Abstract: The Insulin like Growth Factor-1 (IGF-1) receptors are members of the super family of Receptor Tyrosine Kinase (RTKs) implicated in human cancers due to amplification, overexpression or somatic mutation of the gene. The type-1 insulin like growth receptors (IGF-1R) is overexpressed by many tumors and mediates proliferation, motility and apoptosis protection. Tumor growth and metastasis can be blocked by agents that inhibit IGF-1R expression or function, suggesting that the IGF-1R is a promising treatment target. The strategies to block IGF-1R function employed anti-IGF-1R antibodies, small-molecule inhibitors of the IGF-1R tyrosine kinase, antisense oligodeoxynucleotides and antisense RNA, small inhibitory RNA, triple helix, dominant negative mutant and various compounds reducing ligand availability. Studies show that antisense IGF-1R expression in melanoma cells leads to enhanced radio sensitivity and impaired activation of ATM, required for DNA double strand break repair. Antisense and dominant negative strategies also enhance tumor cell chemosensitivity induced by tumor cells and killed in vivo by IGF-1R antisense. However, antisense agents cause only modest IGF-1R down regulation and can affect the insulin receptor. Specificity is an important issue for the development of both kinase inhibitors and molecular reagents. Using an array-based screen to identify accessible region of IGF-1R mRNA, are designed small interfering RNAs (siRNAs) that induce potent IGF-1R gene silencing without affecting the insulin receptor. These siRNAs block IGF-signaling, enhancing radio and chemosensitivity and show a genuine therapeutic potential. The clinical efficacy of IGF-1R targeting will be determined by key factors including the role of receptors in established tumors. The potency of inhibition achieved in vivo and the extent to which other signaling pathways compensate for IGF-1R loss.
Introduction
Tumor cells exhibit abnormally high levels of proliferation, which is promoted and controlled by a variety of growth factors. Among these, the Insulin-like Growth Factors (IGFs) play a major role in regulating cell proliferation and inhibiting apoptosis. IGFs are expressed ubiquitously and act as an autocrine/paracrine manner through binding to the IGF-I receptor (IGF-IR). The bioactivity of IGF in tissues is determined by both local and systemic factors. The local factors included the levels of the receptors that are expressed, various IGF binding protein (IGFBPs) and IGFBP protease. The systemic factors involved are mainly those that regulate the circulating levels of IGFs, such as Growth Hormone (GH) and various nutritional factors. IGF system is comprised of the IGF ligands (IGF-I and IGF-II), cell surface receptors that mediate the biological effects of the IGFs including the IGF-IR, the IGF-II receptor (IGF-IIR) and Insulin Receptor (IR), as well as a family of IGF-binding proteins (IGFBPs). IGFBPs affect the half-lives and bioavailability of the IGFs in the circulation, in extracellular fluids and may exert IGF-independent effect under certain conditions. This review will focus on the structure and function of the components of the IGF axis, their interactions and their role in tumerogenesis (Baxter, 2000; Clemmons, 2001; LeRoith et al., 1995).
IGF Ligands
Structure of IGF-I and IGF-II
The mature IGF-I and IGF-II peptides consist of B and A chain of insulin.
Unlike insulin, the IGF peptides are not prototypically cleaved, but remain
linked in the mature peptides by C domains analogous to the C peptide of insulin.
Both IGF-I and IGF-II contain an additional short D domain that is not found
in insulin. The IGF-I and IGF-II prohormones contain a C-terminal E peptide
that is cleaved in the Golgi apparatus during secretion (Daughaday and Rotwein,
1989).
Expression of IGF-1 and IGF-II
The prenatal expression of the IGF-II gene is widespread in rodent and diminished
dramatically after birth only with the choroids plexus and the leptomeninges
persistently synthesizing IGF-II in adult animals. In contrast, murine expression
of IGF-I is low during the prenatal period and increases significantly during
puberty and adulthood, when hepatic productions become a major contributor to
overall circulating IGF-1 levels. IGF-I exerts endocrine, paracrine and autocrine
effects and is produced by numerous other adult organs, including kidney, lung
and bone. The overall picture of IGF expression in rodent initially led to the
concept of IGF-II as a fetal growth factor and IGF-I as an adult growth factor.
However, this expression pattern is not observed in human; as both IGF-I and
IGF-II are produced in multiple human tissues throughout life. Infact, the circulating
levels of human IGF-II are consistently several fold higher than that of IGF-I,
which is consistent with the concept that IGF-I and IGF-II have potentially
divergent roles in human physiology (Daughaday and Rotwein, 1989).
IGF Receptors
IGF and Insulin Receptors
The IGF-I and IGF-II ligands interact with an array of cell surface receptors
that may be present singly or in various combinations on target cells. Both
IGF-I and IGF-II interact with the IGF-IR, a transmembrane tyrosine kinase that
is structurally and functionally related to the IR. IGF-II can also bound to
the IGF-II-R with high affinity. Cloning of the previously characterized cation-independent
mannose-6-phosphate (M6P) receptor, which plays a role in endocytosis and intracellular
trafficking of M6P-tagged proteins. This molecule is thought to function as
a clearance receptor for IGF-II, thereby enhancing the extra cellular levels
of IGF-II (Daughaday and Rotwein, 1989). Most, if not all, of the effects of
IGF-I result from activation of the IGF-IR. IGF-I does not cross react with
the IR, except at pharmacological doses, since IGF-I has a relative affinity
almost two order of magnitude higher for the IGF-IR, as compared to the IR.
