International Journal of Colorectal Disease

, Volume 20, Issue 3, pp 203–220

The role of the insulin-like growth factor system in colorectal cancer: review of current knowledge

Authors

  • Rajaraman Durai
    • University Department of Surgery, Royal Free and University College Medical SchoolUniversity College London
  • Wenxuan Yang
    • University Department of Surgery, Royal Free and University College Medical SchoolUniversity College London
  • Sharmila Gupta
    • Royal Free Hampstead NHS Trust Hospital
    • University Department of Surgery, Royal Free and University College Medical SchoolUniversity College London
  • Marc C. Winslet
    • University Department of Surgery, Royal Free and University College Medical SchoolUniversity College London
    • Royal Free Hampstead NHS Trust Hospital
Review

DOI: 10.1007/s00384-004-0675-4

Cite this article as:
Durai, R., Yang, W., Gupta, S. et al. Int J Colorectal Dis (2005) 20: 203. doi:10.1007/s00384-004-0675-4

Abstract

Background

The insulin-like growth factor system, which includes insulin-like growth factors (IGF-I and IGF-II), IGF receptors (IGF-IR and IGF-IIR) and IGF binding proteins (IGFBPs), plays an important role in epithelial growth, anti-apoptosis and mitogenesis. There is a growing body of evidence showing that IGFs control growth and proliferation of several types of cancer. This review introduces the latest information on the biology of the IGF system and its pathophysiological role in the development of colorectal cancer.

Discussion

The growth promoting effects of IGF-I and IGF-II on cancer cells are mediated through the IGF-IR, which is a tyrosine kinase and cancer cells with a strong tendency to metastasise have a higher expression of the IGF-IR. Most of the IGFs in circulation are bound to the IGFBPs, which regulate the bioavailability of the IGFs. All IGFBPs inhibit IGF action by high affinity binding, while some of them also potentiate the effects of IGFs. Colon cancer cells produce specific proteases that degrade the IGFBP so that the IGF will be free to act on the cancer cell in an autocrine manner. Therefore, the IGFBPs play a crucial role in the development of the cancer.

Conclusion

The current knowledge about the link between IGFs and colon cancer is mainly based on in vitro investigations. Further in vivo study is needed to understand the exact role of the IGF system, especially its binding proteins, so that they can be manipulated for the prevention and treatment of colorectal cancer.

Keywords

Insulin-like growth factorsInsulin-like growth factor binding proteinsColorectal cancer

Introduction

The incidence of colorectal cancer has been increasing rapidly since 1975 [1] with about 300,000 new cases and 200,000 deaths in Europe and the USA each year [2, 3]. It is the second commonest cause of death from any cancer in men in the European Union [4]. Long-term survival of colorectal cancer is related to the stage of disease. If detected early, it may be curable by surgery. But once metastases develop the prognosis becomes poor. At least 40% of patients with colorectal cancer develop metastases during their illness [5]. Though various combinations of surgery, radiotherapy and chemotherapy are used for advanced cancer the most effective approach is yet to be discovered.

Several growth factors including transforming growth factor-β, epidermal growth factor, insulin and insulin-like growth factors (IGFs) have been implicated in the development of colon cancer. In the last decade the role of the IGFs in cancer has been increasingly recognised. The IGFs are peptides structurally related to insulin that regulate the proliferation of several mammalian cells, including epithelial cells of the breast, prostate, colon and lung [6]. Abnormal expression of the IGFs, their receptors and binding proteins has been linked to several cancers, including colorectal cancer. This review introduces the latest information on the biology of the IGF system and its pathophysiological role in the development of colorectal cancer. In particular, this review highlights the IGF-related properties of colon cancer cells, the role of individual IGF binding proteins in colon cancer and their potential applications in the treatment of colorectal cancer.

Biology of the IGF system

IGF-I and IGF-II

The IGF system includes ligands IGF-I and -II, their receptors, IGF-binding proteins (IGFBPs) and IGF proteases. Figure 1 is a schematic representation of the IGF system. Their biological features and actions are summarised in Table 1. The IGFs are single-chain polypeptides with structural homology to pro-insulin. The IGF-I is a basic peptide with 70 amino acids and the IGF-II is slightly acidic with 67 amino acids. IGF-I and IGF-II act on a variety of mammalian cells in an endocrine, paracrine and autocrine manner [7, 8‘] to regulate cell proliferation, apoptosis, transformation and differentiation [9, 10]. Both IGF-I and -II are essential for normal human growth and development [11]. Growth hormone regulates the expression of the IGF-I gene and it thereby mediates most of its actions through IGF-I, but it does not control the expression of IGF-II [12]. Adults have a higher concentration of IGF-II than IGF-I, but it is the IGF-I that plays an important role in postnatal growth, while IGF-II acts mainly in embryonic and foetal life [12]. The majority of circulating IGFs, in particular IGF-I, are produced by the liver, although various tissues have the capability to synthesise these peptides locally [10, 13]. More than 90% of the IGFs are bound to IGFBP-3 [7, 14] and only 1% of IGF circulates in the free form.
Fig. 1

Schematic representation of the insulin-like growth factor (IGF) system.

Table 1

Insulin-like growth factors (IGFs), their receptors and proteases [13, 106]. IGFBP insulin-like growth factor binding proteins

Components

Nature

Location

Role

IGF-I

70 residue single chain polypeptide

Various body fluids and cells

Growth, anti-apoptosis, mitogenesis and differentiation

IGF-II

67 residue single chain polypeptide

Various body fluids and cells

Foetal and embryonic growth

IGF-IR

Heterotetrameric 2(α)+2(β) glycoprotein-tyrosine kinase

Cell membrane

Mediates the effect of various growth factors including IGF-I and -II and steroids. It also induces VEG- mediated angiogenesis

IGF-IIR

Monomeric glycoprotein

Cell membrane

It binds to IGF-II and mannose-6 phosphate and inhibits action of IGF-II

Hybrid receptor

One alpha + one beta unit with one hemi insulin receptor

Cell membrane

It binds IGFs

IGFBP proteases

Several types (serine protease, metalloproteinases, cathepsins and specific proteinases)

Various body fluids and cells including smooth muscle and fibroblast

It causes proteolysis of IGFBP to free IGF

IGF receptors

The actions of IGFs are mediated via the type I and type II receptors (IGF-IR and IGF-IIR), both of them glycoprotein, found in the cell membrane. The growth-promoting effects of IGFs are mediated through the IGF-IR, which is a tyrosine kinase resembling Insulin receptor [7]. The IGF-IR has two extracellular subunits (α) to bind the IGFs and two intracellular subunits (β) that contain tyrosine kinase [15] to mediate the intracellular effects. In contrast to the insulin receptor, which primarily mediates metabolic function, the IGF-IR mediates mainly growth and differentiation [16]. When activated the IGF-IR stimulates the synthesis of RNA and DNA, cell proliferation, differentiation and increases cell survival [15]. The IGF-IIR is similar to mannose-6 phosphate receptor. It is a single chain polypeptide with a large extracellular domain, a single transmembrane region and a small cytoplasmic domain. It is entirely different from the IGF-IR in that it has no tyrosine kinase activity. The IGF-IIR is thought to function as a clearance receptor for IGF-II, thereby influencing the extracellular levels of the IGF-II [17]. The IGFs can also act via a hybrid receptor, which consists of a single α and β subunit linked by disulfide bonds. These hybrid receptors have half insulin hemi receptor and a dimerised IGF-IR. The IGF-IR/hybrid receptors retain the high affinity for IGF-I, but exhibit a dramatically decreased affinity for insulin [17].

