This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Khandwala, H. M. Articles by Friend, K. E. Search for Related Content PubMed PubMed Citation Articles by Khandwala, H. M. Articles by Friend, K. E. Pubmed/NCBI databases Substance via MeSH Medline Plus Health Information Cancer

The Effects of Insulin-Like Growth Factors on Tumorigenesis and Neoplastic Growth

Section of Endocrine Neoplasia & Hormonal Disorders (H.M.K., K.E.F.) and the Department of Neurosurgery (I.E.M.), The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030; and Medical Department M (Diabetes and Endocrinology) (A.F.), Aarhus Kommunehospital, Aarhus, DK-8000 Denmark

	Abstract

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
Several decades of basic and clinical research have demonstrated that there is an association between the insulin-like growth factors (IGFs) and neoplasia. We begin with a brief discussion of the function and regulation of expression of the IGFs, their receptors and the IGF-binding proteins (IGFBPs). A number of investigational interventional strategies targeting the GH or IGFs are then reviewed. Finally, we have assembled the available scientific knowledge about this relationship for each of the major tumor types. The tumors have been grouped together by organ system and for each of the major tumors, various key elements of the relationship between IGFs and tumor growth are discussed. Specifically these include the presence or absence of autocrine IGF-I and IGF-II production; presence or absence of IGF-I and IGF-II receptor expression; the expression and functions of the IGFBPs; in vitro and in vivo experiments involving therapeutic interventions; and available results from clinical trials evaluating the effect of GH/IGF axis down-regulation in various malignancies.

I. Introduction

II. Overview: IGF Physiology and Gene Regulation

A. IGF-I gene expression

B. IGF-II gene expression

C. IGF-I receptor gene expression

D. IGF-II receptor gene expression

III. Potential Therapeutic Agents

A. GHRH antagonists

B. Somatostatin analogs

C. GH receptor antagonists

D. IGF-I receptor antibodies and analogs of IGF-I

E. IGFBPs

IV. Central Nervous System Neoplasms

A. Gliomas/astrocytomas

B. Meningiomas

V. Gastrointestinal Neoplasms

A. Colon cancer

B. Gastric cancer

C. Pancreatic cancer

D. Other (esophageal/hepatocellular)

VI. Head, Neck, and Pulmonary Neoplasms

A. Lung cancer (small cell/non-small cell)

B. Thyroid cancer

VII. Female Reproductive Neoplasms

A. Breast cancer

B. Ovarian cancer

C. Other (endometrial, vaginal, cervical)

VIII. Male Reproductive Neoplasms

A. Prostate cancer

B. Testicular cancer

IX. Genitourinary Neoplasms

A. Renal cell carcinoma

B. Bladder cancer

X. Bone Neoplasms

A. Osteosarcoma

B. Other (chondrosarcoma/fibrosarcoma)

XI. Skin Neoplasms

A. Melanoma

B. Basal/squamous

XII. Hematological Malignancies

XIII. Summary

	I. Introduction

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
THE ASSOCIATION between the insulin-like growth factors (IGFs) and neoplasia has been a subject of investigation for many years. In this manuscript, we have reviewed the available scientific knowledge relevant to understanding the nature of this relationship. This information has been organized by grouping the tumors together by organ system (based upon tissue of origin). For each type of neoplasm, four key elements of the relationship between the IGFs and tumorigenesis and tumor growth will be discussed. These are as follows: 1) presence or absence of autocrine IGF-I and IGF-II production, particularly in relation to normal tissue; 2) presence or absence of IGF-I and IGF-II receptor expression; 3) effects of IGF-I and IGF-II on tumorigenesis and tumor growth; and 4) relevant clinical studies involving therapeutic interventions. When possible, the mechanisms whereby the IGFs are exerting their growth-promoting effects (e.g., decreased apoptosis, increased cell proliferation, or angiogenesis) are also discussed, as is the importance of the binding proteins. We begin the manuscript with a brief overview of what is known about the function and regulation of expression of the IGFs and their receptors.

	II. Overview: IGF Physiology and Gene Regulation

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
IGF-I and IGF-II share approximately 50% structural homology to insulin. The majority of circulating IGF-I and IGF-II are produced by the liver, although various tissues have the capability to synthesize these peptides locally. The hepatic synthesis of IGF-I is largely GH dependent, whereas the synthesis of IGF-II is relatively independent of GH. The GH/IGF-I axis is the primarily regulator of postnatal growth while IGF-II appears to have an important role during fetal development (1, 2, 3).

The IGFs are present in the circulation in combination with high- and low-affinity binders, which form the IGF-binding protein (IGFBP) superfamily. A total of six high-affinity binding proteins have been identified, IGFBP-1 through IGFBP-6 (4, 5, 6, 7). Although the majority of the circulating IGFBPs are synthesized in the liver, many other organs are capable of production of IGFBPs. IGFBP-3 is the most abundant binding protein in the serum. It forms a ternary complex with IGF-I and an acid-labile subunit. The IGFBPs have higher affinities for IGFs (k_d 10^-10 M) than do the type I IGF receptors (k_d 10^-8 M). The IGF/IGFBP complex is acted upon by proteases at the target organ, whereby IGF is released and is available for biological actions. The IGFBPs have stimulatory and/or inhibitory actions on cell proliferation, and these effects may be dependent or independent of IGF. Chen et al. (8) in 1994 demonstrated that addition of exogenous IGFBP-2 and -3 increased IGF-I stimulated [³H]thymidine incorporation in the estrogen receptor-positive cell line, MCF-7. This was thought to be due to a conformational change in the IGFBP-3 as a result of cell surface association, leading to reduced affinity and increased bioavailability of IGF-I and -II.

In addition to the high-affinity IGFBPs, several low-affinity binding proteins, termed IGFBP-related proteins (IGFBP-rP), have been described. The IGFBP-rPs are cysteine rich, have structural similarities to the N terminus of the IGFBPs, and bind to IGFs with a 100-fold lower affinity compared with the IGFBPs. The IGFBP-rPs have actions that are predominantly independent of IGFs.

The IGF-I receptor is a transmembrane heterotetramer consisting of two - and two ß-subunits. There is approximately 60% sequence homology between the IGF-I receptor and the insulin receptor. The IGF-I receptor, like the insulin receptor, possesses tyrosine kinase activity. The postreceptor signal transduction events include phosphorylation of insulin receptor substrate-1 (IRS-1) and activation of phosphatidylinositol-3 (PI-3) and mitogen-activated protein kinases. IGF-II and insulin also bind to the IGF-I receptor but with 2- to 15- and 1,000-fold lower affinity, respectively (9). The IGF-I receptor has been shown to protect cells from apoptosis in vitro and in vivo. Resnicoff et al. (10) demonstrated that for a number of different cell lines, including a human melanoma cell line, decreasing the number of IGF-I receptors by antisense oligodeoxynucleotides was associated with decreased cell growth in vitro and in vivo. The number of apoptotic cells was significantly increased in these cell lines as measured by various techniques. Valentinis et al. (11) in 1999 suggested that the paradoxical ability of IGF-I to cause unregulated cell growth, on one hand, or lead to terminal differentiation, on the other, may depend on the balance between various intracellular substrates. In their study using murine hemopoietic cells (32D), which lack IRS-1, exogenous IGF-I induced differentiation of these cells along the granulocytic pathway after interleukin-3 withdrawal. The IGF-I-induced differentiation in the absence of IRS-1 required the C terminus of the IGF-I receptor. Expression of IRS-1 in these cells was associated with loss of differentiation and caused unregulated growth, and the degree of inhibition of differentiation correlated with IRS-1 expression. The authors also suggested that overexpression of Shc, another substrate of IGF-I receptor, potentiated 32D cell differentiation while a dominant negative mutant of Shc partially inhibited differentiation.

