Background: Insulin-like growth factor 2 (IGF2) is a protein hormone known to regulate cell proliferation, growth, migration, differentiation and survival. The gene is parentally imprinted in the sense that transcripts are almost exclusively derived from the paternal allele. Loss of imprinting of the IGF2 gene is a recurrent observation in growth disorders that combine overgrowth with a variety of malignant tumours. Moreover, IGF2 has been proposed to play a role in the development of a variety of seemingly unrelated cancers that play an important role in geriatric medicine, e.g. breast cancer, colon cancer and lung cancer. Finally, IGF2 has been implicated in cardiovascular disease, since, for example, IGF2 has been shown to influence the size of atherosclerotic lesions. Objective: To summarize current knowledge about IGF2, its interactions with binding proteins and receptors and connections with key diseases. Methods: The contents of this paper were based on reviews of existing literature within the field. Results: There is a substantial amount of research linking IGF2 to growth disorders, cancer and to a much lesser degree cardiovascular disease. Some of the studies on IGF2 and tumour growth have yielded conflicting results, for instance regarding its effect on apoptosis. Conclusion: Today, our knowledge on how IGF2 is composed and interacts with receptors has come a long way. However, there is comparatively little information on how IGF2 affects tumour growth and cardiovascular diseases such as atherosclerosis. Thus, further research will be needed to elucidate the impact of IGF2 on key diseases.
The existence of insulin-like growth factors (IGFs) or somatomedins was first predicted by Salmon and Daughaday  in 1957, who proposed that pituitary growth hormone exerted its effects on skeletal growth by using an intermediate class of growth-promoting peptides. Further characterization revealed two different molecules, which were able to exert growth hormone-like effects on cartilage explants in vitro . Sequence analysis revealed a significant homology with pro-insulin, and hence they were termed IGF1 and IGF2 [2,3]. Although IGF1 and IGF2 elicit very similar biological responses, there are significant differences in their pattern of expression in vivo. In mammals, IGF1 is preferentially expressed after birth and is produced almost exclusively in the liver. IGF2 is preferentially expressed in early embryonic and fetal development in a wide variety of somatic tissues. The adult expression of IGF2 takes place in the liver and in the epithelial cells lining the surface of the brain, i.e. the meninges and choroid plexus .
IGF2 is a crucial factor for the regulation of cell proliferation, growth, migration, differentiation and survival. IGF2 (as well as IGF1) interacts with several receptors and binding proteins in order to exert its actions. It binds to the non-signaling IGF type 2 receptor (IGF2R) with high affinity. This receptor is homologous to the cation-independent mannose-6-phosphate (M6P) receptor. IGF2 can also bind to different signaling receptors, such as the IGF type 1 receptor (IGF1R) and the insulin receptor, albeit with lower affinity, although an alternatively spliced version of the insulin receptor, named insulin receptor isoform A, may bind IGF2 with higher affinity.
Both IGF1 and IGF2 are present in the circulation and can be readily detected in plasma. As might be predicted from the pattern of synthesis, circulating IGF1 levels rise during juvenile life and then decline after puberty, while circulating IGF2 levels are highest in the fetal circulation . Circulating IGFs are mostly associated with 6 specifically designed binding proteins (IGFBPs) which exhibit tissue- and stage-specific expression. In vitro, all IGFBPs inhibit the biological activity of IGFs, suggesting that part of their function may be to restrict the availability of biologically active IGFs .
Altogether, these findings weave an intricate pattern of biological activity which this article aims to summarize. The IGF2 protein, its receptors and binding proteins will be elaborated on, and an attempt will be made to establish whether or not IGF2 itself or its interacting proteins may have any associations with key diseases.
In mice, the IGF2 gene resides on chromosome 7 and contains 6 exons; the coding region is confined to only 3 exons (fig. 1). The IGF2 gene comprises many alternatively spliced transcripts from different promoters. In the mouse, the expression of the different mouse IGF2 transcripts is dependent on promoter usage, with P0-derived transcripts only expressed in the labyrinthine layer of the mouse placenta, whereas transcripts from P1-P3 are found throughout the developing embryo and placenta .