IGF-binding proteins and IGF proteases

The effects of the IGFs are modulated by at least ten different IGF binding proteins [18, 19], six of which have a high affinity for IGFs (IGFBP1–6) and four of them have a low affinity for IGF [18, 20], also known as IGFBP-related proteins (IGFBP-rp1–4). Table 2 summarises the facts related to IGFBPs 1–7.
Table 2

IGF binding proteins [11, 106, 107]

IGFBPs

Nature (kDa protein)

Locationa

Roleb

IGFBP-1

25–34

Various body fluids including amniotic fluid

Modulates the IGF effects

IGFBP-2

32–34

Various body fluids including CSF

Mediates TNF-α induced apoptosis and stimulates cell proliferation

IGFBP-3

53

Various body fluids

Main carrier protein for IGFs in circulation and growth inhibitor. 90% of serum IGF is bound to IGFBP-3

IGFBP-4

24, 29

Various body fluids and tissues including liver, vascular smooth muscle

Inhibits the action of IGF

IGFBP-5

23

Various body fluids, osteoblast, kidney, etc.

Increases mitogenesis

IGFBP-6

30–32

Serum and CSF

Modulates the action of IGF

IGFBP-7

27

Various tissues including smooth muscle and endothelium

Modulates the action of IGF

All IGFBPs can act as a carrier for IGFs

aAll IGFBPs are found in serum

bAll IGFBPs except IGFBP-4 and -6 have both IGF-potentiating as well as -inhibiting actions under different circumstances

The IGFBPs are a family of homologous proteins that are produced by many different tissues and all have different molecular weight, amino acid composition, binding properties and distribution in biological fluids [19], but they all share cysteine-rich N terminal domain and C-terminal [18]. They also have IGF-independent actions [18]. The IGFBPs control the distribution of IGFs because of their higher affinity to IGFs, which is 2–50 times higher than that of the IGF-IR [21]. IGFBP-1–5 preferentially bind the IGF-I over the IGF-II, but IGFBP-6 has 100-fold higher affinities for IGF-II than for IGF-I [7]. The IGFBPs do not bind to the insulin [17, 19]. The key function of the IGFBPs is regulation of the circulating bioavailability of the IGFs [22]. The IGFBPs serve as carriers, mediators, as well as reservoirs of IGFs by protecting IGFs from degradation and delivering them to appropriate tissue [20, 23]. IGFBP-3 is the main binding protein, which is found in circulation in abundance, and most of the IGFs are bound to it. IGFBP-3 forms a ternary complex with IGF-I and a separate protein namely 80-kDa acid-labile subunit [6] to keep the IGF-I within the vascular system [24], while the other IGFBPs form binary complexes with IGF-I allowing transport of IGFs across the capillary to various tissues [24].

IGFBPs-1–6 inhibit the action of IGFs by high affinity binding, while IGFBP-1, -3 and -5 also potentiate the IGF effects [22]. The potentiation needs a decrease in binding affinity of IGFBP for IGF so that the IGF can act on the IGF-IR. This process involves association of IGFBP with the cell surface or extracellular matrix, and/or postsynthetic processing of IGFBP, e.g. phosphorylation and proteolysis [25]. IGFBP-1, -3 and -5 have well-established effects that are independent of IGF-IR signalling. IGFBP-1 exerts these effects by signalling through [α] 5[β] 1-integrin, whereas IGFBP-3 and -5 may have specific cell-surface receptors with serine kinase activity [22].

The IGFBPs are degraded by several proteases, some of which are specific to that binding protein and others are non-specific proteases that control the function and the tissue availability of IGFBPs. These proteases cleave the IGFBPs into fragments with lower affinity for the IGFs. In the serum proteolysis IGFBP-3, which is the main binding protein, is negligible in the normal population, but it is increased in certain physiological conditions like pregnancy and pathological conditions like postoperative period and malignancies [26], which increase the availability of circulating free IGF. In tissues, IGFBP proteases enhance the IGF-I availability by cleaving IGFBP, thereby increasing free IGF-I concentration [7]. Experiments show that proteolytic activity is highest in proliferating cells [27]. The biological activity of IGF is determined by the integrated actions of circulating IGF-I and IGFBP and by local production of IGF, IGFBP and IGFBP protease.

Studies linking the IGF system and colon cancer

Tables 3 and 4 summarise the results of population-based, in vitro and in vivo studies linking the IGF system and colon cancer. Most of the current knowledge about the IGF system is based on in vitro and population-based studies. Most population-based studies measured a few factors in the serum and compared them with the incidence/development of colorectal adenoma and carcinoma. Some of the studies are larger and others are smaller. Not all of them clearly support a linear association between IGF-I, IGF-II, IGFBPs and colorectal cancer. It may be due to the fact that the relation between these factors may be a complex one. Some studies show no association between plasma IGFBP-1 and IGFBP-2 and the risk of colon and/or rectal cancer [28], while others show that chronically high levels of circulating insulin and IGFs associated with a western lifestyle may increase the colorectal cancer risk, possibly by decreasing IGFBP-1 and increasing the bioactivity of IGF-I [29]. Overall, the studies suggest that an increase in IGF-I and a decrease in IGFBP-3, leading to increased serum-free IGF-I, may be associated with the development of colorectal adenoma and carcinoma [30, 31]. Renehan et al. [32] performed a meta-regression analysis of case-control studies linking the relationship between IGF-I, IGFBP-3 and cancers. Their study showed that higher concentration of IGF-I was associated with colorectal cancer, but high concentration of IGFBP-3 did not protect from colorectal cancer. It is possible that the case-control studies might have overestimated the relationship between IGFs, IGFBP-3 and colorectal cancer.
Table 3

Clinical, epidemiological and population-based studies linking IGF, IGFBPs and colorectal cancer (CRC). NA not applicable

Reference

Study type

Population

Cases/control

Parameters

Results/conclusions

[30]

Nested case-control

9,345

282/282

Serum IGF-I and IGFBP-3

No significant relation between IGF-I, IGFBP-3 and CRC

[96]

Nested case-control

32,826

276/236

Plasma IGF-I and IGFBP-3

↑IGF-I and ↓IGFBP-3 are associated with ↑ risk of large or tubulovillous/villous colorectal adenoma and cancer

[29]

Case-control

14,275

102/200

C-peptide, IGF-I, and IGFBP-1, -2 and -3

CRC risk was not related to IGF-I but ↑ for highest quintile of IGFBP-3 (OR=2.46, 95% CI=1. 09–5.57)

[108]

Nested case-control

14,916

193/318

Plasma IGF-I, IGF-II and IGFBP-3

Circulating IGF-I and IGFBP-3 are related to future CRC but IGF-II is not related

[28]

Nested case-control

NA

168/336

Plasma IGFBP-1, IGFBP-2 and insulin

IGFBP-1 and IGFBP-2 showed no association while chronic hyperinsulinaemia is moderately associated with CRC risk

[109]

Cohort

18,244

135/661

Serum IGF-I, IGF-II and IGFBP-3

Serum IGF-I was not associated with risk of CRC but ↑ circulating IGF-II and IGFBP-3 may indicate impending CRC

[31]

Case-control

803

157/311

Plasma IGF-I and IGFBPs

Nonsignificant ↑ in the prevalence odds of colorectal adenomas for highest versus lowest quartile level of IGF-I

[86]