The IGF-II/cation-independent mannose-6-phosphate (IGF-II/ Man-6-P) receptor is a monomeric receptor that binds IGF-II with a 500-fold increased affinity over IGF-I. The IGF-II receptor does not bind insulin. Most of the biological actions of IGF-II are thought to be mediated via the IGF-I receptor (12, 13). Four classes of ligands are currently known to bind to the extracytoplasmic receptor domain. They include mannose 6-phosphate-containing lysosomal enzymes, IGF-II, retinoic acid, and urokinase-type plasminogen activator receptor. The IGF-II receptor is thought to function primarily as a scavenger receptor, regulating the internalization and degradation of extracellular IGF-II, thus regulating the circulating IGF-II levels. IGF-II receptor mutant animals have been demonstrated to have increased circulating and serum levels of IGF-II along with an increased birth weight, organomegaly, and perinatal mortality. The IGF-II receptor also regulates intracellular trafficking of lysosomal enzymes including cathepsin, which serves as an IGFBP proteolytic enzyme. Recently, O’Gorman et al. (14) described enhanced tumor growth in a choriocarcinoma cell line, JEG-3, after IGF-II receptor expression was decreased by antisense IGF-II cDNA constructs. The resultant decrease in cathepsin degradation, yielding an increase in IGFBP proteolysis and bioavailable IGFs, may account for this enhanced tumor growth. Although most of the mitogenic/metabolic effects of IGF-II signaling are mediated via the IGF-I receptor, some effects, including Na⁺/H⁺ exchange and production of inositol triphosphate in renal tubular cells, stimulation of Ca⁺ influx, and DNA synthesis in BALB/c 3T3 cells, may be mediated by binding of IGF-II to its receptor (12, 13).

A. IGF-I gene expression
The prepro-IGF-I gene consists of six exons in most mammalian species and is located on chromosome 12 in humans. The majority of circulating IGF-I is produced in the liver, and hepatic production is principally regulated by GH. Although the liver-specific deletion of the IGF-I gene using the Cre/loxP recombination system reduced circulating IGF-I concentration by approximately 80% in mice, the growth rates of these transgenic animals was not significantly different when compared with the wild-type animals. This was demonstrated by Yakar et al. (15) and suggests the importance of extrahepatic, autocrine/paracrine production of IGF-I in growth regulation. In some tissues, IGF-I gene expression is predominantly under the control of other hormones: estradiol in the endometrium, gonadotropins in the gonadal tissue, TSH in the thyroid, etc. (16, 17, 18). IGF-I mRNA levels have been also demonstrated to be altered by nutritional state and developmental stage. For instance, fasting reduces serum IGF-I concentrations as much as 75% and negates the effects of GH stimulation. Refeeding rapidly restores parameters to normal (19).

The coding region of prepro-IGF-I is flanked by complex 5'- and 3'-untranslated regions that result in considerable heterogeneity in mature IGF-I transcripts. For instance, in the rat it has been demonstrated that there are separate start sites present in both exons 1 and 2. In exon 1, transcription can be initiated from several different sites over a several hundred base pair region. This broad range of sites exists because there are no core promoter elements such as TATA and CAAT box motifs in exon 1. TATA and CAAT box motifs elements, however, are present upstream of the cluster of start sites present in exon 2. Nevertheless, in most tissues, the majority of transcripts arise from sites in exon 1. As a result of different start sites and alternative splicing, there are a variety of different 5'-untranslated regions that can be present in the mature mRNA transcript. Translation of prepro-IGF-I can be initiated from codons in exons 1, 2, or 3. Alternate polyadenylation sites in the 3'-untranslated region of the molecule also contribute to differences in transcript size. On Northern analysis, IGF-I transcripts ranging from <1 kb to approximately 7.5 kb are observed (20).

B. IGF-II gene expression
The human prepro-IGF-II gene consists of nine exons and is located on chromosome 11. The first six exons are noncoding. There are four promoters present (P1–4). There is one promoter each in the regions upstream of exons 1, 4, 5, and 6. During fetal development, IGF-II expression is much higher than in the postnatal period or in the adult. A distinct pattern of promoter use correlates with expression levels during development. In the fetus, promoters P2–4 are active in the liver. After birth, the use of these three promoters declines and P1 becomes dominant. There is some degree of tissue-specific regulation (21). The IGF-II gene is one of the few genes known to have parental allele-specific expression. As such, it is referred to as an imprinted gene. In normal cells, IGF-II is maternally imprinted in that it is expressed only from the paternal copy of the gene. The IGF-II gene is located on chromosome 11p15.5 close to H19, a paternally imprinted gene. The imprinting process is an early event, taking place at the time of gametogenesis. Loss of IGF-II imprinting has been reported in a variety of tumors, including Wilms’ tumors, gastric adenocarcinomas, lung cancers, gliomas, hepatoblastomas, leiomyosarcomas, cervical cancer, prostate cancer, choriocarcinomas, rhabdomyosarcomas, seminomas, and ovarian carcinomas (22, 23, 24). When loss of imprinting occurs, biallelic expression of IGF-II results, ultimately leading to overexpression of this potent growth factor. The precise role of loss of IGF-II imprinting in tumorigenesis and tumor growth is unknown at this time. In some neoplasms, such as Wilms’ tumor, loss of imprinting is an event that occurs early in carcinogenesis. In others, such as cervical carcinoma, it is not an early event. The precise relationship with H19, which may have tumor suppressor function, is also not completely known. In Wilms’ tumor, the loss of imprinting is often linked with a reduced expression of H19. This reciprocal relationship, however, has not been commonly detected in other tumor types, such as hepatoblastomas, gliomas, testicular tumors, and cervical cancers.

C. IGF-I receptor gene expression
The highest levels of IGF-I receptor mRNA expression occur during fetal development and in the early postnatal period. Although IGF-I receptor expression is significantly down-regulated in the adult, it is present in most types of tissue. Its expression is up-regulated to some degree by fasting; up-regulation has been reported in the kidney in an experimental diabetes model (25). IGF-I decreases IGF-I receptor expression in a dose-dependent manner in FRTL-5, IM-9, and endothelial cells (26, 27).

The IGF-I receptor gene promoter lacks both TATA and CAAT box motifs. The 5'-untranslated region is very GC-rich and contains multiple Sp1 consensus-binding sequences. Accordingly, Sp1 has been demonstrated to potently activate IGF-I receptor gene transcription (28). Transcription is initiated from a single start site. Basic fibroblastic growth factor (bFGF) has also been demonstrated to increase IGF-I receptor mRNA levels and activate the IGF-I receptor promoter. The effects of bFGF have been localized to a region of the IGF-I promoter located between nucleotides -476 and -188 (29). Transfection experiments have demonstrated that the tumor suppressor p53 inhibits activity of the IGF-I receptor promoter. Mutant versions of p53, frequently present in malignant states, result in increased IGF-I receptor gene activation (30). The increased level of IGF-I receptor expression present in some malignant tumors may enhance response to autocrine, paracrine, or circulating IGFs.

D. IGF-II receptor gene expression
The IGF-II receptor binds IGF-II and ligands containing a mannose 6-phosphate recognition marker (lysosomal enzymes). Unlike the IGF-I receptor gene, this receptor is a large single-chain peptide that has no intrinsic tyrosine kinase activity (31). Its primary function seems to be transport of its ligands to liposomes, resulting in either their activation or degradation. The majority of IGF-II receptors are located on intracellular membranes. Most of the physiological actions of IGF-II are thought to be mediated through binding to the IGF-I receptor. However, binding of IGF-II to its receptor has been demonstrated to trigger interaction with a membrane-bound GTP-binding protein, Gi-2, that mediates the influx of calcium into the cell (32).