Until recently, it was thought that IGF-II like IGF-I, bound at significant
levels to the IGF-IR but not to the IR. Studies of knockout mice lacking various
components of the IGF and insulin receptor systems suggested that IGF-II acted
through the IR during the early stages of development, before IGF-IR gene expression
was detectable. The molecular basis for the phenomenon was revealed when it
was discovered that a splice variant of the IR displayed high affinity for IGF-II.
The IR transcript is subject to alternative splicing of exon 11, which encodes
a 12-aminoacid segment in the C-terminus of the extra cellular subunit. Previous
studies had shown that the IR isoform encoded by the mRNA lacking the exon 11
sequence (IR-A) displayed a 2 fold higher affinity for insulin than the IR-B
form, which includes exon 11. More recently, it has been reported that the IR-A
isoform functions as a high affinity receptor for IGF-II and mediates predominantly
proliferative effects as compared to the principally metabolic effects elicited
by insulin stimulation of the IR-B isoform (Rother and Accili, 2000). Thus,
IGF-1 functions primarily by activating the IGF-IR, whereas IGF-II can act through
either the IGF-IR or through the IR-A isoform.
Hybrid Receptors
The complexity of IGF signaling is increased by the formation of hybrid
receptors that result from the dimerization of IGF-IR and IR hemireceptors.
Each hybrid receptor consists of a single α and β subunits linked
by disulfide bonds, which are formed in the Golgi apparatus of cells expressing
both the IGF-IR and the IR. These could be due to the preferential formation
of disulfide bonds between cysteine residue in IGF-IR and IR subunit themselves.
Thus, in some circumstances, hybrid receptors may outnumber homoreceptor molecules
at the cell surface. IGF-IR/IR hybrid receptors retain high affinity for IGF-I,
but exhibit a dramatically decreased affinity for insulin. This is thought to
reflect the ability of IGF-I to efficiently bind to either IGF-IR α. Whereas
tight insulin binding requires interaction with both of the β subunit found
in the IR. Thus the presences of a significant number of hybrid receptor may
selective diminish cellular responsiveness to insulin, but not to IGF-I. Indeed,
this has been proposed as a mechanism by which up-regulation of IGF-IR expression
could cause insulin resistance in cells that express the IR. The effect of hybrid
receptors are further complicated by the presence of the IR-A and IR-B isoforms
and their different IGF-II binding characteristics. It is likely that hybrids
formed between both IR-A/IGF-IR and IR-B/IGF-IR form, since most cells express
both splice variants. It has been recently demonstrated that IGF-IR/IR-A hybrids
bind IGF-I, IGF-II and insulin, whereas IGF-IR/IR-B hybrids bind IGF-I with
high affinity, IGF-II with low affinity and do not bind insulin (Frasca et
al., 1999). Therefore, the relative expression of the IGF-IR and IR genes
and the degree of alternative splicing of exon 11 of the IR gene governs the
ability of a given cells respond to IGF-I, IGF-II and insulin. It confirms potential
receptor hybrid that may be involved in IGF signaling.
IGF-IR and IR Signal Transudation
IGF-IR is an evolutionary conserved, ubiquitous transmembrame tyrosine kinase, structurally similar to the Insulin Receptor (IR) (Pandini et al., 2002). IGF-IR is composed of two extracellular α subunits and two extracellular β subunits. The α subunit have ligands (IGF-I, IGF-II and insulin at supraphysiological doses) binding site, while β subunit contain three major domains; the juxtamembrane domain, tyrosine kinase domain and C-terminus. The tyrosine kinase domain share high (85%) homology with its counterpart in IR, while the C-terminus is only 40% homologous with the C-terminus of IR (Pandini et al., 2002). Binding of ligands to IGF-IR induces its autophosphorylation and tyrosine phosphorylation of IGF-IR substrate, especially the IR substrate 1 (IRS-1) and src and collagen-homology (SHC) protein. Tyrosine-phosphorylated IRS-1 and SHC bind different effector proteins (enzymes and/or adaptors) inducing multiple signaling cascades, among them several interconnecting pathways controlling cell survival and proliferation (Ullrich, 1986; Shepherd et al., 1998; White, 1994, 2002; Adams et al., 2000; Surmacz, 2000). The critical survival pathway activated by IGF-I stems from IRS-1. IRS-1 recruits and stimulates the PI-3 kinase (PI-3k), which then transmits signal to the serine/threonine kinase Akt. The activated Akt phosphorylates and block a variety of propoptotic proteins, including BAD, caspase-9, beta kinase. Furthermore, Akt induces the expression of antiapoptotic proteins, for example, Bcl-2 (Brazil et al., 2002; Hill and Hemmings, 2002; Nicholson and Anderson, 2002). Other mitogenic survival of IGF-IR pathways involves signal transducers and activators of transcription (STATs) that are phosphorylated and activated by IGF-I through JAK-1/2 and PI-3k/Akt pathway (Zong et al., 2000). In addition, IGF-IR can prevent cell death or induce proliferation via the SHC/Ras/ERK1/2 pathway (Zong et al., 1998; Peruzzi et al., 1999). While antiapoptotic and growth pathways of IGF-IR have been extensively studied, the signals controlling non-mitogenic functions of IGF-IR, such as well substrate adhesion, migration, invasion, or intracellular interactions are not well understood. There is increasing evidence that IGF-IR pathways interconnect with interin and cadherin signaling system (Vuori and Ruoslahti, 1994). In some experimental model, IGF-IR has been shown to mediate metastasis, possibly through-enhanced migration (Doerr and Jones, 1996; Bartucci et al., 2001) reduced cell-cell adhesion (Mauro et al., 2003) and upregulation of Plasminogen Activator (PA) and matrix metalloproteinase (Long et al., 1998; Mira et al., 1999; Dunn et al., 1998; Zhang and Brodt, 2003).