Nested case-control

14,275

102/200

Serum IGF-II

Possible ↑ in CRC risk with ↑IGF-II

[65]

Nested case-control

14,916

193/318

Plasma IGF-I and IGFBP-3

Dairy products cause a modest ↑ in IGF-I levels. Low fat milk lowers the risk of CRC particularly among individual with high IGF-I/IGFBP-3

[77]

Cross-sectional retrospective

NA

81/26

Serum IGF-I, IGF-II, IGFBP-2 and -3

IGFBP-2 may regulate the bioavailability of IGF, which modulates colonic cell proliferation and/ or differentiation

[110]

Prospective

442

53/47

Serum IGF-I and IGFBP-3

IGF-I and IGFBP-3 levels are related to future CRC risk and may predict adenoma progression

[92]

Comparative

NA

52/293

IGF-I, IGF-II, IGFBP-2 and -3

Serum IGFBP-2 may have an adjuvant role in CRC surveillance (↓ in IGFBP-2 level post-surgery)

[78]

Prospective

NA

66/0

Serum IGF-I

Development of new adenomas, but not hyperplastic polyps, was associated both with ↑ serum IGF-I (P<0.005)

[45]

Case-control

NA

41/50

IGF-I, IGF-II and IGFBP-3

Individuals with IGF-I and -II values in upper 2 tertiles had ↑ odds ratio for CRC (OR=5.2, 95% CI 1.0–26.8) compared with those in lower tertile

[111]

Comparative

NA

17/0

IGFs and IGFBPs

Simultaneous elevation of serum IGFBP-2 and -3 in CRC

[112]

Case-control

NA

60/306

Serum IGF-I, IGF-II IGFBP-2 and -3

IGF-II and IGFBP-2 ↑ in acromegalics but presence of colorectal neoplasia did not contribute to ↑ in serum IGF-II and IGFBP-2

[113]

Case-control

NA

92/57

Serum IGF-I, IGF-II, IGFBP-2 and -3

Significant association between adenoma occurrence, ↑ IGF-II (P<0.0001) and serum IGFBP-2 (P<0.0001). Removal of adenoma led to ↓ in IGF-II (P<0.001) and IGFBP-2 (P<0.001)

[26]

Comparative

NA

13/0

IGFBP-3 and IGFBP-3 protease

Inhibition of IGFBP-3 proteolysis and invasion of colon cancer may be related

Table 4

In vitro and in vivo experimental studies linking the IGF system and colon cancer. CLA conjugated linoleic acid

Reference

Study

Type

Cell line

Parameters

Result/conclusion

[94]

Effect of TGF-β on IGFBP-3

In vitro

A variety of cell lines

Proliferation

TGF β ↑ IGFBP-3; antisense inhibition of IGFBP-3 blocks growth-promoting effect of TGF-β in all cell lines

[114]

Identification of a growth inhibitor from serum-free conditioned medium

In vitro

HT-29

Proliferation

Isolated growth inhibitor identical to human IGFBP-4

[115]

Effect of IGFBP-3 cDNA

In vitro

Caco-2

Apoptosis

Colony formation ↓ (50%) in IGFBP-3 transfected cells compared with control. Exogenous IGF-I ↑ colony formation in both transfected and control cells

[2]

Effect of antisense IGFBP-6 cDNA

In vitro

Caco-2

Proliferation

Antisense clones grew faster and the final density was 31±3% higher than the control

[53]

Influence of IGF-II on butyrate- and trichostatin A-induced apoptosis

In vitro

LIM 2405

Cell growth and apoptosis

IGFBP-3 levels ↑ by butyrate and trichostatin and IGF-II inhibited apoptotic effects of both agents

[1]

Effect of episomal expression of sense and antisense IGFBP-4 cDNA

In vitro

HT-29

Proliferation

Antisense cells showed ↑ basal and IGF-I-stimulated growth but Sense cells did not show any such increase, suggesting that IGFBP-4 over-expression was not inhibitory to HT-29 cells

[116]

Effect of IGFBP-3 on cell proliferation

In vitro

HT-29

Proliferation and apoptosis

No ↑ in cell proliferation in response to any concentration of IGFBP-3

[85]

Expression of IGF-II and IGFBPs

In vitro

Caco-2

IGF-II, IGFBP and differentiation

Expression of IGF-II, IGFBP-2, and IGFBP-6 is regulated in a differentiation-dependent manner

[103]

Effect of 1α,25-(OH)(2)D(3) and vitamin D analogues (EB1089, CB1093), 1β,25-(OH)(2)D(3)

In vitro

HT-29

Proliferation

All analogues except 1β,25-(OH)(2)D(3) inhibited cell proliferation, but relative potencies of EB1089 and CB1093 were greater than that of native vitamin

[117]

Effect of transretinoic acid(Tra): role of IGFBP-6

In vitro

Caco-2

Proliferation

Dose-dependent ↓ proliferation caused by Tra may result from IGFBP-6-mediated inhibition of IGF-II

[41]

Role of IGF-I and IGF-II

In vitro

NIH-3T3

Differentiation and proliferation

IGF-I was initially mitogenic, then caused growth arrest and differentiation. Concurrent IGFBP-3 addition blocked growth inhibition by IGF-I and IGF-II

[104]

Effect of polyunsaturated fatty acid

In vitro

Caco-2

Proliferation, IGF-II and IGFBP-6

↓ Proliferation of Caco-2 may be due to ↑ IGFBP-6 which binds to IGF-II

[118]

Effect of trans-10, cis-12-CLA

In vitro

Caco-2

Proliferation

CLA inhibits Caco-2 cell growth by ↓ IGF-II secretion

[119]

Effect of plasma from patients undergoing major surgery

In vitro

HT-29

Proliferation

Major open surgery led to depletion of IGFBP-3 in blood that promotes HT29 tumor cell proliferation in vitro

[39]

Role of IGFBP-3 in sodium butyrate (NaBt)-induced apoptosis

In vitro

Adenoma cells

Apoptosis

IGFBP-3 may act as a positive regulator of NaBt-induced apoptosis by IGF-independent mechanism

[120]

Effect of CLA isomers (trans-10, cis-12 CLA)

In vitro

HT-29

Cell proliferation

t10c12 CLA ↓ IGF-II level in a dose-dependent manner, whereas c9t11 CLA had no effect

[9]

Effect of IGFBP-6 and IGF-II

In vitro

LIM 1215

Proliferation and adhesion

IGFBP-6 inhibits IGF-II-induced but not basal proliferation and adhesion of LIM 1215 colon cancer cells

[83]

Assay of IGF-I receptor, IGF-I and IGFBP-2

In vitro

Cancer tissue

IGF-I IGF-IR and IGFBP-2 mRNA

IGFBP-2 mRNA ↑ 4- to 8-fold in CRC than controls and was highest in Dukes C samples. IGFBP-2 mRNA was localized to malignant cells and not to stroma

[56]

Expression of IGFs and IGFBPs

In vitro

Cancer tissue

mRNAs of IGFs and IGFBPs

Both IGF-I and IGFBP-2 mRNA were expressed by all normal and cancer samples, but IGF-II mRNA was only detected in cancer tissue

[25]

Effect of surface-bound plasmin

In vitro

HT29-D4

IGFBP-4 proteolysis

Proteolysis by plasmin can promote autocrine/paracrine IGF-II bio-availability in colon-cancer cells

[91]