The mouse IGF-II receptor gene is 93 kb in size and contains 48 exons. The gene contains a strong minimal promoter of 266 bp or less. An extended 54-bp footprint within the proximal promoter containing two E-boxes and probable binding sites for Sp1, nerve growth factor-IA (NGF-IA), and related proteins has also been identified. Deletion of the 54-bp segment resulted in an 8-fold decline in promoter activity. Mutational analyses demonstrated that each E box contributed to more than half of the enhancers activity (33).

	III. Potential Therapeutic Agents

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
Currently, modulating the GH/IGF-I axis remains an experimental antineoplastic strategy. Nevertheless, a number of different agents are available that offer potential benefits. Virtually every level of the GH/IGF-I axis, from the hypothalamic hormones to the receptors mediating response on the tumor tissues, can be targeted (Fig. 1). The major classes of therapeutic approaches are summarized in the following section.

View larger version (13K): [in this window] [in a new window]   Figure 1. Sites of action of various pharmacological agents available to modify GH/IGF-I axis. 1, GH-releasing hormone (GHRH) antagonists; 2, somatostatin analogs; 3, GH receptor antagonists; 4, IGF-I receptor antibody; 5, IGF-I/II mRNA antisense vector strategies; and 6, IGFBPs.

The GHRH antagonists have also been shown to inhibit autocrine/paracrine production of IGF-I/IGF-II by acting directly on the tumor tissues. In a study by Csernus et al. in 1999, the GHRH antagonists, MZ-4–71 and MZ-5–156, were shown to inhibit autocrine production of IGF-II, reduce IGF-II mRNA expression, and decrease cell proliferation in a number of cell lines including the breast cancer cell lines MDA-MB-468 and ZR-75–1, prostate cancer cell lines PC-3 and DU-145, and the pancreatic cancer cell lines MiaPaCa-2, Capan-2, and SW-1990 (35). Because IGF-II is a potent mitogen for a variety of cell lines, and overexpression of IGF-II mRNA has been demonstrated for some tumors, reduction of IGF-II by GHRH antagonists could be of potential therapeutic use in the management of such tumors.

Thus, the GHRH antagonists could inhibit tumor growth through direct or indirect pathways. The indirect mechanism would be secondary to a reduction in pituitary GH release and subsequent hepatic IGF-I production, leading to a decrease in circulating, "endocrine" IGF-I. The GHRH antagonists may also directly inhibit intratumor IGF-I and/or IGF-II mRNA expression and decrease the autocrine/paracrine production of these growth factors.

B. Somatostatin analogs
Somatostatin is a peptide synthesized as part of a large precursor molecule that is cleaved and processed to yield several mature products, two of them being somatostatin-14 and somatostatin-28, the latter a congener of somatostatin extended at the N terminus. Somatostatin acts as a neurotransmitter in the central nervous system, regulates the release of GH and TSH, and has a regulatory role in the gastrointestinal system and endocrine and exocrine pancreas. The actions of somatostatin are mediated through somatostatin receptors, five of which have been characterized. These membrane receptors are present in multiple organ systems. The receptors are linked to adenylate cyclase through a coupling mechanism involving guanine nucleotide-binding (G) protein. As the half-life of natural somatostatin is very short, longer acting analogs have been developed, octreotide being the first such analog to come into clinical use. Other analogs include octreotide-LAR, lanreotide, and vapreotide. Somatostatin analogs can be expected to reduce circulating IGF-I levels by 30–50% in nonacromegaly patients (36, 37).

The antineoplastic activity of the somatostatin analogs could result from a reduction of GH and IGF-I secretion, which would, in turn, have an inhibitory effect on cell proliferation. As the somatostatin analogs also inhibit secretion of many gastrointestinal hormones, including gastrin, cholecystokinin (CCK), secretin, and bombesin, which have been implicated as mitogens for a variety of neoplastic cells, modulation of these gastrointestinal hormones may also be responsible for the antineoplastic activity of somatostatin analogs. Furthermore, a more "direct" inhibitory effect of these compounds on cell proliferation has been demonstrated. Somatostatin analogs have been shown to stimulate tyrosine phosphatase activity in certain cell lines, an effect that antagonizes the mitogenic effect of growth factors acting on tyrosine kinase receptors such as epidermal growth factor (EGF), bFGF, and IGF-I (38, 39). This direct mode of action could likely account for the antiproliferative effect of somatostatin analogs observed in vitro.

C. GH receptor antagonists
Pegvisomant (B2036-PEG) is a recombinant protein that is structurally similar to natural human GH with the exception that it contains 9 amino acid substitutions. It binds to the GH receptor but does not initiate signal transduction, thereby functioning as a competitive antagonist with natural GH (Fig. 2). Although the half-life of natural GH is only approximately 20 min, pegvisomant has several covalently attached polyethylene glycol molecules that significantly prolong its half-life (100 h), allowing it to be administered subcutaneously on a once-daily basis. Only a limited number of studies have been performed with pegvisomant in respect to analyzing effects on tumor growth. Our laboratories, however, have demonstrated that it is possible to achieve at least a 75% reduction in circulating IGF-I concentrations when pegvisomant is administered to immunocompromised mice (Fig. 3).

View larger version (14K): [in this window] [in a new window]   Figure 2. Pegvisomant functions as a competitive antagonist at the GH receptor. Normal GH has two binding sites, each of which must bind to a separate GH receptor to initiate signal transduction (left part of inset). Pegvisomant functions as a competitive antagonist because binding site 1 is functional and binding site 2 is not (right part of inset). Because pegvisomant is present in molar excess when compared with normal GH, it occupies the majority of available binding sites, resulting in a dose-dependent decrease in GH-stimulated IGF-I production by the liver.

View larger version (15K): [in this window] [in a new window]   Figure 3. Dose response to the competitive GH receptor antagonist, pegvisomant. Serum IGF-I concentration in mice after 1 week of treatment with vehicle or 50–500 mg/kg/day of pegvisomant.

The expression of dominant negative mutants of the IGF-I receptor have also been demonstrated to have antitumor effects in vitro and in vivo. Reiss et al. (43) transfected human colon, lung, prostate, kidney, and ovarian cancer cell lines with plasmids expressing the dominant negative mutant of the IGF-I receptor, 486/STOP, which has a frameshift mutation resulting in a stop codon at residue 486. They demonstrated a >75% inhibition in colony formation in soft agar in all transfected clones. The in vivo tumorigenicity of the cells expressing the negative mutant of the IGF-I receptor was also significantly decreased when compared with the wild-type cells. Furthermore, when the transfected cells were coinjected with a tumor forming cell line, there was significant inhibition of tumor growth, which suggested that the dominant negative mutant has a bystander effect.

A monoclonal antibody to the IGF-I receptor (IR3) was developed in 1983 (44). This antibody binds to the ligand-binding domain on the IGF-I receptor and inhibits IGF-I-mediated effects. This antibody acts as a competitive inhibitor and does not have a direct cytotoxic effect as its inhibitory effect can be overcome by increasing IGF-I concentration. Because IGF-II and insulin also bind to the IGF-I receptor, the use of this antibody has also been shown to inhibit cell growth in response to these growth factors.

IGF-I peptide analogs that inhibit IGF-I receptor autophosphorylation have been developed (45, 46). Blakesley et al. (47) in 1996 reported a significant reduction in tumor growth in mice injected with fibroblasts which expressed mutations at the tyrosine residue of the carboxy terminus of the IGF-I receptor.