Insulin like Growth Factor Binding Protein (IGFBP)
The biological activities of the IGF ligands are also modulated by a family of high affinity IGFBP-1 to 6 that are found in the circulation and in extra cellular fluids. IGFBP-3 is the predominant IGFBP in serum and the most circulating IGF-I and IGF-II are not found in free form, but as a ternary complex with IGFBP-3 and the acid-labile subunit (ALS), in 1:1:1 molar ratio. IGFBP-5 also form ternary complexes with IGFs and ALS. While IGFBPs-1 through 4 generally has similar affinity for IGF-I and IGF-II. IGFBP-5 and 6 bind IGF-II with 10 and 100-fold greater affinities respectively, than IGF-I. The IGFBPs do not bind with insulin (Guvakova and Surmacz, 1999; Mauro et al., 1999). The IGFBPs control IGF action by increasing the half lives of circulating IGFs, by controlling their availability for receptor binding and in the case of cell surface associated IGFBPs; by potentially influencing their direct interaction with receptors (Baxter, 2000). Thus ligand receptor interactions in the IGF system are subjected to complex regulation as a result of the levels of IGFBPs, their expression profile, the degree of cell surface association and the extent of proteolysis. A series of studies performed over the past several year has established that certain actions of the IGFBPs are IGF-independent. IGFBP-3 and IGFBP-5, in particular have been shown to exhibit effects on proliferation migration and sensitivity to apoptosis that are independent of their effects on IGF-signaling per se. Some of these IGF-independent effects are still modulated by IGF-binding to the particular IGFBP, so ‘IGF-receptor-independent actions’ may be a more accurate term for these novel functions. The cell surface and/or intracellular molecule that participate in these effects have not been well characterized, but exposure to exogenous recombinant IGFBP-3 and IGFBP-5 proteins has been shown to induce nuclear localization of these proteins. A better characterization of these IGF-receptor-independent actions of IGFBPs will provide an important new dimension to our understanding of the IGF signaling system in general (Reiss et al., 2001; Shaw, 2001).
Physiological and Pathophysiological aspects of IGF Action
The IGF system plays an important role in normal growth and development as well as in a variety of pathological situations, particularly tumerogenesis (Khandwala et al., 2000). IGF action is also important in the development of specific organ, such as in the nervous system, in which IGF-signaling regulate neuronal proliferation, apoptosis and cell survival. IGF action plays a critical role in the development and progression of human cancer. A growing body of epidemiological data suggested that high levels of circulating IGF-1 constitute the risk factor for the development of breast, prostate, colon and lung cancer outcomes. As a result of some experimental findings, intensive efforts is being directed towards investigating the utility of the IGF system as both a diagnostic marker and a therapeutic target in cancer therapy.
Role of IR Signaling Cascade in Cancer Cell Function
Regulation of IGF-IR gene expression is closely associated with the function of a variety of tumor suppressor genes and oncogenes. The p-53 tumor suppressor protein protects mammalian cells against cancer. A large number of human cancer cells exhibit mutations within the p-53 gene that either impair its tumor suppressor function or provide it with oncogenic potential. Expression of wild-type p-53 inhibits IGF-IR gene expression, whereas mutant p-53 upregulates IGF-IR gene expression (Werner et al., 1996). Mdm-2 targets p53 for degradation: Mdm-2 mediated reduction of p-53 could thereby induce upregulation of the IGF-IR and increases the survival of cancer cells (Girnita et al., 2000). Expression of the IGF-IR is also regulated by the src tyrosine kinase, PKB/Akt serine-threonine kinase and the PTEN protein phosphates. Constitutively active Akt or src- activated Akt upregulates IGF-IR gene expression, whereas PTEN counteracts this β effect in pancreatic cancer cells and render the cell more invasive (Tanno et al., 2001). Neuronuclear factor (NF) -κβ is a transcription factor that can function in both cytokine signaling and in cell survival. NF-κβ mediates antiapoptotic effect of IGF-I in colon cancer cells (Garrouste et al., 2002), whereas it can mediate proapoptotic effects under other circumstances such as its role in the effects of tumor necrosis factor-α (Cheshire and Baldwin, 1997). Thus, the specific cellular response to NF-κβ depends on the original stimulus. Migration of epithelial colonic cells is dependent on IGF-IR induced alteration in integrins and cell adhesion complexes. While IGF-IR activation did not alter integrin expression levels, most of the integrin relocalized to the leading edge of migrating cells in response to IGF-I stimulation. Blocking integrin function with specific antibodies inhibited IGF-I induced migration. Furthermore, activation of the IGF-IR disrupts the E-cadherin/catenin complex, which associated with the cytoskeleton (Andre et al., 1999). Similarly, in MCF-7 breast cancer cells, the IGF-IR was showed to directly interact with the cells adhesion complex comprised of E-cadherin, β-catenin and p120 catenin. When IGF-IR antisense mRNA was expressed in MCF-7 cells, the cells exhibited a more malignant phenotype that was associated with a reduction in cell-cell adehision complex. This reduction was proposed to arise from a p120 catenin-induced decrease in E-cadherin and activation of Rac and Cad 42 activity (Pennisi et al., 2002). Certain tumor cells exhibit growth factor dependence in early, during the progression of tumorigenesis. During later stages, such cells may become growth factor, which independent for continued progression. For example, early stage melanoma cells have recently been shown to be