Effect of IGF-I and Des(1,3)IGF-I

In vitro

COLO205 HT29 SW620

Cell sensitivity to IGF-I and IGFBPs

In all 3 cell lines, cell-conditioned media ↓ sensitivity to IGF-I but not to Des(1,3)IGF-I, suggesting that IGFBPs inhibit action of IGF-I

[72]

Effect of short-chain fatty acids

In vitro

Caco-2

IGFBP-2 and -3

Butyrate ↑ the secretion of IGFBP-2 in a dose-dependent and reversible manner and ↓ the secretion of IGFBP-3

[37]

Effect of IGF-I, Des-(1–3)-IGF-I

In vitro

HT29

VEGF mRNA

IGF-I ↑ the activity of VEGF promoter. IGF-I and Des-(1–3)-IGF-I had similar effects on VEGF mRNA

[70]

Effect of IGFBP-2 on colon cancer cells

In vitro

LS513, HT-29

Proliferation

IGFBP-2 inhibits proliferation in IGF-responsive colon carcinoma cell lines

[75]

IGF-I receptor antagonism in chemoradiation therapy

In vitro

SW 480 cells

Inhibition

IGF-IR antagonism ↑ the cytotoxic effect of chemoradiation therapy

[69]

IGF-IR and COX-2 activity

In vitro

Caco-2

COX-2 and IGF-II mRNA and PGE-2

Antibody to IGF-IR inhibited COX-2 mRNA expression and dominant negative IGF-IR ↓ COX-2 expression and activity

[95]

Effect of IGFBP-3

In vitro

Colonic epithelium and cancer cells

IGFBP-3, apoptosis and differentiation

IGFBP-3 alone have no effect on growth of colon cancer cells, but it enhances p-53-dependent apoptotic response to DNA damage

[80]

Expression of IGF-II and IGFBPs by different colon cancer cell lines

In vitro

COLO 205, COLO 320, Caco-2, HCT 116, HT-29, DLD-1

mRNA and protein of IGFs and IGFBPs

All cell lines expressed IGFBP2 and/or IGFBP4 mRNA and secreted IGFBP4 and/or IGFBP2; IGFBP1 was not detected in any cell line. IGFBP3 mRNA was detected only in IGF responsive cell lines

[40]

Expression of the IGFs and their receptors

In vitro

Cancer tissue

IGF-I, II and IGF-IR

IGF-I polypeptide, not mRNA, was present in small amounts in normal and malignant tissue. IGF-II was expressed 40 times, IGF-IR 2.5 times and IGF-IIR 4 times more in colonic tumours

[35]

Transfection with a truncated dominant-negative form of the IGF-IR (IGF-IRDN)

In vitro and in vivo

KM12L4 nude mouse

VEGF expression

IGF-IRDN cells showed ↓ level of VEGF mRNA and protein. In Nude mouse it led to ↓ tumour growth, VEGF expression, and vessel count (P<0.05)

[15]

Blockade of the IGF-IR and angiogenesis

In vitro and in vivo

HT-29 nude mouse

VEGF expression, proliferation and vessel count

IGF-IR blockage inhibited HT29 growth in both monolayer and soft agar (P<0.05). In nude mice it led to ↓ tumour growth (P<0.05)

The levels of circulating IGFs and IGFBPs have been shown to vary significantly between studies. The medium of collection (serum or EDTA) has been suggested to be a source of heterogeneity [33]. It has been found that levels of IGFs were about 10% higher in serum than in plasma, the differences were statistically significant (P<0.001), and IGFs were not statistically different in EDTA and heparin plasma (P=0.211 for IGF-I and P=0.654 for IGF-II). IGF levels correlated well between serum and plasma samples (r≥0.95, P<0.001). Freeze–thaw treatment for up to five cycles had little impact on plasma levels of IGFs and IGFBP-3. The IGFBP-3 levels were not different in serum and heparin plasma (P=0.75), but were significantly higher in these two specimens than in EDTA plasma (P<0.001). IGFBP-2 levels were significantly different in the three types of specimen (P<0.001).The values were 83% higher in the EDTA specimen than in heparin plasma. IGFBP-6 levels were significantly higher in serum than in plasma (P<0.01). The authors also measured the levels of IGFs and IGFBP-3 in heparin plasma after five freeze–thaw cycles, but the levels did not decline. These indicate the fact that the medium of collection can influence the results of the study related to the IGF system. Thus, there is a need to standardise the medium of collection, which will enable us to compare different studies and their significance.

Pathophysiological role of the IGFs

The IGF system plays a critical role in all phases of mammalian growth, including intrauterine, childhood and puberty [34]. This is confirmed by the evidence that targeted disruption of the mouse gene for IGF-II resulted in a reduction in foetal growth but normal postnatal growth, and disrupted IGF-I gene leads to a similar decrease in birth weight, but also persistent postnatal growth failure [34]. The liver is the most frequent site of secondaries in colon cancer [35], which may be due to the fact that it is the liver that produces most of the IGFs in circulation. IGF-I has a profound impact on tumour growth by stimulation of cellular proliferation and inhibition of apoptosis. Wu et al. [36] conducted a study involving liver-specific IGF-I-deficient (LID) mice in whose serum the IGF-I level was 25% of that in control mice. In LID mice, the growth of orthotopically transplanted colon adenocarcinoma was significantly less than in controls (31.3 vs. 56.8%, P<0.01) and the appearance of a palpable caecal tumour was not only slower (P<0.05) but also smaller (P<0.01) than that in control mice. The frequency of hepatic metastasis was also significantly lower in LID mice (31.3% vs. 44.0% in control mice, P<0.05). These results support the hypothesis that circulating IGF-I levels play an important role in tumour development and metastasis [36]. Increased systemic IGF-I may lead to colonic tumours as evidenced in acromegaly patients. Malnutrition and calorie restriction may lower the level of IGF-I, which may delay tumour progress. IGF-I also plays an important role in tissue angiogenesis [37]. It regulates VEGF expression in human colon cancer cells by induction of transcription of the VEGF gene, which is mediated through hypoxia-inducible factor-1 [38] without involving IGFBP [39]. Colorectal cancers often express 10–50 times higher levels of IGF-I and IGF-II than adjacent uninvolved colonic mucosa [35], although some studies found normal IGF-I levels [40]. IGF-I has a biphasic effect on colon cancer cells through transient inactivation of forkhead1, initially mitogenic, then mediating growth arrest and differentiation [41]. The IGF-I may increase risk of cancer, either by its anti-apoptotic activity or by modulating the effects of sex steroids [42].

The IGF-II also has a place in the development of colon cancer. Experiments conducted by Zarrilli et al. showed high levels of IGF-II RNA in proliferating Caco-2 cells, which decreased by more than 10-fold when the cells ceased to proliferate and differentiate [43]. Reduced IGF-II expression was associated with a decrease in IGF-I receptor number that was high in proliferating cells. Exogenously added IGF-II was able to stimulate proliferation of serum-deprived cells in a dose-dependent fashion [43]. Increased systemic IGF-II level does not have any predictive influence on susceptibility to colorectal cancer [44]. The IGF-II acts on the IGF-I receptor in an autocrine/paracrine manner and IGF-IR (α-IR3) antibody inhibits both basal and IGF-II-stimulated cell proliferation [45].