E. IGFBPs
As reviewed previously, the IGFs circulate in combination with six high-affinity binding proteins, IGFBP 1–6. The mammalian IGFBPs contain three distinct domains of roughly equivalent sizes: the N-terminal, the midregion, and the C-terminal. The N- and C-terminal domains are conserved, and the midregion is highly variable. The human IGFBPs share approximately 36% similarity. The C-terminal domain is responsible primarily for high-affinity binding, whereas the N-terminal and perhaps the midregion are involved in low-affinity binding. The high-affinity IGFBPs modulate IGF bioavailability by undergoing proteolysis and generating fragments with reduced or no affinity for the IGFs. In the case of human IGFBP-1, posttranslational phosphorylation has been demonstrated to enhance IGF binding (7). In addition to modulating IGF bioavailability and thus IGF-dependent functions, the IGFBPs also have functions that are unrelated to IGFs and thus are IGF independent. The C-terminal domain and the midregion seem to be involved in IGF-independent actions. In addition to the IGFBPs, the IGFBP-rPs can have actions that are independent of IGFs on cell growth. Recombinant forms of these proteins, therefore, are potential therapeutic tools.

	IV. Central Nervous System Neoplasms

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
A.Glioblastomas/astrocytomas
1. IGF-I expression. Sandberg et al. (48) in 1988 demonstrated that IGF-I mRNA was present in both adult and fetal human brain tissue using slot blot and Northern analysis. Within the adult brain, the following rank order of IGF-I expression was observed: pons > cerebellum > cerebral cortex > thalamus. In comparison, the concentrations measured in three human gliomas and one anaplastic astrocytoma were approximately 1.1- to 4.0-fold greater than the highest levels observed in normal brain tissue. Antoniades et al. (49) in 1992 confirmed the presence of IGF-I expression in 10/10 human astrocytoma specimens with in situ hybridization and immunohistochemistry. Using an IGF-I immunoassay, Glick et al. (50) in 1991 demonstrated the presence of IGF-I in cystic fluid aspirated from 5/5 astrocytomas or glioblastomas (range 3.4–178 µg/liter). Corresponding normal values were 3 µg/liter in cerebrospinal fluid and 388 µg/liter in serum. In a follow-up study, however, primary cultures from 12 human glioma surgical specimens did not demonstrate the presence of IGF-I in the conditioned media (51).

Sandberg-Nordqvist et al. (52) in 1993 demonstrated IGF-I immunoreactivity in 6/9 gliomas and suggested that there appeared to be a positive correlation between IGF-I immunoreactivity and the histopathological grade of the tumor. In a study of 39 astrocytomas (World Health Organization grades II-IV), Hirano et al. (53) in 1999 demonstrated that there is indeed a positive correlation between IGF-I immunoreactivity and histopathological grade (Table 1). They also determined that there was a positive correlation between IGF-I and cell proliferation rates by examining the relationship between IGF-I and Ki-67 (MIB-1) labeling indices. Additional findings of note include the observation that in the high-grade tumors, IGF-I immunoreactivity was greatest in perivascular areas. Proliferating microvessels exhibited more intense staining than nonproliferative vessels. Reactive astrocytes at the margins of tumor infiltration also demonstrated high levels of IGF-I expression.

2. IGF-II expression. Sandberg et al. (48) in 1988 demonstrated that IGF-II mRNA is abundantly expressed in fetal brain but only minimally detectable in normal adult brain. Northern analysis identified approximately a 5- to 50-fold increase in IGF-II mRNA levels in 4/4 human glial tumors as compared with fetal brain. Glick et al. (50) in 1991 identified IGF-II in cyst fluid aspirated intraoperatively in 3/5 glial tumors. The concentration in the cyst fluid in the IGF-II positive tumors ranged from 3.9–131 µg/liter. IGF-II concentrations were 3.3 and 564 µg/liter in cerebrospinal fluid and serum, respectively. In a later study, Glick et al. (51) identified IGF-II in the conditioned media in 4/12 glioma primary cultures. Lichtor et al. (55) in 1993, using Northern blot analysis, did not identify IGF-II mRNA expression in two intermediate grade astrocytomas, three anaplastic astrocytomas, or two glioblastomas. No significant IGF-II mRNA expression was found in the two glioma cell lines, U-87 MG and T-98 G. Hultberg et al. (56) in 1993, also by Northern blot analysis, were unable to detect IGF-II mRNA expression in any of the 9 glial tumors studied. In contrast, Antoniades et al. (49) using in situ hybridization demonstrated the presence of IGF-II mRNA in 10 human glial tumors.

3. IGF receptor expression. Using [¹²⁵I]-labeled ligand, Sara et al. (57) in 1986 identified IGF-I specific binding in plasma membrane preparations from both normal brain tissue and glioblastomas. No statistically significant binding differences were observed between the two groups, each of which were comprised of 6 specimens. Gammeltoft et al. (58) in 1988, also using a competitive ligand binding assay, identified IGF-I binding in 6/6 human glioma cell lines. Glick et al. (59) in 1989 also confirmed IGF-I receptor expression in human glioma specimen by [¹²⁵I]-labeled IGF-I binding studies. Merrill and Edwards (60) in 1990 studied 18 glioma specimens for IGF-I receptor expression and found an increase in IGF-I binding sites in some of the gliomas. The mean number of IGF-I binding sites was 68 (20–133) pmol/g in normal brain tissue and 195 (60–356) pmol/g in the glioma specimens. No difference was detected in affinity characteristics. Cross-linking studies demonstrated that glial tumors expressed the same lower molecular mass ( 118 kDa) -subunit as is expressed in normal brain tissue. In the cell lines derived from the glioma specimens, however, a larger ( 133 kDa) -subunit was present.

Using [¹²⁵I] -labeled IGF-II binding studies, Sara et al. (57) also demonstrated an approximately 2-fold increase in IGF-II binding as compared with normal brain tissue. Gammeltoft et al. (58) found 2 to 5 times more type II IGF binding sites than type I sites. Type II sites bind IGF-II with 10 times greater affinity than IGF-I; type 1 binding sites bind IGF-I and IGF-II with approximately equal affinity.

4. IGF effects on tumor growth. The mitogenic effect of exogenous IGF-I on human glioma cell lines was studied by Merrill and Edwards (60). Using [³H]thymidine incorporation as a tumor growth marker, they reported that 9/10 glioma cell lines responded to IGF-I with increases in [³H]thymidine incorporation that ranged from 1.3 to 4 times the control value. A dose-dependent response to IGF-I was observed. The effect of IGF-I receptor expression on the in vitro and in vivo growth of glioma cell lines has been analyzed by many investigators (61, 62, 63, 64, 65). Resnicoff et al. (66) in 1994 reported complete inhibition of IGF-I-stimulated growth in vitro in C6 rat glioblastoma cells incubated with antisense oligodeoxynucleotides to the IGF-I receptor mRNA. Cells stably transfected with a plasmid expressing an antisense IGF-I receptor mRNA also demonstrated near-complete inhibition of IGF-I- and IGF-II-stimulated growth. This was associated with a reduction in IGF-I receptor phosphorylation and correlated with approximately 50% reduction in IGF-I binding sites. BD-IX rats injected with C6 glioma cells stably transfected with the antisense IGF-I receptor plasmids did not develop tumors, whereas all animals injected with wild-type or sense cells developed tumors within a week. They also demonstrated inhibition of tumor formation when IGF-I receptor antisense C6 cells were injected 3 weeks before wild-type cells, whereas prior injection of sense C6 cells provided no protection against subsequent tumor development. Furthermore, complete regression of established wild-type tumors was seen within 2 weeks of injecting IGF-I receptor antisense cells in the opposite flank.