exquisitely sensitive to IGF-I. At these early stages, IGF-I activates the MAP kinase pathway, which triggers proliferation and the PI kinase pathway, which promotes cell survival and stabilization of β-catenin. At the later stages of development i.e., in malignant melanoma cells, Erk 1 and Erk 2 were constitutively activated and β-catenin become more stabilized; IGF-I was unable to further activate these system (Satyamoorthy et al., 2001). Cross talk between receptors and their signaling pathway has been recently shown to play a critical role in various cellular responses to ligands. Such cross talk may occur between receptors within the same family, such as the Epidermal Growth Factor (EGF) and IGF-I receptor, both are tyrosine kinase receptors (Gilmore et al., 2002) or between different families such as nuclear steroid receptors and the IGF-IR (Dupont et al., 2000) or G-protein coupled receptors and the IGF-IR (Dalle et al., 2001). For example, the GBM(R) glioblastoma cell line is insensitive to AG1478, an anti-EGF therapeutic agent that acts as a specific EGF- receptor tyrosine kinase inhibitor. GBM (R) cells exhibited compensatory upregulation of IGF-IR levels in response to AG1478 treatment. This resulted in persistent signaling through the PI3 kinase pathway and was associated with an antiapoptotic and proinvasive phenotype. Both Akt1 and p70S6K appeared to play a role in this process (Chakravarti et al., 2002). In another example, the IGF-IR also protects mammary epithelial cells from apoptosis. Activation of this IGF-IR induces serine phosphorylation of BAD in this cell type, but this is mediated via transactivation of the IGF receptor, as this effect was blocked by ZD1839, a specific EGFR tyrosine kinase antagonist (Gilmore et al., 2002). Motility is an important process that plays a role in the spread of cancer cells. Activation of the IGF-IR and subsequent activation of the PI3 kinase pathway induces extension of lamellipodia in neuroblastoma cell lines (Meyer et al., 2001). Migration of melanoma cells is also stimulated by IGF-I. This effect is mediated by upregulation of interleukin-8 gene suppression via IGF-I induced activation of MAP kinase and phosphorylation of c-Jan (Satyamoothy et al., 2002). Various strategies have been used to block the IGF-IR in order to prevent tumor cell growth and to increase apoptosis of malignant cells. Expression of a dominant- negative truncated IGF-IR in colon cancer cells reduces the level of vascular endothelial growth factor expression, impaired tumor progression in nude mice and increase tumor cell apoptosis (Reinmuth et al., 2002). Scotlandi et al. (2002a, b) overexpresses a dominant negative form of this IGF-IR with a mutated ATP-binding site in enhanced apoptosis, decreased tumorigenesis and increased sensitivity to chemotherapeutic agents. Other techniques that have been used to inactivate the IGF-IR include expression of truncated soluble receptors to prevent ligand-receptor interaction (Pietrzkowski et al., 1993) and expression of peptides that could interfere with these interaction (D’Ambrosio et al., 1996). Perhaps the most exciting potential therapeutic modalities will arise from the recent crystallographic studies of the tyrosine kinase domain of the IGF-IR (Faveylyukis et al., 2001; Munshi et al., 2002; Pautsch et al., 2001). The production of small molecule that can act as a specific antagonist for IGF-IR and it also inhibit its anti-apoptotic effects (Katia et al., 2005; De Meyts and Whittaker, 2002).
IGF-IR Structure-function Studies
The extensive mutational analysis of IGF-IR identified that the receptor domains required for initiation of specific function, that is, proliferation (measured as cell growth in monolayer), survival (usually measured as the ability of cell to survive under anchorage-independent conditions) and transformation (assessed as the ability to grow in soft agar or to form foci). The experiments by using different cell models unequivocally demonstrated the mutation in the ATP binding site of the IGF-IR tyrosine kinase domain produced ‘dead’ receptors, incapable of signal transmission. Mutations at other residues of the tyrosine kinase impair partially the IGF-IR. For instance, the substitutions of tyrosine (tyr1131, 1135 and 1136) into phenylalanine abrogate signal transformation and mitogenesis, but not survival signaling. Mutations in either Tyr1131 or Tyr1135 downregulate transformation without reducing cell growth. Tyr 950 in the IGF-IR juxtamembrane domain was found necessary for IRS and SHC binding and for induction of mitogenic and transforming activity, but the IGF-IR Tyr-950 mutant still transmitted in addition to the classic IRS-1 dependent PI-3k (Akt pathway). Other survival pathway(s) emanate from IGF-IR (Hongo et al., 1996; O’Connor et al., 2000; Romano et al., 1999). Deletion of the entire C-terminus at aa1229 totally abrogated transforming function, without inhibiting mitogenic and antiapoptotic ability (Surmacz et al., 1995). The ‘transforming domain’ was mapped between residues 1245 and 1310 with Tyr1251, Ser 1280-1283, His 1293 and Lys 1294 that are required for transformation. It is worth nothing that C-terminal deletions (at residue 1229 or 1245) appeared to amplify antiapoptotic effects and intrinsic inhibitor of IGF-IR survival signaling. The mutation in the C-terminus at Tyr 1250, 1251, His 1293 and Lys 1294 reduce survival, implying that these residues act as neutralizers of the C-terminus proapoptotic function (Hongo et al., 1996). The important practical implication of the above studies is that transformation by IGF-IR does not occur without activated IGF-I survival pathway. Thus targeting the survival function of IGF-IR, which should be the optimal approach to inhibit tumorigenicity, as evidenced by the mutational analysis, the best way to achieve this effects is to inactivate totally the IGF-IR tyrosine kinase.