The IGF-IR is crucial for normal growth and development because IGF-IR knockout mouse embryos suffer generalised organ hypoplasia and invariably die at birth [15]. The IGF receptors are found in both mucosal and muscular layers of intestine and are mainly concentrated in the basolateral region of crypt enterocytes [46]. Experimental administration of IGF-I in rats showed an increase in crypt cell population, as well as linear and cross-sectional increase in the mucosal and muscular layers of intestine leading to an increase in gut weight [46]. The IGF-IR is essential for cell transformation that is induced by tumour-virus proteins and oncogene products [12]. Colorectal cancer cells often over-express IGF-I receptors [47] and when activated by IGF-I they inhibit apoptosis and allow progression through the cell cycle. Thus, IGF-I can influence both pre-malignant and cancerous stages [14]. Cancer cells with a strong tendency to metastasise have a higher expression of the IGF-IR [12, 48]. Hakam et al. [48] investigated the expression of the IGF-IR in colonic adenomas, adenocarcinomas and in corresponding metastases. Immunostaining for the IGF-IR showed strong cytoplasmic positivity in 96% (34 out of 36) of carcinomas, and in 93% (25 out of 27) of metastases. Only a faint cytoplasmic stain of the IGF-IR was identified in 83% (10 out of 12) of adenomas, while normal mucosa showed negative immunostaining. But Adenis et al. did not find any difference in IGF-IR concentration when they compared 46 frozen sections of colon cancer with 26 controls [49]. IGF-IR activation up-regulates components of the TGF-α autocrine loop resulting in TGFα-mediated EGFr activation, which was critical for IGF-IR-mediated re-entry into the cell cycle from the growth-arrested state [50].

The IGF-IIR antagonises the growth-promoting effect of IGF-II and loss of the IGF-IIR is expected in cancer [12]. The IGF-IIR acts like a clearance receptor and removes the IGF-II from circulation. Experiments show that there is a rise in the IGF-II/Man-6-P receptor message in colorectal cancer and the increase in IGF-II message is accompanied by a doubling of the IGF-II protein in the tumour tissue compared with the adjacent normal tissue [40]. Garrouste et al. isolated soluble IGF-II/mannose 6-phosphate receptor in the culture medium from HT-29 colonic cancer cells [51]. These findings suggest that the IGF-II/Man-6-P receptor may also be involved in the development of adenocarcinomas of the colon.

IGF-I, IGF-II and the IGF-IR are often over-expressed by many cancers so that IGF-I and IGF-II can act in an autocrine manner [17, 5254]. Most colonocytes secrete IGF-II as well as express the IGF-IR, but it is the sequestration of IGF-II by IGFBP that prevents the establishment of an operative IGF autocrine loop. One study examined IGF-IR expression by the colon of 344 Men Fischer rats. Colonic IGF-IR mRNA levels declined with aging (P≤0.05), while colonic IGF-I mRNA levels were unchanged, which may be a protective adaptive mechanism against aging colon [8]. Insulin, IGF-I and IGF-II induce expression of hypoxia-inducible factor-1, which is required for the expression of genes encoding IGF-II, IGFBP-2 and IGFBP-3 [55]. Colon cancer cells produce specific proteases that degrade the IGFBP secreted by these cells [56].

Loss of imprinting of the IGF-II gene and colorectal cancer

All human genes have two copies, one is paternal and the other is maternal in origin. Of these two copies of the gene, only one is functional and this phenomenon is termed genomic imprinting [57]. Imprinting is vital for normal development, and disruption of imprinting mechanisms gives very similar phenotypes in mice and humans [58]. The IGF-II gene is imprinted and expressed only from paternal allele except from the adult liver where the expression is biallelic because of promoter switching after birth [43, 59]. The IGF-IIR is also imprinted but expressed only from maternal allele. This may mean that maternally-expressed genes can act as growth suppressors because the function of the IGF-IIR is to remove IGF-II from circulation and destroy it so that less IGF-II is available to promote cell growth [57, 60]. Loss of genomic imprinting (LOI) of the IGF-II gene involves abnormal activation of the normally silent maternally-inherited allele and has been associated with personal and family history of colorectal neoplasia [61]. LOI of the IGF-II gene is found in normal colonic mucosa of about 30% of colorectal cancer patients, but it is found in only 10% of healthy individuals.

Nakagawa et al. [62] performed a quantitative analysis of IGF-II imprinting and examined the methylation pattern of the IGF-II/H19 differential methylated region (DMR) in tumour and normal colonic mucosa from randomly selected sporadic colorectal patients. Colorectal cancer samples without LOI showed complete methylation on one allele and low-level or absent methylation on the other, whereas the samples with LOI exhibited complete methylation on one allele and partial methylation on the other, indicating the relaxation of allele-specific methylation. As this methylation-dependent LOI was present in both tumors and normal colonic mucosa, it is possible that hypermethylation creates a field defect predisposing to cancer [62].

However, a later study showed contrasting results from those reported by Nakagawa. In this study, the imprinting status of IGF-II was examined in 20 CRC tissue samples by reverse transcriptase-polymerase chain reaction (RT-PCR), 12 with LOI and 8 with normal imprinting [63]. All 8 of the CRC tissue samples with normal imprinting showed the normal half-methylation pattern at the IGF-II DMR, and all 12 of the CRC tissue samples with LOI showed marked hypomethylation of the IGF-II DMR (P=0.000007). In tumours with normal imprinting, the fraction of CpG sites that were methylated was 43.6±10.9%, whereas in tumours with LOI the fraction of sites methylated was 10.9±9.4% (P<0.0001). These data suggested that normal imprinting in the colon and LOI in CRC is specifically linked to the methylation status of a DMR within IGF-II and not H19.

Imprinting of the IGF-II gene has been also used as a marker to predict the risk of CRC [64]. In a pilot study of 172 patients with adenoma or CRC, it was found that the odds ratio for LOI in lymphocytes was 5.15 for patients with a positive family history of CRC (95% CI, 1.70–16.96; probability P=0.002), 3.46 for patients with adenomas (95% CI, 1.14–11.37; P=0.026), and 21.7 for patients with CRC (95% CI, 3.48–153.6; P=0.0005). Persons with colorectal adenomas/cancer had a 5.1-fold (95% CI: 1.92–13.6) increased risk of having LOI of IGF-II in PBL compared with those without CRC.

Factors influencing the IGF system with respect to colon cancer

Dairy products

Intake of dairy products is associated with an increase in circulating IGF-I levels, but intake of low fat milk is associated with a lower risk of colorectal cancer particularly among individuals with high IGF-II/IGFBP-3 [65]. In a nested case-control study [65], it was found that there was a moderate but statistically non-significant inverse association between intake of low-fat milk or calcium from dairy food and colorectal cancer risk. Intake of dairy food (especially low-fat milk) was positively and moderately associated with plasma levels of IGF-I, IGFBP-3, and IGF-I/IGFBP-3. Non-drinkers of low-fat milk with IGF-I/IGFBP-3 in the highest tertile had a 3-fold higher risk than non-drinkers of low-fat milk with IGF-I/IGFBP-3 in the lowest tertile (RR=3.05; 95% CI=1.29–7.24), but no such increase was seen among frequent low-fat milk drinkers (RR=1.05; 95% CI=0.41–2.69). There was a statistically significant interaction between low-fat milk intake and IGF-I/IGFBP-3 in association with risk of colorectal cancer (P=0.03). Men with high IGF-I/IGFBP-3 who were frequent low-fat milk drinkers had a 60% lower risk (95% CI=0.17–0.87; P=0.02) than non-drinkers of low-fat milk. Chronic energy restriction reduces plasma levels of IGF-I and IGFBP-3, increases plasma IGFBP-1 and IGFBP-2, and thereby protects from various cancers [28]. On the other hand, obesity causes hyperinsulinaemia and reduction in plasma IGFBP-1 and IGFBP-2 [28]. In one study the mean serum IGF-I concentration was 13% lower in 92 vegan women compared with 99 meat-eaters and 101 vegetarians (P=0.0006) [66]. The mean concentrations of both serum IGFBP-1 and IGFBP-2 were 20–40% higher in vegan women compared with meat-eaters and vegetarians (P=0.005 and P=0.0008 for IGFBP-1 and IGFBP-2 respectively). Cohort studies show that dairy products and milk protect against colon cancer, but the evidence has not been supported by case-control studies [67]. No proven evidence is available regarding whether cheese and yoghurt protect against colon cancer [67] through regulation of the IGF system.