5. Summary. Many studies have demonstrated that most gliomas express the necessary receptors to respond to IGF stimulation and do so in a wide variety of in vitro and in vivo models. The recent study by Hirano et al. (53), demonstrating that there is a correlation between endogenous IGF-I expression and histopathological grade, is important confirmation of earlier studies suggesting that autocrine IGF production is of functional significance in the clinical setting. Animal studies targeting the IGF-I receptor, such as those by Resnicoff et al. (66), have been sufficiently promising that clinical trials using this strategy have been initiated. How effectively the beneficial effects observed in animals translate into patients should therefore be at least partially answered as the results of these trials become known.

B. Meningiomas
1. IGF-I expression. Glick et al. (50) in 1991 measured IGF-I by RIA in the cystic fluid aspirated from two meningiomas, obtaining concentrations of 216 and 32 ng/ml, respectively. In a subsequent study, IGF-I concentrations, indicative of autocrine production, were measurable in the conditioned media of 5/12 human meningioma primary cell cultures (51). Antoniades et al. (49) detected IGF-I mRNA and IGF-I protein expression in three meningioma specimens and one control human meninges specimen. IGF-I immunoreactivity was detected in 6/12 human meningioma specimens by Lichtor et al. (55) in 1993.

2. IGF-II expression. Lichtor et al. (67) in 1991 demonstrated IGF-II mRNA using Northern blot analysis in 2/2 human meningioma specimens. Glick et al. (50) did not detect IGF-II in cyst fluid aspirated intraoperatively from two meningiomas. In a subsequent in vitro study using human meningioma cell lines, they detected IGF-II in the culture media of 6/11 tumors (51). Antoniades et al. (49) demonstrated IGF-II mRNA expression in three meningiomas and one control meningeal specimen using in situ hybridization. Hultberg et al. (56) demonstrated IGF-II mRNA in all four meningioma specimens by Northern blot analysis and in 3/4 specimens by RIA. In the IGF-II-positive tumors, the levels measured by RIA were 40, 144, and 160 ng/g. Sandberg- Nordqvist et al. (68) in 1997, studied the expression of IGF-II mRNA and IGFBP-2 mRNA. In all of the anaplastic or atypical meningiomas, there was a high ratio of IGF-II to IGFBP-2 mRNA levels. In a 5-yr follow up period of the 8 patients with a high IGF-II to IGFBP-2 ratio, 4 died from complications of the tumor, 2 had recurrence, and 2 were lost to follow-up. Of the 14 patients with a low IGF-II to IGFBP-2 ratio, no tumor-associated deaths or recurrences were observed, leading the authors to postulate that a high IGF-II to IGFBP-2 mRNA ratio is a sign of biologically aggressive behavior in meningiomas.

3. IGF receptor expression. Kurihara et al. (69) in 1989, using [¹²⁵I]-labeled ligand, demonstrated the presence of high- affinity IGF-I binding sites in 8/8 of the meningioma specimens examined. Antoniades et al. (49) identified IGF-I and IGF-II receptor mRNA in 3 meningioma specimens using in situ hybridization. Lichtor et al. (55) demonstrated IGF-I receptor immunoreactivity in 4/12 meningiomas.

4. IGF effects on tumor growth. Kurihara et al. (69) demonstrated that IGF-I increased [³H]thymidine incorporation in primary meningioma cultures in a dose-dependent manner. A maximal response of approximately 350% of control was observed at an IGF-I concentration of 10^-8 M in serum free conditions. Tsutsumi et al. (70) in 1994 also confirmed that IGF-I was a mitogen for meningiomas. [³H]thymidine incorporation was increased 200-225% over control by IGF-I in a dose-dependent manner. Friend et al. (71) in 1999 analyzed the mitogenic effect of IGF-I on primary cultures of 14 human meningioma specimens. [³H]thymidine incorporation was increased by 21, 43, and 176% of control in response to IGF-I doses of 1, 5, and 10 µg/liter, respectively. The GH receptor antagonist, B2036-PEG (pegvisomant), decreased [³H]thymidine incorporation in vitro. Friend et al. (72) in 1999 reported that the GH receptor antagonist, pegvisomant, decreased meningioma xenograft growth in nude mice. After 8 weeks of treatment, the tumor volume in the pegvisomant group had decreased by 38% of the initial volume, whereas the tumor volume in the vehicle-treated group increased by 23%. The serum IGF-I concentration in the pegvisomant group decreased by 20%.

5. Summary. As with gliomas, a number of studies have demonstrated that meningiomas express the necessary receptors to respond to IGF stimulation; the GH receptor also appears to be ubiquitously expressed. Accordingly, activation of these receptors, particularly the IGF-I receptor, has been demonstrated to stimulate mitogenesis. Endogenous IGF-II production by meningiomas, specifically in association with low IGFBP-2, appears to be an accurate predictor of aggressive behavior and increased risk of mortality. This observation, made by Sandberg-Nordqvist et al. (68) in 1997, is likely an illustration of the importance of autocrine IGF-II production in propagating rapid tumor growth. High levels of IGFBP-2 production would appear to ameliorate the stimulatory actions of IGF-II, suggesting that the binding protein environment is perhaps as critical a factor as the IGF production itself.

In vivo attempts to influence tumor growth by modulating GH or the IGFs have been limited, until recently, by the relative lack of reliable animal models. Using a model developed by Jensen et al. (73), in which primary cultures of human tumors are xenografted into the flank of immunocompromised mice, the authors have demonstrated that the GH receptor antagonist, pegvisomant, significantly inhibits tumor growth (72). A newly developed meningioma model, in which human tumors are implanted orthotopically, should permit more extensive analysis concerning the efficacy of targeting GH and the IGFs in these tumors (74).

	V. Gastrointestinal Neoplasms

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
A. Colon cancer
1. IGF-I expression. Tricoli et al. (75) in 1986 demonstrated the presence of IGF-I mRNA in normal colonic epithelium using Northern blot analysis. They analyzed 20 human colon cancer specimens and found that IGF-I mRNA levels were mildly elevated (3- to 5-fold) in 20% of the tumors. Culouscou et al. (76) demonstrated IGF-I production by RIA in the conditioned medium from the human colonic adenocarcinoma HT-29.

2. IGF-II expression. Tricoli et al. (75) demonstrated the presence of IGF-II mRNA by Northern blot analysis in normal colonic epithelium and colon cancer. In approximately 40% of colon cancer specimens, the IGF-II mRNA was increased 10- to 50-fold compared with the normal colonic mucosa. This increase in IGF-II mRNA levels was limited to cancers involving the rectum and rectosigmoid with 60% of rectal and 50% of rectosigmoid cancers overexpressing IGF-II. None of the tumors of cecal or sigmoid origin demonstrated enhanced IGF-II mRNA expression. Among the IGF-II- expressing cancers, the Dukes C tumors possessed higher IGF-II levels as compared with Dukes B tumors, leading the authors to suggest that increased IGF-II mRNA expression may be a marker for aggressive distal colon cancers. Lambert et al. (77) in 1990 examined 21 surgical specimens and found that, compared with the normal colonic epithelium, 30% of colon cancer specimen had IGF-II mRNA overexpression. The amount of overexpression ranged from 2–800 times that seen in normal tissue. No correlation between overexpression of IGF-II mRNA and the site of the tumor was observed in this study.