Role of Circulating IGF in Cancer
Prostate Cancer
The potent mitogenic activity of IGF-I in cell culture made it an obvious
candidate risk factor in cancer development, but it was not until 1998 that
several prospective studies suggested that high circulating levels of IGF-I
were associated with an increased risk of developing prostate cancer (Cohen,
1998; Wolk et al., 1998). A significant amount of data had been accumulated
suggest that the IGF system plays an important role in the prostate. Prostatic
stromal cells and epithelial cells in primary culture secrete IGFBPs and stromal
cells, which produce IGF-II and both stromal and epithelial cells express the
IGF-I receptor, which are responsive to IGF-I with respect to proliferation
(Cohen et al., 1994a-c). In vivo, it is likely that the prostate
epithelial cells that are precursors to prostatic intraepithelial neoplasia
and prostatic adenocarcinoma, respond to both locally produced IGF-II and circulating
IGF-I through paracrine and endocrine mechanisms, respectively. Further support
for the role of IGF-action in prostate growth has come from recent reports that
systemic administration of IGF-I increases rat prostate growth (Torring et
al., 1997), that modulate the ventral prostate weight by finasteride, is
associated with altered levels of IGF-I receptors and IGFBPs gene expression
(Huynh et al., 1998). The IGF-I deficient mice exhibit decreased prostate
size and complexity of prostate structure (Ruan et al., 1999). The validity
of the association between IGF-I levels and prostate cancer risk was questioned
by subsequent cross sectional studies (Cutting et al., 1999; Djavan et
al., 1999), in a prospective study, it was found that the IGF-I/PSA ratio
was superior to IGF-I or PSA measurement alone for predicting prostate cancer
risk. Finne et al. (2000), in a screening trial, did not find an association
between serum IGF-I levels and prostate cancer risk, while Baffa et al.
(2000) actually found that circulating IGF-I levels were lower in a group of
patients undergoing radical prostatactomy as compared to age-matched controls.
In prospective studies, however, Harman et al. (2000) and Stattin et
al. (2000) found that IGF-I levels were associated with prostate cancer
risk and this association was especially evident in younger men. While the conclusion
of this extensive series of studies conducted over the last 4 years appears
contradictory, there is, in fact, some consistency. Prospective studies consistently
demonstrated an association between high circulating IGF-I levels and prostate
cancer risk, while cross-sectional studies consistent have generated variable
results. These data are consistent with the hypothesis that high serum IGF-I
levels in younger men predict the occurrence of advanced prostate cancer. Years
later, while IGF-I levels at the time of diagnosis are not especially informative.
This hypothesis suggests that long–term exposure of prostate epithelial
cells leads to high levels of serum-derived IGF-I, which increases the probability
of initiating hyperplasia in the cellular precursors of prostatic intraepithelial
neoplasia and subsequent prostate adenocarcinoma. Molecular corporation of the
relationship between IGF-I levels and prostate carcinogenesis has now come from
the analysis of transgenic mice with targeted expression of IGF-I in the basal
prostatic epithelium. This dysregulated IGF-I biosynthesis regulated in the
appearance of hyperplastic lesions resembling PIN by 6 months of age and prostatic
adenocarcinomas or small cell carcinomas were eventually seen in 50% of the
transgenics. Specifically, desregulated expression of IGF-I and constitutive
activation of IGF-I receptor in basal epithelial cells resulted in tumor progression
similar to that seen in human disease. These studies also provide additional
evidence for the prostate basal epithelial cell as a precursor to prostate cancer.
Breast and Other Cancer
Hankinson et al. (1998) reported that premenopausal, but not post-menopausal
women in the highest fertile of serum IGF-I levels had a significantly increased
risk of developing breast cancer. These findings have been generally supported
by most (Vadgama et al., 1999; Jernstrom and Barret-Connor, 1999) but
not all subsequent studies. Racial factors may play a role in the IGF-I breast
cancer association, in the Agar-Collins et al. (2000) found that high
serum IGF-I levels were strongly associated with breast cancer risk in post
menopausal African-American women. With respect to the colorectal cancer, Ma
et al. (1999) and Palmqvist et al. (2002) have reported positive
association between serum IGF-I and colorectal cancer risk in US, Greek and
Swedish cohorts, while Probst-Hansch et al. (2001) found an association
between IGF-I or IGFBP-3 levels and colorectal cancer risk in a Chinese cohorts.
The role of IGF-II is also unclear, being positively associated in the Greek
and Chinese studies, but not in the US cohorts. Yu et al. (1999) reported
a positive association between high IGF-I and low IGFBP-3 levels (but not IGF-II)
and lung cancer risk. Lukanova et al. (2001), however found no correlation
between serum IGF-I or IGFBPs in a large female cohort. Collectively, these
studies continue to suggest a role for IGF-I at a risk factor for breast, colorectal
and lung cancer, but its utility as a pragmatic marker is potentially limited
by ethnic and (for colorectal and lung) gender factors.
In breast cancer cells, estrogens enhance the mitogenic effect of IGF-I, induced expression of IGF-I and stimulate production of IGF-IR (Herbert and Thomas, 2000). Estrogens also repress synthesis of some IGFBPs in breast tissue. In breast cancer cells, estrogens decrease the expression of IGF-IIR and increase the level of IGFBP proteases (Clarke et al., 1997). The interaction between estrogens and IGF is reciprocal. IGF-I enhances expression of estrogen receptor (ER) in breast cancer cells and ER levels in breast tissue are associated with the levels of some IGFBPs (Mathieu et al., 1991).