The human IGF2 gene is located on chromosome 11p15.5 and stretches over approximately 30 kb of DNA. There are 4 promoters and 10 exons (fig. 1), giving rise to different transcripts depending on which promoter the transcript stems from .
Although the human IGF2 gene contains 10 exons, only the last 3 contain coding sequences. The IGF2 gene is active in nearly all human embryonic, extraembryonic and fetal tissues. The quantity of the transcripts differs between organs, but in all cases the transcription is driven from promoters P2-P4, with P4 being predominantly active. The transcription of the IGF2 gene declines rapidly after birth in most tissues. In rodents, in most tissues where IGF2 is expressed at high levels throughout embryogenesis and fetal development, all 3 promoters are downregulated after birth, and the transcriptional activity continues only in exchange tissues surrounding the central nervous system .
In human adult life, transcripts derived from P1 are found exclusively in liver and choroid plexus-leptomeninges. These transcripts contain an internal ribosomal entry site in their leader sequence. Human P2-derived transcripts are usually found in low quantities in fetal liver and only reach higher levels in transformed cell lines or in neoplastic tissues. P3 and P4 transcripts are found in fetal as well as adult tissues, with the P4 promoter being predominantly active. In the mouse, transcripts derived from the P2 promoter disengage from polysomes during development and in a variety of cultured cells .
Genomic imprinting is a form of developmental gene regulation whereby only one of the parental alleles is expressed. As more examples of imprinted genes are being discovered, it is becoming obvious that these sequences are clustered into chromosomal domains, implying that imprinting may be regulated in a regional fashion. The IGF2 gene was one of the first genes shown to be imprinted, and it was clearly demonstrated that the paternal IGF2 allele is transcribed whereas the maternal allele is silent. Interestingly, silencing of the paternal allele resulted in a reduced size of the offspring, which was otherwise normal and fertile . This principle is persistent in rodents as well as in humans, with some notable exceptions. In adult life, both alleles are transcribed in human liver as well as in the central nervous system. In humans, there appears to be a fundamental difference between different promoters, since the 3 fetal promoters are clearly subject to imprinting whereas the adult promoter is not.
The biological significance of parental imprinting has been the subject of a great deal of discussion. One elegant hypothesis postulates that imprinting reflects an ongoing struggle between maternal and paternal genomes. The finding that the IGF2 and IGF2R genes are oppositely imprinted supports this hypothesis. Since the maternally produced IGF2R acts as a scavenger for paternally expressed IGF2, its effects are neutralized before the growth factor can reach the signal-transducing IGF1R. Whatever the underlying reason, the existence of imprinting must confer a selective advantage that outweighs the susceptibility of imprinted genes to loss of imprinting mutations.
In the most common case, IGF2 is first synthesized as a pro-hormone/precursor hormone containing 180 amino acids which is subsequently processed and finally appears as a 67-amino acid bioactive IGF2 protein (fig. 2). At first, a signal peptide containing 24 amino acids is removed from the N terminus, generating pro-IGF2 (156 amino acids). Subsequent cleavage of pro-IGF2 then results in a 104-amino acid peptide product [IGF2(1-104)]. Endoproteolyzation generates IGF2(1-87), and as a result, the mature IGF2 peptide consists of 67 amino acids. Both IGF2 precursor forms, namely IGF2(1-104) and IGF2(1-87), collectively named ‘big IGFs', have been found circulating in human and bovine serum .
The 180-amino acid preproprotein stems from the bona fide transcript (isoform 1). There is also a variant transcript (isoform 2) which contains two alternate 5′ coding exons that has a different 5′ untranslated region. Both isoform 1 and 2 variants encode the 180-amino acid preproprotein. However, there is a transcript variant which contains 2 alternate exons at the 5′ end, one non-coding and the other coding, compared to the bona fide transcript. This results in the use of an upstream AUG and a longer 236-amino acid protein isoform with an N terminus which is different from isoform 1 . The quantitative determination of isoform expression remains to be elucidated.