Zarrilli et al. (78) in 1994 demonstrated the presence of IGF-II mRNA in human colon cancer cell line CaCo-2 by Northern blot analysis. IGF-II levels were high in proliferating cells and decreased more than 10-fold when the cells stopped proliferating and differentiated. IGF-II concentrations in the conditioned medium demonstrated a similar pattern in regard to proliferating and differentiated cells. Guo et al. (79) in 1995 identified IGF-II mRNA expression in six human colon cancer cell lines, HCT 116, COLO 205, COLO 320DM, LoVo, DLD-1, and HT 29. They also demonstrated autocrine production of IGF-II by a specific RRA in the conditioned medium of COLO 205 and HCT 116 cells. Kawamoto et al. (80) in 1998 performed IGF-II immunohistochemical staining on 92 colon cancer surgical specimens and 38 normal colon specimens. They found that 74% of cancer specimens stained for IGF-II compared with 11% of normal. There was a positive correlation between IGF-II staining and size, depth of tumor invasion, and proliferative cell nuclear antigen staining. IGF-II-negative tumors were also associated with a statistically significant increased chance of survival (Fig. 4). Kawamoto et al. (81) in 1999, using in situ hybridization and immunohistochemical staining, demonstrated that normal liver tissue at the invasive margins of metastases overexpresses IGF-II. This finding led them to suggest that hepatic IGF-II might be an important paracrine factor in propagating the growth of hepatic metastases. Freier et al. (82) in 1999, using a RNase protection assay, reported that IGF-II mRNA was increased 40-fold in six human colon cancer specimens compared with adjacent normal colonic mucosa, while protein levels were twice as abundant.

View larger version (14K): [in this window] [in a new window]   Figure 4. Survival curve in patient groups with colon cancer stratified according to IGF-II staining. The survival curve was significantly higher in the IGF-II-negative group than in the IGF-II-positive group. [Reproduced with permission from K. Kawamoto et al.: Oncology 55:242–248, 1998 (80 ) © Karger, Basel.]

4. IGFBPs. Michell et al. (88) in 1997 analyzed the effect of IGFBPs on [³H]thymidine incorporation in the human colon cancer cell lines, COLO205, HT29, and SW620. In fresh medium, IGF-I was more potent in stimulating DNA synthesis than its analog des(1, 3)IGF. This analog differs from IGF-I primarily in respect to its having a much lower affinity for the IGFBPs. In the 24-h cell-conditioned media, however, IGF-I was much less potent when compared with its analog. When the conditioned medium was analyzed for the presence of binding proteins by Western blot, it was demonstrated that all the cell lines secreted IGFBP-4 and COLO205, and SW620 secreted IGFBP-2 as well. The authors suggested that the binding proteins in the culture media formed complexes with IGF-I and therefore reduced the amount of free IGF-I available, thus decreasing IGF-I-mediated DNA synthesis. As the des(1, 3)IGF-I has reduced affinity for the binding proteins, no significant change was seen in des(1, 3)IGF-I-induced DNA synthesis. Singh et al. (89) in 1994 reported that antibody to IGFBP-4 increased both basal and IGF-I-stimulated growth in the colon cancer cell line HT-29. They also demonstrated that cells transfected with an antisense vector to IGFBP-4 cDNA had a higher basal and IGF-I stimulated growth rate when compared with cells transfected with the sense or control vectors. Transfection with the sense vector to IGFBP-4 cDNA was associated with a 5- to 10-fold increase in IGFBP-4 expression; however, it was not associated with a further inhibition of basal and IGF-I stimulated cell growth. The observation that exogenous IGF-I was unable to overcome the inhibitory effect of IGFBP-4 overexpression led the authors to suggest that the effect of IGFBP-4 on inhibition of cell growth may be IGF-I independent.

Macdonald et al. (90) in 1999 demonstrated that the human colon carcinoma cell line, Caco-2, which was stably transfected with an IGFBP-3 expression construct, grew more slowly in vitro than the cells transfected with a control vector.

5. IGF effects on tumor growth. The effect of IGF-I on the growth of two mouse colon cancer cell lines was studied by Koenuma et al. in 1989 (91). IGF-I, added to serum-free media at a concentration of 10 µg/liter, increased the growth of two variants of mouse colon adenocarcinoma 26. The growth of a highly metastatic variant (NL-17) was increased to 490% of control while a variant with lower metastatic potential (NL-44) was increased to 279% of control. The number of binding sites, 1.37 x 10⁵/cell (NL-17) and 1.26 x 10⁵/cell (NL-44) was similar between the two cell lines. Lahm et al. (92) in 1992 studied the effects of IGF-I and IGF-II on eight human colon cancer cell lines. IGF-I and IGF-II were roughly equipotent in stimulating tumor growth; half-maximal responses were observed in the responsive cell lines at IGF concentrations ranging from 1.9–6.5 µg/liter. In a follow-up study, Lahm et al. (93) in 1994 demonstrated that a neutralizing monoclonal antibody against the human IGF-I receptor (IR3) decreased the growth of 7/12 cell lines. As some responsive lines expressed only IGF-II, not IGF-I, the authors concluded that autocrine IGF-II, by binding to and activating the IGF-I receptor, was propagating tumor growth. Accordingly, the growth-inhibitory effects of IR3 could be overcome by adding either exogenous IGF-I or IGF-II to the culture media.

Smith and Solomon (94) in 1988 studied the effect of somatostatin on the growth of human colon cancer cell lines in vivo. Nude mice bearing xenografts of the human colon cancer cell line CX1 developed significantly smaller tumors when treated with 100–300 µg/kg of somatostatin 14 twice daily. In 1991, Qin et al. (95) studied the effects of the somatostatin analog RC-160 (vapreotide) on in vivo growth of the rat colon cancer cell line, DHD/K12, using a syngeneic model. In the RC-160-treated rats (100 µg/kg/day), the final tumor volume was approximately 36% of control. Protein and DNA concentration in the tumors of the treatment group were decreased to 70% and 69%, respectively, of the control values. The percent of bromodeoxyuridine-labeled cells in the treatment group was 35% lower than that in the control group. No similar inhibitory effect was observed on tumor growth in vitro, leading the authors to propose that indirect mechanisms, such as effects of the somatostatin analog on GH/IGF-I, CCK, or gastrin, were mediating the antitumor effects. The in vivo effect of another somatostatin analog, octreotide (SMS 201–995), was studied in mice implanted with the colon cancer cell line CT 26 by Alonso et al. (96) in 1992. Tumor volume, tumor weight, and DNA content were significantly reduced in the mice who received the somatostatin analog. Tumor growth was inhibited by 40% and was accompanied by a prolonged survival (42.5 vs. 48.5 days).

The effect of vapreotide (25 µg/twice daily) on the growth of hepatic metastases of the human colon cancer cell lines, 320 DM and WidR, was studied by Qin et al. in 1992 (97). The mice were randomized to receive either RC-160 or vehicle after intrasplenic tumor injection. There was a decrease in the incidence (25% vs. 38%) and mean number (60 vs. 177 in 320 DM; 77 vs. 135 in WidR) of hepatic metastases and a corresponding increase in survival times (7 days vs. 20 days) in the mice receiving RC-160. When the same cell lines were implanted subcutaneously into the flank, similar beneficial effects on tumor volumes in the treatment groups were observed. In 1992 Dy et al. (98) studied the effects of octreotide (50 µg/kg/day) on the growth of LIM 2412 and LIM 2405 cells implanted subcutaneously in mice. Octreotide inhibited tumor growth by approximately 50%. As inhibitory effects were also observed in in vitro studies, direct effects of octreotide on the tumor, independent of the GH/IGF-I axis, were considered potential mechanisms of action. Duan et al. (99) in 1999 analyzed the effect of pegvisomant on human colon cancer cell line, COLO 205, xenografted in nude mice. After treatment for 16 days, they reported a 39% reduction in tumor volume and a 44% reduction in tumor weight in the pegvisomant group along with reductions in circulating IGF-I and IGFBP-3 levels.