Role of the IGF-IR in Human Cancer
Numerous studies performed over the last 20 year have suggested that transformed cells express the IGF-IR at higher levels than normal cells. However, the molecular mechanism which IGF-IR gene increases expression in tumors remains largely unidentified. Amplification of the IGF-IR locus at band 15q26 has been reported in a small number of breast cancer and melanoma cases (Almeida et al., 2001). During tumorigenesis, overexpression of the IGF-IR is presumed to increases the cellular responsiveness to the IGFs, in terms of proliferation and inhibition of apoptosis. This picture is probably most accurate with respect to the pediatric tumors associated with chromosomal translocation such as Wilm’s tumor and rhabdomyosarcoma. However, the role of the IGF-IR in the progression of epithelial tumors that are most prevalent in adults is likely to be more complex (De Pinho, 2000). It has been suggested that the IGF-IR itself can function as an oncogene, based upon the phenotype of fibroblasts overexpressing the IGF-IR (Kaleko et al., 1990). However, the relevance of this system to human cancer in general is unclear. Other studies have used IGF-IR overexpression in fibroblasts to show that the IGF-IR can modulate radiosensitivity (Turner et al., 1997). Nevertheless, it should be noted that a recent report demonstrated that inhibition of IGF-IR activity by a selective kinase inhibitor in MCF-7 breast cancer cells increases radiosensitivity (Wen et al., 2001). Many studies describing overexpression of the IGF-IR in breast, prostate and other tumor cells have been largely based on analysis of tissue homogenates or established cancer cell lines. Unfortunately, there are no appropriate normal controls for these samples that can be used for comparison. This apparent IGF-IR content of tissue homogenates, in particular, can be affected by contamination with stroma, which would dilute the IGF-IR content in normal epithelium or small tumors. More focused studies of IGF-IR expression in breast and prostate that employed immunohisochemistry or matched cell lines corresponding to normal epithelium and early stage tumors both express abundant levels of the IGF-IR and that IGF-IR expression is significantly reduced in advanced, metastatic cancer (Tennant et al., 1996; Happerfield et al., 1999; Chott et al., 1999; Schnarr et al., 2002; Damon et al., 2001). A recent report Hellawell et al. (2002) challenged this view, reporting that IGF-IR expression was decreased in certain metastatic prostatic cancer. Sample, as compared with benign or carcinoma tissue, but that IGF-IR expression was increased in the majority of samples studied (eight out of 12). However, it should be noted that IGF-IR immunostaining with a single β-subunit antibody was diffusely cytoplasmic in most samples, in contrast to the expected membrane localization reported by Chott et al. (l999) who used two different subunit antibodies. Thus, the levels of IGF-IR expression in the progression of prostate cancer have not been clearly established. Activation of the IGF-IR present in normal epithelium in response to elevated levels of circulating IGF-I may underlie the epidemiological data described above. In contrast, the subsequent decrease in IGF-IR (if this is an attempt by established cancers cells to counteract the potential differentiating effects of IGF-I at the sites of metastasis. Alternatively, decreased expression of the IGF-IR may protect tumor cells from a novel, non-apoptotic form of programmed cell death that has been recently described as being triggered by the unliganded IGF-IR (Sperandio et al., 2002).
The IGF Receptor as Anticancer Treatment Target
Two factors underpin the concept of the IGF-IR as anticancer treatment target, relating to the function of the IGF-IR and its pattern of expression. The IGF-IR is expressed on the surface of most normal cells, but it is frequently overexpressed by tumors, including melanoma and cancers of the colon, pancreas and prostate (Hellawell et al., 2002; Bohula et al., 2003).
IGF-IR Inhibitors
IGF signaling can be inhibited by blocking either the expression or function of the IGF-IR.
Inhibitors of IGF-IR Function
Small Molecule Kinase Inhibitors
Chemical inhibitors have many advantages; they can be designed for solubility
and stability and they can often be administered orally with high bioavailability
(Dreves et al., 2003). Specificity is a major design hurdle for development
of IGF-IR inhibitors, given the high degree of homology with the insulin receptor.
Recent structural studies reveal regions of divergence within the IGF-IR and
insulin receptor kinase domains, suggesting that it may be possible to design
specific IGF-IR inhibitors (Garcia-Echeverria et al., 2003; De Meyts
and Whittaker, 2002). High throughput technology combined with computer modeling
is currently used to identify low molecular weight compounds blocking the IGF-IR
tyrosine kinase. The first described IGF-IR inhibitors, tyrphostin AG538 and
I-OmeAG, were modeled on the IR tyrosine kinase. The compounds inactivated the
IGF-IR tyrosine kinase by blocking the substrate-binding site; however, cross
reactivity with the IR tyrosine kinase was reported (Blum et al., 2000).
Recent advances in the characterization of the three dimensional structures
of IGF-IR and IR greatly facilitate the design of specific IGF-IR inhibitors
I (De Meyts and Whittaker, 2002). Most importantly, crystallographic studies
revealed conformational differences in the phosphorylate forms of IGF-IR and
IR kinases, the factors allowing the development of selective therapeutics (Faveylyukis
et al., 2001; Pautsch et al., 2001). Several new compounds with
enhanced specificity towards IGF-IR and low cross reactivity with IR entered
into preclinical studies. The examples include derivatives of pyrimidine and
podophyllotoxin, disclosed in patent application WO 02/092599 and WO 02/102804,
respectively, specific small inhibitors of IGF-IR are likely candidates to become
anti-IGF-IR drugs. The positive experience with similar therapeutics, especially
the possibility of oral delivery and low toxicity, makes this approach especially
attractive.