Non-steroid anti-inflammatory drugs

Non-steroid anti-inflammatory drugs (NSAIDs), which may inhibit the activity of cyclooxygenase-2 (COX-2), reduce IGF-IR expression in colon cancer lines and inhibit IGF-II-stimulated growth and invasion in a dose-dependent manner [68]. These changes are reversible when treated with PGE2 or angiotensin II [68]. Therefore, combination therapy with NSAIDs and ACE inhibitors targeting IGF-IR might be a novel and potentially promising strategy for the chemoprevention of colon cancer. There is also a direct correlation between COX-2 and IGF-II expression in Caco-2 cells. The IGF-II up-regulates COX-2 expression and PGE2 synthesis in Caco-2 cells, which is mediated by IGF-IR [69]. This is supported by evidence that when IGF-IR is blocked with antibody α-IR3 it results in not only inhibition of COX-2 expression and PGE2 synthesis, but also induction of apoptosis in Caco-2 cells [69].

Other chemicals

Retinoic acid and dexamethasone cause growth inhibition of colon cancer cells in vitro, which is accompanied by increased IGFBP-2 expression and specific IGFBP-2 proteolysis, resulting in reduced affinity for IGF-II [70]. Suramin releases IGF-II from the IGF-II-IGFBP complex so that IGF-II can act on the IGF-IR of the colon cell line [71]. Intestinal epithelial cells respond to short-chain fatty acids by altering secretion of IGFBPs [72]. Short-chain fatty acids are bacterial metabolites from unabsorbed carbohydrate. Transretinoic acid (Tra) inhibits Caco-2 cell proliferation in a dose-dependent manner via increased expression of IGFBP-6 [50], whereas it decreases the concentrations of IGFBP-2 and IGFBP-4. IGF-IR expression on colorectal adenocarcinoma cell lines is decreased by the chemopreventive agent N-acetyl-l-cysteine [73]. Conjugated linoleic acid (CLA) inhibits cell proliferation and induces apoptosis in HT-29 cells [74]. Levels of IGFBP-3, which may induce apoptosis by IGF-dependent or -independent mechanisms, are increased by butyrate and trichostatin A and IGF-II augments this effect [53]. Butyrate also increases the secretion of IGFBP-2 in a dose-dependent and reversible manner [53, 72].

Chemoradiation

Cell culture studies show that IGF-I receptor activation blocks the expected cytotoxic effect of 5FU and external beam irradiation [75]. IGF-IR antagonism using a monoclonal antibody combined with chemoradiation produces an increased cytotoxic response [75], which may be useful in humans in the near future.

Diabetes

Patients with type-2 diabetes mellitus are at high risk of developing colorectal cancer [76]. Elevated levels of postprandial insulin and C-peptide increase the colorectal cancer risk [76]. Insulin may directly activate its own receptor, the receptors for IGF-I, hybrid insulin/IGF-I receptor or it may act indirectly by increasing the bioavailability of IGF-I by altering IGFBP levels [24].

Acromegaly

Acromegaly is associated with increased proliferation of normal colonic epithelium [28] and increased prevalence of tubulovillous adenoma and colon carcinoma [77]. In a retrospective study [77], serum IGF-I, IGF-II, IGFBP-2 and IGFBP-3 in patients with acromegaly and in those with colonic neoplasia without acromegaly were analysed. Mean serum IGF-I and IGFBP-3 levels were significantly elevated in patients with acromegaly without colonic neoplasia and in those with acromegaly and colonic neoplasia, and significantly reduced in those with colonic neoplasia without endocrine disease, compared with controls (P<0.001). However, median serum IGFBP-2 levels were significantly elevated in patients with acromegaly and colonic neoplasia (P<0.01) and in those with colonic neoplasia without endocrine disease (P<0.0001). The precise mechanism of the development of colon cancer in acromegaly patients remains unknown. It may be due to the stimulatory role of excess growth hormone/IGF-I or excess IGF-II [77]. In one study involving acromegaly 5% of patients had colorectal cancer and 25% had colorectal adenomas [78]. Antagonists of growth hormone-releasing hormone (GH-RH) inhibit growth of HT-29 human colon cancers, both in vitro and in vivo [5]. The effect of GH-RH antagonists may be mediated through reduced production and secretion of IGF-II by cancer cells.

IGF-related properties of human colon cancer cells

Heterogeneity

Different cancer cells behave in a different way with respect to IGF responsiveness. Virtually all human colon cancers express IGF-I receptors, but only 50% respond to growth and the mitogenic effects of exogenous IGF-I [79, 80]. Therefore, cancer cell lines can be divided into IGF responsive (e.g. COLO205, Caco-2) and IGF-non responsive (e.g. HT-29, DLD-1) groups [80]. Some colon cancer cells (e.g. LIM 1215) are IGF responsive, but IGF-II is not a major autocrine factor for these cells. All these suggest the existence of heterogeneity between colon carcinoma cell lines with respect to the role of the IGF system [9].

Polarity

Colon cancer cells behave in a polarised fashion and secrete different members of the IGFBPs from different areas of the cell surface. Caco-2 cells secrete IGFBP-3 from the apical surface, IGFBP-2 predominantly from the basolateral surface and IGFBP-1 and -4 from both surfaces. However, epidermal growth factor (EGF) induces secretion of IGFBP-4 more from the apical surface than from the basolateral aspect. It does not affect the polarity of the other IGFBPs [81]. HT29-D4 human colonic carcinoma cells secrete endogenous IGF-II predominantly (66%) from the basolateral cell surface where type I IGF receptors are almost all (>96%) localised. IGFBP-2 and IGFBP-4 are secreted primarily into the basolateral side (71 and 87% respectively), whereas IGFBP-6 is targeted at the apical surface (76%). The differential sorting of the various forms of IGFBPs may play a modulatory role in the maintenance of functional polarity in the differentiated colon cancer cells [82].

Differential expression

The majority of human colon cancer cells express IGF-II, IGFBP-2 and IGFBP-4 and less frequently IGFBP-3, IGFBP-5 and IGFBP-6; the levels of IGFBP-2 and IGFBP-6 decrease as differentiation proceeds (differentiation-dependent manner) [10, 8385]. Singh et al. measured the expression and secretion of IGF-II, IGFBP-2 and IGFBP-4 in relation to growth and differentiation of Caco-2 human colon cancer cells, which undergo spontaneous enterocytic differentiation in culture. Caco-2 cells demonstrated an initial rapid phase of growth followed by a significant retardation in the growth. Changes in growth and differentiation were accompanied by >80% decline in the relative concentration of IGF-II mRNA. In contrast, the relative mRNA concentrations of inhibitory binding proteins (IGFBP-2 and IGFBP-4) increased rapidly to 200% of day 2 values by days 5–7 before returning to baseline levels by day 13 [84]. The quantity and the type (differential expression) of IGFBPs and IGF-II may play a critical role in both proliferation and differentiation of colonocytes [80, 84]. HT29-D4 cancer cells remain in an undifferentiated state in culture medium when they cannot use endogenous IGF-II as an autocrine regulatory factor [82]. The IGF-IR primarily controls the differentiation of colonic cells.

