest that IGF-I may have significant anabolic effects on bone minerali­
zation in adults. Clinically, it has been difficult to isolate the effects of IGF-I on
bone from those of other important effectors of bone mineral homeostasis, such as
sex steroids, thyroid, and glucocorticoid hormones, which also may be disturbed
in patients with pituitary disease. However, a recent study employing a primate
model of menopausal bone loss showed that GH treatment prevented osteoporosis
induced by suppression of ovarian flmction in rhesus monkeys (Mann et al, 1992).
Thus, although IGF-I's promotion of linear growth is its most clinically obvious
effect, this anabolic hormone may also have an important role in maintaining
skeletal mass in adults.
This review addresses primarily the endocrine role(s) and regulation of circulat­
ing IGFs. IGFs and IGFBPs are also synthesized locally for presumed autocrine or
paracrine functions in many different tissues. Significant interspecies differences
are apparent in local patterns of IGF system expression. The many studies directed
at elucidating specific local IGF system functions are beyond the scope of this
review.
C. IGF-I and Intermediary Metabolism

IGF-I has significant insulin-like metabolic effects, including the stimulation of

glucose and amino acid uptake, protein and glycogen synthesis (Guler et al., 1987;
Boulware et al., 1992). However, compared to insulin, IGF-I promotes substrate
uptake preferentially in lean tissue (Cascieri et al., 1986; Werner et al., 1989) and
promotes protein as opposed to fat synthesis (Bolinder et al., 1987; Sinha et al.,
1989). Although its anabolic effects are most notable during childhood and adoles­
cence when IGF-I secretion and growth are maximal, this hormone has continuing
effects in the adult, with a strong correlation found between decreasing IGF-I levels
and loss of musculoskeletal mass during aging (Juul et al., 1994).
There is complex interplay between GH, IGF-I, and insulin in regulating fuel
homeostasis and specific tissue anabolism (Figure 5). Portal insulin appears neces­
sary for GH to maximally stimulate IGF-I synthesis, because, when portal insulin
is reduced or absent, as in starvation, hyperalimentation, and insulin-dependent
diabetes mellitus (IDDM), hepatic IGF-I output in response to GH is reduced. This
may be due to a downregulation of hepatic GH receptor expression reflected by a
reduction in circulating GH binding protein levels (HoU et al., 1993). Insulin
priming of hepatic IGF-I synthesis is also deficient in patients with severe genetic