The therapeutic potential of a novel kinase inhibitor of IGF-IR, NVP-AEW541, in Ewing's sarcoma, osteosarcoma and rhabdomyosarcoma, the three most frequent solid tumors in children and adolescents. NVP-AEW541 may be combined with vincristine, actinomycin D and ifosfamide, three major drugs in the treatment of sarcomas.
Blocking Antibodies
Monoclonal antibodies to the IGF-IR have been shown to inhibit the growth
of a range of tumors in vivo. Efficacy can be limited by poor penetration
and antimouse immune responses, although these problems may be avoided by use
of single chain and for humanized antibodies (Ludwig et al., 2003). The
initial approach to inhibit IGF-IR signaling was based on the use of IGF-IR
blocking antibodies. The mouse Mab alpha-1R3 raised against the alpha domain
of IGF-IR (Jacobs et al., 1986) inhibited IGF-IR activation and IGF-IR-dependent
mitogenicity in several cell types in vitro, including breast carcinoma
(Artega et al., 1989; Artega, 1992), rhabdomyosarcoma (Kalebic et
al., 1994), NSCLC (Zia et al., 1996) and Ewing’s sarcoma (Scotlandi
et al., 1998). However, in some cases alpha-1R-3 was ineffective in blocking
IGF-1 sensitive tumors in animal models (Artega, 1992). Furthermore, it has
been reported that alpha-1R-3 may exhibit agonistic abilities towards IGF-IR
(De Leon et al., 1992; Kato et al., 1993). The mouse anti-IGF-IR
Mab 391 inhibited IGF-IR autophosphorylation and signaling to Akt in several
human cancer cell lines. Chronic treatment with Mab 391 resulted in down regulation
of receptor through lysosome-dependent pathways (Hailey et al., 2002).
Several other mouse anti-IGF-IR Mab were described (Li et al., 1993,
2000). One of them Mab 1H7, which blocks IGF-IR/IGF-I binding and IGF-IR dependent
DNA synthesis was used to engineer a single chain humanized anti-IGF-IR sc Fv-Fc
Ab and contains the Fc domain of human IGF-1 fused to the Fv region of 1H7 (Li
et al., 2000). Treatment of MCF-7 breast cancer cells with Sc Fv-Fc for
2-24 h downregulated the levels of IGF-IR through the lysosomal endocytic pathways,
rendering the cells refractory to IGF-I stimulation (Sachdev et al.,
2003). Importantly, down regulation of IGF-IR by scFv-Fc occurred also in MCF-7
xenografts and was paralleled by reduced tumor growth (Sachdev et al.,
2003). These are similar humanized Mabs will likely become a model for future
drug development. Once their specificity towards IGF-IR and lack of IR-cross
reactivity is demonstrated in vivo.
An antagonistic monoclonal antibody, designated EM164, inhibits the proliferation and survival functions of the IGF receptor in cancer cells. EM164 was initially selected by a rapid cell-based screen of hybridoma supernatants to identify antibodies that bind to IGF-IR but not to the homologous insulin receptor and that show maximal inhibition of IGF-I-stimulated autophosphorylation of IGF-IR. However, the inhibition of MCF-7 cell growth in vitro by EM164 can be attributed principally to a cytostatic effect, where EM164 treatment causes cells to accumulate in the G0-G1 state of the cell cycle (Erin et al., 2003).
Dominant Negative Protein
IGF-IR dominant negatives have been constructed as protein transacted within
the β-subunit, capable of forming inactive heterodimers of mutant and wild
type receptors unable to transduce downstream signals. Soluble IGF-IR dominant
negative lack the transmembrame region and compete with wild type receptors
for ligand binding. This strategy has been shown to suppress IGF-IR function,
resulting in inhibition of growth and tumorigenicity (Dreves et al.,
2003).
IGF-I Mimetic Peptide
A series of small IGF-I peptide analogues was designed by molecular modeling
of the IGF-I protein (Pietrzkowski et al., 1992, 1993) to compete with
IGF-IR ligands. The synthetic peptides were modeled on domains of IGF-I, as
these domains contain the least similarity between IGF-I and insulin. One of
the peptides JB-1 (modeled on the domain) effectively inhibit IGF-I dependent
IGF-IR autophosphorylation and proliferation in several tumor cell lines. The
analogues used at nano or micro molar concentrations exhibited good specificity
for IGF-IR and low toxicity for cells in cell culture (Pietrzkowski et al.,
1992, 1993) compounds against experimental tumors in vivo has never been
assessed.
Inhibitors of IGF-IR Expression
Because of difficulties in designing specific small molecule IGF-IR kinase inhibitors other molecular approach have used to block IGF-IR expression. Antisense is the best characterized of these, but this is now being surprised by the recent demonstration that profound gene silencing can be induced in mammalian cells by small interfering RNAs (Elbashir et al., 2001).