6. Clinical studies. Glass et al. in 1994 measured IGF-I levels in more than 300 healthy people scheduled for a colonoscopy for occult gastrointestinal (GI) bleeding (100). No significant difference in the serum IGF-I concentrations were observed between the group with a normal colonoscopy and the group with colonic polyps or colon cancer. Ma et al. (101) reported the results of a case-control study, nested in the physicians’ health study in 1999. IGF-I and IGFBP-3 levels were measured in more than 14,000 subjects at the beginning of the study. The 193 patients who subsequently developed colon cancer, some up to 15 yr later, had significantly higher baseline IGF-I levels and lower IGFBP-3 levels than age- and weight-matched controls from the same cohort.

Iftikhar et al. (102) studied the effects of the somatostatin analog octreotide on tumor kinetic measurements in 12 patients with primary rectal cancer. Four of the 12 patients who received octreotide had a lower Ki67 immunostaining in the posttreatment biopsy compared with the pretreatment biopsy. None of the six untreated control patients had a lower Ki67 index at the time of the second biopsy. In a randomized trial by Cascinu et al. (103) in 1995, 107 patients with advanced colon cancer refractory to chemotherapy were randomized to receive octreotide 200 µg three times per day 5 days/week or best supportive care. The patients in the octreotide group had a median survival time of 20 weeks compared with 11 weeks in the supportive care group. Although there were no objective responses, 45% of the patients randomized to octreotide had stable disease compared with 15% in the supportive care group. Goldberg et al. (104) in 1995 reported the results of an Eastern Cooperative Oncology Group randomized trial in which 260 patients with advanced colon cancer received either octreotide 150 µg subcutaneously three times per day or placebo/no treatment. No significant difference was observed in time to progression or survival between the two groups.

In neither the 1995 Cascinu study or the Goldberg studies were serum IGF-I levels measured. Cascinu et al., however, in a follow-up study in 1997 (105), did measure IGF-I concentrations. They randomized 75 patients with newly diagnosed colon cancer to receive either octreotide, 200 µg subcutaneously three times per day, or placebo for 2 weeks before surgery. There was approximately a 50% reduction in circulating IGF-I levels in the octreotide-treated group (179 vs. 86). They also studied effects of octreotide treatment on tumor cell kinetics. [³H]thymidine labeling and flow cytometry were used to assess S-phase fraction in the tissue obtained from the baseline endoscopic biopsy specimen and the tissue obtained at the time of surgery. The percentage of cells in the S-phase fraction, as measured by [³H]thymidine incorporation, decreased from 18% to 3% in the octreotide group whereas no difference was observed in the placebo group. By flow cytometry, the decrease in S-phase fraction in the octreotide group was decreased from 27 to 22%. Again, no change was observed in the placebo group.

7. Summary. The increased risk of colonic polyps and colon cancer in the acromegalic population is well known. The recent epidemiological study by Ma et al. (101) suggests that even within the normal population, a high IGF-I/IGFBP-3 ratio may increase the risk of colon cancer development. Once tumors develop, the overexpression of IGF-II that occurs in a substantial subset, originally described by Tricoli et al. (75) and subsequently confirmed by Lambert et al. (77), has been clearly demonstrated by Kawamoto et al. (80) to correlate with an aggressive phenotype. Attempts to understand the utility of agents that modulate GH or the IGFs should be analyzed with respect to whether or not the tumors express high levels of IGF endogenously. Tumors without high autocrine IGF production may be the most responsive to agents that only modify circulating or paracrine IGF production. IGF-II-overexpressing tumors, which appear to be more aggressive, will likely require strategies that decrease tumor IGF production, bioavailability, or action at the receptor level.

B. Gastric cancer
1. IGF-I and IGF-II expression. Thompson et al. (106) in 1990 reported the detection of autocrine IGF-II production in the human gastric carcinoma cell line, LIM-1839. IGF-II mRNA was present and IGF-II was measurable in the conditioned medium of cells grown to confluence. No IGF-I mRNA was detectable. Chung and Antoniades (107) in 1992 demonstrated the presence of IGF-I mRNA by in situ hybridization in three human gastric cancer surgical specimens. Expression was also observed in the adjacent normal epithelium. Guo et al. (108) demonstrated IGF-II mRNA in AGS and SIIA gastric cancer cell lines in 1993 by Northern analysis. No IGF-I mRNA was detected in either cell line.

2. IGF receptor expression. Thompson et al. (106) demonstrated the presence of IGF-I and IGF-II receptors by affinity cross-linking and competitive binding studies in LIM-1839 gastric carcinoma cells. Durrant et al. (109) demonstrated the presence of IGF-I receptors by competitive binding in the gastric cancer cell lines, St16, St42, and MKN45. The number of receptors per cell were 250,190 and 310, respectively. Chung and Antoniades (107) demonstrated the presence of IGF-I receptor mRNA by in situ hybridization in human gastric cancer specimens.

3. IGF effects on tumor growth. The mitogenic effect of exogenous IGF-I and IGF-II on human gastric cancer cell line LIM-1839 was reported by Thompson et al. (106). These growth factors increased cell growth 1.6- to 2.0-fold when added to serum-free medium at 20 and 50 ng/ml. The growth-promoting effect was inhibited by IR3, indicating that both IGF-I and IGF-II stimulate growth via the IGF-I receptor. Durrant et al. (109) also demonstrated that IGF-I has mitogenic effects on the St16, St42, and MKN45 gastric cancer cell lines.

C. Pancreatic cancer
1. IGF-I and IGF-II expression. Ohmura et al. (110) in 1990 demonstrated the autocrine production of IGF-I in MIA-PaCa 2 cells, a human pancreatic cancer line, by measuring IGF-I by RIA in conditioned culture medium. Bergmann et al. (111) did not detect IGF-I mRNA by Northern analysis in four human pancreatic cell lines, including ASPC-1, COLO 357, T3 M4, and PANC-1 cells. In 12 pancreatic surgical specimens, a 32-fold increase was present as compared with the low levels observed in the normal pancreas.

2. IGF receptor expression. Ohmura et al. (110) demonstrated ¹²⁵I-labeled IGF-I binding in the human pancreatic cancer cell line, MIA-PaCa 29. Ishiwata et al. (112) in 1997 reported the presence of IGF-II receptor mRNA in six human pancreatic cancer cell lines by Northern analysis (ASPC-1, CAPAN-1, COLO 357, MIA-PaCa-2, PANC-1, and T3 M4). They also found that 7 of 12 pancreatic cancer surgical specimens overexpressed IGF-II receptor mRNA (5.6-fold) in comparison to normal pancreatic tissue.

3. IGF effects on tumor growth. Bergmann et al. (111) demonstrated that IGF-I stimulated cell growth in human pancreatic cell lines, ASPC-1 and COLO 357. For the ASPC-1 cells, the one-half maximal and maximal stimulation occurred at 1.8 and 5 nM, respectively. In the COLO 357 cells, the corresponding values were 0.3 and 1.3 nM. They also observed inhibition of IGF-I stimulated growth by IR3. An antisense oligonucleotide to the IGF-I receptor inhibited the growth of ASPC-1 and COLO-357 by 34% and 35%, respectively.