The Mature Peptide
IGF2R is a type I transmembrane glycoprotein composed of a large extracellular region, a small 23-residue transmembrane region and a 167-residue cytoplasmatic tail (fig. 3). The extracellular region consists of a 40-residue amino acid signal sequence and 15 homologous extracellular repeat domains, each of them containing between 124 and 192 amino acids .
Each extracellular repeat domain contains approximately 147 amino acids. Binding of IGF2 occurs primarily in domain 11. The affinity of this domain for IGF2 is believed to be further enhanced by a fibronectin type II-like insert in domain 13, though the exact mechanisms of this enhancement remain unclear. There are also unanswered questions regarding the precise binding sites within domain 11; analyses so far have pointed towards 3 different proposed locations, namely a hydrophobic pocket, a cluster of residues and a disparate set of residues .
IGF2R regulates the amount of circulating and tissue IGF2 by transporting the ligand into the cell and degrading it . The receptor is multifunctional and binds not only IGF2 but also M6P-marked lysosomal enzymes at domains 3, 5 and 9, enabling the transfer of newly synthesized lysosomal enzymes from the trans-Golgi network to late endosomes. Lysosomally destined enzymes are recognized by an M6P tag, which prompts them to bind to the M6P receptor. They are then transported to late endosomes via clathrin-coated vesicles. After reaching the late endosomes, the enzymes are released and transported to their final destination, the lysosomes, whereas the M6P receptors are either headed for the cell surface or back to the Golgi network .
Like IGF2, IGF2R is an imprinted gene (fig. 4), but while IGF2 is only expressed from the allele inherited from the father, IGF2R is exclusively expressed from the allele inherited from the mother. In mice, the imprinting of IGF2R is regulated by the intron 2 region. The paternal allele contains an antisense transcript, which stems from intron 2, mediating the silencing of the paternal IGF2R allele. Deletion of this intron 2 region disrupts the silencing, leading to biallelic expression of paternally inherited IGF2R .
Mice inheriting a disrupted IGF2R gene from their mother, thereby not expressing IGF2R in tissues, have been shown to suffer from overgrowth and perinatal lethality, presumably because of cardiorespiratory failure (due to malformed lungs and abnormalities in cardiac muscle). However, when the same gene was inherited from the father, no abnormalities in development were recorded, confirming that the IGF2R gene is paternally imprinted . Thus, studies conducted on sheep revealed that an imprinting defect of the IGF2R gene leading to loss of IGF2R expression causes plasma levels of IGF2 to rise. The result of this is overgrowth .
Single-nucleotide polymorphisms in IGF2R lead to an increased risk of cancer, and thus IGF2R has been referred to as a tumour suppressor gene. Several recent studies point to IGF2R single-nucleotide polymorphisms as risk factors for breast cancer, brain tumours and osteosarcoma [15,16,17]. Associations with hepatocellular, gastrointestinal, ovarian and prostate cancer have also been shown .
IGF1R and Insulin Receptor
The insulin receptor and IGF1R are both members of a receptor subfamily of transmembrane tyrosine kinases. They are structurally homologous, with the tyrosine kinase domains sharing 84% sequence identity and the juxtamembrane and C-terminal regions possessing 61 and 44% sequence identity, respectively .
Although these receptors mainly bind insulin and IGF1, they also bind IGF2 with high affinity, although the affinity is slightly higher for IGF1. However, an alternatively spliced version of the insulin receptor lacking exon 11, insulin receptor isoform A, has been shown to bind IGF2 with high affinity. Consequently, cell proliferation, differentiation, migration and survival may be mediated via this receptor as well. The insulin receptor B isoform is the classical insulin receptor, which can bind IGF2 but will mainly result in a metabolic response .
In addition to the homodimer forms of IGF1R and the insulin receptor, there are IGF1R/insulin receptor heterodimers that have been considered at least as frequent as the homodimeric variants .