Role of individual IGFBPs in colorectal cancer

IGFBP-1

IGFBP-1 is mainly produced by the liver, which is regulated by insulin [24]. It has equal affinity for IGF-I and IGF-II. It can either potentiate or inhibit the action of the IGFs [86]. It can also act independent of the IGFs by interacting with α5β1 integrin to cause cell adhesion and migration [87]. Kaaks et al. [29], in their nested case-control study analysed serum IGFBP-1, -2, and -3 levels of 102 women who subsequently developed colorectal cancer and 200 matched control subjects. They observed a statistically significant decrease in colorectal cancer risk with increasing levels of IGFBP-1 (P=0.02; OR in the upper quintile=0.48; 95% CI=0.23–1.00), as well as for the highest quintile of IGFBP-2 levels (P=0.06; OR=0.38; 95% CI=0.15–0.94). Colorectal cancer risk was significantly increased for the highest quintile of IGFBP-3 (OR=2.46; 95% CI=1.09–5.57) [29]. In vitro studies show that colon cancers do not express IGFBP-1 [88]. Therefore, the exact role of IGFBP-1 in colon cancer is not clearly known.

IGFBP-2

IGFBP-2 is up-regulated in different pathological and unphysiological situations like trauma, certain tumours and starvation. Serum transgenic mice over-expressing IGFBP-2 show significantly reduced body weight gain, demonstrating that IGFBP-2 is a negative regulator of normal somatic growth, most probably by sequestering the IGFs from their receptors [89].

In vitro studies show that all IGFBP-secreting colon cell lines express the IGFBP-2 gene [80, 90] and IGFBP-2 inhibits cell proliferation in several IGF-responsive colon carcinoma cell lines, but, in contrast, Miraki-Moud et al. [77] have shown that Caco-2 cells that over-express IGFBP-2 grow at a faster rate, indicating that IGFBP-2 stimulates cell proliferation. Therefore, IGFBP-2 can either inhibit or potentiate cancer cell growth under different circumstances.

Most of the IGFBPs are synthesised and secreted by cells of the colonic mucosa, but the proteolysis of secreted IGFBP-2 occurs mainly in the colon cancer tissue. This selective degradation may confer a growth advantage [91]. Colon cancer extracts are able to degrade exogenous IGFBP-2, IGFBP-3 and IGFBP-4, whereas normal tissue extracts have no effect on IGFBP-2 [56]. A study carried out by Mishra et al. shows that IGFBP-2 mRNA is localised to malignant cells and not to the surrounding stromal cells, suggesting the autocrine role of IGFBP-2 [83]. Immunostaining for IGFBP-2 shows strong areas of immunoreactivity in the cytoplasm of malignant colonic epithelium compared with benign epithelium.

In human colon cancer, IGFBP-2 mRNA is often over-expressed and circulating IGFBP-2 is often elevated [2]. IGFBP-2 is increased in patients with colonic neoplasia irrespective of whether the patient has acromegaly or not [77]. It is due to increased expression of the protein itself. Serum IGFBP-2 levels reflect the tumour load and their levels fall after curative resection [92]. The IGFBP-2 level often reflects the tumour grade and the stage of the disease [89]. In one study, IGFBP-2 mRNA levels were increased 4- to 8-fold in patients with colon cancer compared with controls. Patients with Dukes’ stage C disease had the highest levels of IGFBP-2 mRNA. Therefore, IGFBP-2 may be implicated in colon cancer metastases and prognosis. Its sensitivity as a tumour marker increases when used in combination with carcinoembryonic antigen (CEA) [92]. Serum IGFBP-2 may have an adjuvant role in cancer surveillance in patients with colorectal cancer.

IGFBP-3

Like IGF-I, most of the IGFBP-3 found in circulation is produced by the liver under the influence of growth hormone. It has both IGF-I inhibiting and potentiating actions. Human colon carcinoma cell line, Caco-2, produces IGFBP-3, the secretion of which correlates with markers of enterocyte differentiation. Endogenous IGFBP-3 expression in Caco-2 cells inhibits colony formation and increases apoptosis. IGFBP-3 appears to decrease the tumorigenic potential of colon cancer cells in vitro. IGFBP-3 may directly inhibit target cells. One study shows that IGFBP-3 may inhibit proliferation and induce early differentiation of Caco-2 cells [93], but another study shows that IGFBP-3 mediates the transforming growth factor-β1 (TGF-β1)-induced proliferation of three metastatic or highly aggressive colon carcinoma cell lines [45]. TGF-β1 increases IGFBP-3 abundance, while phosphorothiolated antisense oligonucleotides to IGFBP-3 block the growth-promoting effect of TGF-β1 in three different cell lines. IGFBP-3 induces carcinoma cell growth in a dose-dependent and time-dependent manner in vitro [94]. There is one more study that mentions that treatment of human colonic tumour cell lines with IGFBP-3 alone was shown to have no effect on growth [95]. IGFBP-3 may act as a positive regulator of butyrate-induced apoptosis in colonic epithelial cells, and represents a potentially important mechanism whereby the sensitivity of colonic epithelial cells to sodium butyrate-induced apoptosis can be increased [39].

High levels of circulating IGF-I and particularly low levels of IGFBP-3 are associated independently with an elevated risk of large or tubulovillous/villous colorectal adenoma and cancer [96].

Treatment of human colonic tumour cell lines with IGFBP-3 alone was shown to have no effect on growth. However, an increase in p53-dependent apoptosis was observed in the presence of IGFBP-3 24 h after the induction of DNA damage by γ-irradiation in colonic epithelium [95]. The level of IGFBP-3 protein expression in colonic epithelial cells correlates with the p53 status of the cells [39]. The IGFBP-3 acts independent of IGF by enhancing the p53-dependent apoptotic response of colorectal cells to DNA damage [46, 95]. The IGF-I-independent growth-inhibitory effects in colon cancer [10] may also occur through IGFBP-3-specific cell surface association proteins or receptors and it involves nuclear translocation [7]. The IGFBP-3 may also act through its own receptors and mediates IGF-independent growth inhibitory action [97].

IGFBP-3 may confer a selective growth advantage on tumour cells in vivo because levels of IGFBP-3 are elevated at least 2-fold in 7 out of 10 resected colon cancers compared with adjacent normal tissue [41, 94]. Matrix metalloproteinase-7 (MMP-7) secreted by cancer cells has been implicated classically in the basement membrane destruction associated with tumour cell invasion and metastasis. Addition of IGFBP-3 inhibited IGF-I-mediated IGF-IR phosphorylation and activation of the downstream molecule Akt in two human colon cancer cell lines (COLO201 and HT29). Co-incubation of the IGF-I/IGFBP-3 complex with MMP-7 restored IGF-I-mediated IGF-IR phosphorylation and activation of Akt in these cell lines [98]. These results indicate that MMP-7 proteolysis of IGFBP-3 plays a crucial role in regulating IGF-I bioavailability, thereby promoting cell survival. This mechanism may contribute to the tumorigenesis of MMP-7-producing IGF-IR-expressing tumours in the primary site and to organ-specific metastasis in a paracrine manner.