D. Other (esophageal, hepatocellular)
a. Esophageal cancer. Oku et al. (113) in 1991 demonstrated that both IGF-I and IGF-II were potent mitogens for the human squamous esophageal cancer cell line TE-3-OS. Both IGF-I and IGF-II increased [³H]thymidine uptake by approximately 300% in a dose-dependent manner with maximal effects being observed at 100 µg/liter concentrations of both peptides. The anti-IGF-I receptor antibody, IR3, inhibited growth stimulated by both IGF-I and IGF-II. Chen et al. (114) in 1991 demonstrated IGF-I mRNA by Northern analysis in the human esophageal cancer cell line CE48T/VGH. IGF-I increased cell growth in serum-free media about 3.5-fold, with maximal stimulation being observed at 10^{- 9} M. [¹²⁵I]-IGF-I binding studies demonstrated 2.5 x 10⁵ receptors per cell.

b. Hepatocellular carcinoma.
1. IGF-I expression.
Tsai et al. (115) demonstrated the presence of IGF-I mRNA by Northern analysis in 10 of 10 human hepatoma cell lines in 1988. Su et al. (116) analyzed 7 human hepatoma surgical specimen for the presence of IGF-I mRNA in 1989. Compared with nontumorous hepatic tissue from an adjacent area, the tumors expressed relatively low levels of IGF-I mRNA. Because the expression of GH receptor mRNA in the tumors was also low on a relative basis, the authors proposed that IGF-I mRNA was low as a result of reduced GH stimulation of transcription.

2. IGF-II expression.
Cariani et al. (117) in 1988 observed a 40- to 100-fold increase in IGF-II mRNA expression in 9 of 40 liver cancer surgical specimens as compared with normal adult liver. They reported that many liver cancers expressed fetal transcripts (6.0 and 5.0 kb) instead of the 5.3-kb transcript present in the normal adult. Su et al. (116) in 1989 also reported that 4 of 7 human liver cancer surgical specimens had expression of IGF-II fetal transcripts (5.6 and 4.5 kb). The authors suggested the presence of the fetal transcript of IGF-II mRNA was reflective of cellular dedifferentiation. The amount of IGF-II mRNA expression in the tumors was highly variable. Lamas et al. (118) in 1991, using in situ hybridization and immunohistochemistry, demonstrated an overexpression of IGF-II mRNA and protein in malignant cells as compared with surrounding nontumorous cells.

D’Errico et al. (119) in 1994 studied IGF-II immunoreactivity in liver sections from 54 patients with hepatocellular carcinoma. IGF-II immunoreactivity was detected in 60% (9 of 15) of the specimens from hepatitis B virus-positive patients and in 26% (10 of 39) of the specimens obtained from hepatitis B virus-negative patients. Rogler et al. (120) in 1994 analyzed the effects of IGF-II overexpression by constructing transgenic mice in which human prepro-IGF-II cDNA was placed under control of the major urinary protein promoter. This led to an increase of up to 30-fold in serum IGF-II concentration in the transgenic animals compared with controls. Over the course of 18 months, the transgenic mice developed an increased number of tumors in general and hepatocellular carcinoma in particular when compared with control animals. Scharf et al. (121) in 1998 using a Northern blot technique could not detect IGF-II mRNA in the human hepatoma cell line PLC. Ng et al. (122) in 1998 reported repression of the expression of the normal adult IGF-II transcript in 93% (28 of 30) of human hepatoma surgical specimens. Fetal transcripts, however, were present in 40% of the specimens. The nontumorous hepatocytes expressed the adult IGF-II mRNA transcript in 93% of the cases. The tumors from older patients were more likely to express IGF-II. Sohda et al. (123) in 1997 detected an overexpression of IGF-II mRNA by in situ hybridization and immunohistochemistry in 50% (5 of 10) of human hepatocellular carcinoma specimens. Cells that expressed IGF-II were more likely to express the Ki-67 antigen (expressed during active cellular proliferation), suggesting that IGF-II acts as an autocrine or paracrine growth factor for these tumors.

3. IGF receptor expression.
Tsai et al. (115) detected the presence of IGF-I receptor mRNA in 10 of 10 human hepatoma cell lines. Scharf et al. (121) identified IGF-I and IGF-II receptor mRNA in the hepatoma cell line PLC.

4. IGF effects on tumor growth.
The effects of exogenous IGF-I and IGF-II on DNA synthesis were also demonstrated by Scharf et al. (121). Both growth factors increased [³H]thymidine incorporation in the PLC cell line in a dose-dependent manner. Lin et al. (124) in 1997 observed that high levels of IGF-II were produced by the human hepatoma cell lines HuH-7 and HepG2. Antisense oligonucleotides complementary to IGF-II mRNA led to reduction in IGF-II mRNA and protein content in the cell lines. There was also a decrease in [³H]thymidine incorporation and cell-proliferative activity. The inhibitory effect on cell growth was not observed in those cell lines that did not overexpress IGF-II.

	VI. Head, Neck, and Pulmonary Neoplasms

Top Abstract I. Introduction II. Overview: IGF Physiology... III. Potential Therapeutic... IV. Central Nervous System... V. Gastrointestinal Neoplasms VI. Head, Neck, and... VII. Female Reproductive... VIII. Male Reproductive... IX. Genitourinary Neoplasms X. Bone Neoplasms XI. Skin Neoplasms XII. Hematological Malignancies XIII. Summary References

 
A. Lung cancer (small cell/non-small cell)
1. IGF-I expression. Minuto et al. (125) in 1986 measured IGF-I concentrations by RIA in tissue extracts from 10 human lung surgical specimens (7 epidermoid/3 adenocarcinoma). The IGF-I concentration in the cancerous tissue was significantly higher than the IGF-I concentration in the surrounding normal lung tissue (615 ± 123 mU/g vs. 234 ± 51). The same group of investigators (126) detected autocrine IGF-I production by RIA in the conditioned culture medium of the human lung cancer cell line CALU-6. Nakanishi et al. (127) in 1988 reported the synthesis of an IGF-I precursor molecule by Western blot analysis by two small cell lung cancer (SCLC) cell lines, NCI-H345 and NCI-N417. Jaques et al. (128) in 1988 detected IGF-I immunoreactivity in the cell pellets and culture media of 11 of 14 separate SCLC cell lines. Macaulay et al. (129) in 1990 demonstrated that the "classic" SCLC cell line HC12, which has neuroendocrine features, produced IGF-I. The "variant" cell line ICR-SC17, which does not have neuroendocrine features, did not express IGF-I. Reeve et al. (130) in 1990 also measured IGF-I immunoreactivity in the conditioned media of lung cancer cell lines. IGF-I was detected in 2 of 2 "classic" SCLC cell lines (COR-L51, COR-L47) and 3 of 3 "variant" cell lines (COR-L27, COR-L24, COR-L103). No IGF-I immunoreactivity was detected in a large cell (COR-L23) or adenocarcinoma (MOR) line.

2. IGF receptor expression. Nakanishi et al. (127) reported the presence of two high-affinity specific binding sites for ¹²⁵I-labeled IGF-I [dissociation constant (K_d) 1.3 and 4.0 nM] in the human SCLC cell line, NCI-H345. Jaques et al. (128) also demonstrated high-affinity IGF-I binding sites in 14 of 14 SCLC cell lines (K_d 0.89–5.21 nM). The maximum binding (B_max) ranged from 131 to 1,230 fmol/mg protein. Rotsch et al. (131) in 1992 described the presence of high-affinity IGF-I binding sites in a number of SCLC and NSCLC lines. They also demonstrated the presence of IGF-I receptor mRNA by Northern blot analysis in all of the