Antisense nucleotides, Antisense RNA, siRNA, Triple helix
A variety of experiments employing antisense oligodeoxynucleotides (ODNs), antisense RNA and small interfering RNA (siRNA) demonstrated that IGF-IR-dependent tumorigenicity can be decreased or eliminated by blocking IGF-IR mRNA, thus inhibiting IGF-IR protein synthesis. Most of the reported anti-IGF-IR ODNs contained sequences complementary to the IGF-IR translation initiation site. The association of these reagents with IGF-IR mRNA produced heteroduplex that was cleaved by RNase. Multiple studies documented that anti IGF-IR ODNs (regular or phosphorotioate chemistry) at nanomolar concentrations decreased IGF-IR expression reduced cell proliferation and rodent cancer cell type in cell grown in culture (Pietrzkowski et al., 1993; Dreves et al., 2003; Resnicoff et al., 1994, 1995a, b; Muller et al., 1998; Coppola et al., 1999; Macauley et al., 2001; Pavelic et al., 2002). Furthermore, in some instances, treatment with ODNs in induced massive apoptosis and tumor regression in animal models was reported (Resnicoff et al., 1995a, b). However the reduction of IGF-IR expression was often incomplete even with high concentrations (100-500 nM) of target sequence (Macauley et al., 2001). Moreover, interactions of anti-IGF-IR ODNs with IR synthesis were reported (Bohula et al., 2003). To address these problem, Bohula et al. (2003) used scanning oligonucleotide array to probe the secondary structure of IGF-IR mRNA in order to identify target sequences that are accessible for ODNs and do not appear in IR mRNA. This strategy enabled selection of specific ODNs that effectively and selectively downregulated IGF-IR in human cancer cell lines. Furthermore, the accessible sequence were suitable target for anti-IGF-IR siRNAs. Indeed, some of the designed siRNAs were able to silence paralleled by repression of IGF-IR signaling (Bohula et al., 2003). In addition to ODNs and siRNA, different antisense IGF-IR RNA vectors containing fragments of IGF-IR cDNA cloned in 3’5’ orientation were generated to inhibit IGF-IR expression. The vectors (plasmids or viruses) produced antisense RNA that hybridized with complementary sequences in IGF-IR mRNA, blocking IGF-IR synthesis. For instance, an antisense IGF-IR RNA against the first 309 bases of IGF-IR mRNA, delivered to cells by transfection or adenoviral infection, reduced IGF-IR expression. IGF-I dependent proliferation and survival in a number of human and rodent cell models, including endometrial cancer (Nakamura et al., 2000), Ewing’s sarcoma (Scotlandi et al., 2002a, b) and rat glioblastoma (Resnicoff et al., 1994). The expression of this IGF-IR antisense efficiently inhibit tumorigenicity of cells grown as explants in experimental animals, most probably by induction of massive apoptosis (Resnicoff et al., 1994, 1995b; Scotlandi et al., 2002a, b). The same antisense construct inhibit metastasis of murine lung carcinoma cells (Brodt et al., 2000). Another antisense IGF-IR plasmid containing ~300bp DNA complementary to the region surrounding the IGF-IR translation initiation site was used to inhibit IGF-IR expression and function in breast cancer cells (Neuenschwander et al., 1995). In many cases, the induction of cell death with antisense IGF-IR strategies was much more pronounced in vivo (animal models) than in vitro (monolayer tissues culture or soft agar), suggesting that in vivo tests may be superior in screening for anti-IGF-IR compounds (Resnicoff et al., 1994, 1995b; Scotlandi et al., 2002a, b). Oligonucleotide-directed triple helix formation is an approach to block transcription of specific genes by inhibiting the passage of RNA polymerase along with target DNA. The third effector strands (oligopyrimidine) contains oligopurine and/or oligopyrimidine sequenced in target DNA. The triple helix strategy has been reported to be effective in down regulation of IGF-IR. Specifically, a plasmid encoding the homopurine RNA sequences designed to form a triplex with a homopurine and homopyridine sequence present in 3’ to the termination codon of the IGF-IR gene suppressed IGF-IR transcription in rat (6 glioblastoma cells). The triple helix reagent induced dramatic reduction of IGF-IR transcripts and IGF-IR expression can inhibit tumor formation in nude mice (Rininsland et al., 1997). Interestingly, in some case of rat C6 glioblastoma and some other cellular models, down regulation of IGF-IR by antisense approaches were associated with the induction of an immune host response leading to elimination of untreated established tumors (Resnicoff et al., 1994). This peculiar effects perhaps related to the induction of immune response by the presence of apoptotic cells (Trojan et al., 2000) was further explored in pilot studies involving patients with astrocytomas treated with autologous glioma cells exposed to anti-IGF-IR ODNs (Andrews et al., 2001).
Modulators of IGF-IR Internalization and Recycling
Following ligand binding, the IGF-IR ligand complex is internalized and the ligand is degraded by endosomal proteinase and the receptor is returned to the membrane, one way to reduce IGF-I effects is to block IGF-IR re-expression on the cell surface. Recent studies suggested that IGF-IR trafficking could be substantially blocked by the inhibition of IGF-I degrading enzymes for example cathepsin. The cathepsin inhibitors, E-64 and CA074 methyl ester, reduced IGF-IR expression on the cell surface and impaired several IGF-I dependent effects, including DNA synthesis and synthesis of matrix metalloproteinase in human breast cancer and murine lung carcinoma cells (Brodt et al., 2000).
Perspective for Anti-IGF-IR Pharmaceuticals
IGF-IR is a promising target in cancer therapy because (1) IGF-IR expression is easily measurable by conventional techniques. (2) Tumor cells may be more sensitive to targeting IGF-IR than normal cells and (3) IGF-IR is often required for the tumorigenic effects of other oncogenic agents. Thus targeting IGF-IR can be combined with other therapies. Unlike with HER2 and EGFR, the development of anti-IGF-IR pharmaceutical is still in early discovery phases. Similarly to HER2 and EGFR, however, the most advanced strategies are those involving small inhibitors of the IGF-IR tyrosine kinase and anti-IGF-IR antibodies. Other approaches such as siRNA, antisense and triple helix strategies are also promising, but they will require optimization of specificity in vitro and efficient and safe delivery system.
Summary
In summary, the IGF signaling system plays a central role in many aspects of tumorigenesis. A better understanding of this complex system will facilitate the development of novel approaches to diagnose and treat various human cancers.