IGFBP-4

It is unique in that it is the smallest IGFBP and exists in both glycosylated and non-glycosylated forms [11]. IGFBP-4 is secreted by almost all colon cancer cell lines and it primarily functions as an inhibitory protein for colon cancer cell lines [88] by binding both IGF-I and IGF-II with equal affinity [11] and overproduction of IGFBP-4 could induce cachexia by preventing IGF-mediated anabolic effects in muscle and adipose tissue [99]. Approximately 15% of IGFBP-4 is associated with the extracellular matrix. Endogenous IGFBP-4 is a potent inhibitor of the mitogenic effects of endogenous and exogenous IGFs [1]. The role of endogenous IGF-I in regulating IGFBP-4 degradation was confirmed by the ability of an IGF-I antagonist to inhibit IGF-I-activated IGFBP-4 proteolysis in intact cells. There is a significant up-regulation of IGFBP-4 expression in a human colon cancer cell (Caco-2) line on spontaneous differentiation of the cells in culture, which suggests that the expression of IGFBP-4 may be related to growth and differentiation of colon cancer cells [88]. IGFBP-4 proteolysis by cell-bound plasmin can promote autocrine/paracrine IGF-II bioavailability in colon-cancer cells [82]. Inhibition of IGFBP-4 mRNA confers a growth advantage to the cells in response to endogenous and exogenous IGFs. In human intestinal smooth muscle cells levels of secreted IGFBP-4 are determined by the confluence-dependent production of a cation-dependent serine protease that is activated by endogenous IGF-I [27]. IGFBP-4 also has IGF-IR-independent actions. Cell culture studies carried out by Diehl et al. show that IGFBP-4 reduces colony formation independent of IGF-IR, but not cell proliferation and migration, both of which are mediated via IGF-IR [100].

IGFBP-5

IGFBP-5 can either stimulate or inhibit cell proliferation [101]. IGFBP-5 expression is up-regulated by antiproliferative agents such as retinoic acid and vitamin D-related compounds. The exact role of IGFBP-5 in colorectal cancer is not known. Studies involving prostate cancer cell lines show that IGFBP-5 stimulates growth via IGF-dependent and -independent mechanisms [102]. But it inhibits the growth of osteosarcoma and cervical cancer cell lines [102]. It has both stimulatory and inhibitory effects on breast cancer cells depending upon the stage. This means that it has a complex role in cancer, which needs further research.

IGFBP-6

IGFBP-6 has a high affinity for IGF-II [103]. IGFBP-6 was found to inhibit both HT-29 and Caco-2 cell proliferation by binding to endogenously produced IGF-II, thereby preventing IGF-II from interacting with the IGF-I receptor to stimulate cellular proliferation by the autocrine mechanism [2]. Kim et al. [2] transfected Caco-2 cells with the antisense IGFBP-6 expression construct and the antisense clone grew at a rate faster than that of the control and reached a final density that was 31±3% higher than the control. Analyses revealed that accumulation of IGFBP-6 mRNA and concentrations of IGFBP-6 peptide produced by the antisense clone were significantly decreased. Exogenous IGF-I and IGF-II stimulated proliferation of both the control and antisense clones in a dose-dependent manner, but the relative potency and efficacy of IGF-II was higher in the antisense clone than in the control. This indicates that suppression of IGFBP-6 secretion correlates with an increase in the basal rate of Caco-2 cell growth and indirectly shows that IGFBP-6 inhibits Caco-2 proliferation by binding to endogenously-produced IGF-II.

In one study IGFBP-6 had no effect on basal proliferation in LIM 1215 colon cancer cells, but co-incubation of IGFBP-6 decreased IGF-II but not IGF-I-induced proliferation [9]. Synthetic analogue of vitamin D3 (EB1089 and CB1093) and polyunsaturated fatty acids(eicosapentaenoic acid and docosahexaenoic acid) increase the level of IGFBP-6 in colon cell lines. In vitro study shows that polyunsaturated fatty acids inhibit Caco-2 cell proliferation by decreasing IGF-II and increasing IGFBP-6, resulting in less free IGF-II [104].

IGFBP-7 (IGFBP-rP-1)

IGFBP-7 is also known as ‘mac 25’ and it shows 20–25% similarity to other IGFBPs [97]. The IGFBP-7 specifically binds to both IGF-I and IGF-II. In comparison with IGFBP-3, IGFBP-7 has at least a 5- to 6-fold lower affinity for IGF-I and a 20- to 25-fold lower affinity for IGF-II [97]. IGFBP-7 is normally expressed by colonic mucosa. The level of IGFBP-7 expression is reduced in several cancer cells, including colon cancer, which suggests that it may function as a direct tumour growth-suppressing factor [44, 97]. But the study conducted by Shao et al. showed an increase in expression of IGFBP-rP1 in colonic adenoma and colon cancer [105]. This shows that it may act both as tumour promoter as well as a suppressor in different circumstances.

Summary and future recommendations

The IGF system has a complex role in the development and progress of colorectal cancer. This system is essential for human day-to-day tissue regeneration. The biological activity of IGF is determined by the integrated actions of circulating IGF-I and IGFBP and by local production of IGF, IGFBP and IGFBP protease. So far, the in vitro, in vivo and population-based studies have shown that increased IGF-I and -II may be related to the development of colon cancer. Since the association is not clearly proven consistently it appears that the ligands may play an adjuvant role rather than inducing carcinogenesis primarily. When activated, the IGF-IR not only stimulates the synthesis of RNA and DNA, cell proliferation, differentiation and increases cell survival, but it also increases angiogenesis. All IGFBPs bind to free IGFs and inhibit their action on IGF-IR, which is vital for carcinogenesis. Thus, IGF-IR is more important than any other factor because it mediates growth of cancer. IGF-IR antagonism using monoclonal antibody, when combined with chemoradiation, produces increased cytotoxic response, which may be useful in a clinical situation. Manipulation of IGF-IR may help to control several cancers, and this needs to be studied properly. Loss of imprinting of the IGF-II gene is responsible for at least some colorectal cancers. Thus, it may be possible to cure this LOI by insertion of the correct gene copy after further study. We do not have enough evidence to recommend the use of IGFBPs in clinical therapeutic trials for cancer. Serum IGFBP-2 reflects tumour load, so it could be used in cancer surveillance in patients with colorectal cancer after further large studies. Its sensitivity as a tumour marker increases when used in combination with CEA. IGFBP-3 is the main binding protein that carries and binds more than 90% of the IGFs in circulation. IGFBP-3 is the binding protein that carries and binds more than 90% of the IGFs in circulation. Some of the binding proteins like IGFBP-5, -7 etc. have both stimulatory and inhibitory effects under different circumstances. If future studies show the exact circumstances under which they inhibit cancer growth they can be utilised to arrest cancer growth. As stated in the meta-regression analysis by Renehan et al. [32], case-control studies might have overestimated the magnitude of association between IGF and cancer. The medium of collection is a source of heterogeneity. It has to be standardised so that all available data can be compared, otherwise the results of studies carried out by different individuals will show different results. The medium of collection can influence the value of IGF and IGFBP levels, leading to variation in the results. Much of our current knowledge is from in vitro studies. There is not much information available regarding the usage of these proteins in the prevention of cancer in vivo. There is, therefore, a need for further in vivo studies to enable us to manipulate these binding proteins to prevent and control various cancers.

Copyright information

© Springer-Verlag 2005