Interaction with the IGF1R is associated with cell proliferation, differentiation, migration and survival. When activated, the insulin receptor and IGF1R initiate a phosphorylation cascade via two different pathways. Upon ligand binding, the receptors initially undergo autophosphorylation, which enables different adapter molecules such as insulin response signalling (IRS)-1 and -2 and Scr homology 2 domain containing (SHC) to bind. The two pathways then activated are the phosphatidylinositol 3-kinase (PI3K)-AKT/protein kinase B (PKB) pathway and the mitogen-activated protein kinase (MAPK) pathway, which in this case involves Ras activity (fig. 5). The PI3K/PKB pathway leads to metabolic activity, and the MAPK pathway regulates cell growth and differentiation, in addition to controlling the expression of certain genes .
In the PKB pathway, ligand binding induces phosphorylation of IRS-1. PI3K then catalyzes the conversion of phosphatidyl inositol phosphate (PIP)2 to PIP3, which activates AKT. Activation of AKT increases the uptake of glucose in the cell but also affects other proteins along the pathway. Eukaryotic initiation factor 4E (eIF4E) is normally inhibited in a complex with its binding protein eIF4E-binding protein 1 (4EBP1), but phosphorylation of 4EBP1 frees eIF4E, leading to protein synthesis. mTOR phosphorylates S6 protein kinase, activating it, which also leads to protein synthesis. This pathway can also inactivate pro-apoptotic transcription factors and proteins, leading to decreased apoptosis.
In the MAPK pathway, SHC becomes phosphorylated and binds growth factor receptor bound protein 2 (GRB2) and son of sevenless homolog 1 (SOS1) in a complex which activates Ras, MAP kinases (MEK) and finally serum response factor (SRF) and E twenty-six-like transcription factor 1 (ELK1), which lead to mitogenic activity. This pathway can also regulate apoptosis [21,22].
However, it should be noted that signaling via the IGF1R is only one of several parallel intracellular pathways that are involved in the intracellular mediation of mitogenic as well as other biological messages. Moreover, the interaction between the PKB and MAPK pathways and other regulatory pathways is being found to be more and more intricate. An increasingly complex web of signaling molecules is continuously being revealed [10,23].
When IGFs circulate freely, they are unstable and subject to degradation. To achieve functional stability, they require specific IGFBPs for transportation in the blood stream. There are 6 classical IGFBPs which bind IGFs with high affinity and have a large part of their amino acid sequence in common. More recently, a group of proteins binding IGFs with lower affinity were discovered. Although they are structurally related to the classical IGFBPs and considered to be part of the IGFBP superfamily, due to their low binding affinity they are referred to as IGFBP-related proteins .
The availability and distribution of IGF2 in different tissues is controlled through IGFBPs , as well as their half-life in blood. This function makes them powerful modulators of IGF2 action. Different cells and tissues synthesize different combinations of IGFBPs.
IGFBPs are in turn controlled by proteases secreted by various tissues. Proteolytic cleavage of IGFBPs negatively affects their IGF binding affinity. Phosphorylation of IGFBPs could also reduce protein binding activity, but the exact biological significance of these findings is unknown .
IGFs in Disease
The growth-promoting and beneficial functions of IGF2 during embryonic development and placental growth have been well documented. In certain instances, these functions may also work to the disadvantage of the individual. One such scenario is if the mechanisms that regulate expression of IGF2 are dysfunctional.
Cardiovascular disease and cancer are two large disease groups that play an immense role in geriatric medicine. If the quality of clinical management in ageing patients is to be capable of meeting the demands of the future, it is necessary to integrate our knowledge about basic mechanisms into the development of future therapies.
IGF2 plays a pivotal role in fetal growth. Interest has therefore been focused on clinical syndromes that display aberrant growth properties. One of the first syndromes in which IGF2 expression was linked to a growth disorder was Beckwith-Wiedemann syndrome (BWS) [13,14]. This syndrome is particularly interesting since it provides a link between aberrant growth and tumour development. Embryonal tumours occur in about 5% of BWS patients , Wilms' tumour being the most prominent, but adrenocortical carcinoma, hepatoblastoma and rhabdomyosarcoma may also develop in BWS patients . Overexpression of IGF2 is frequently observed in cases of Wilms' tumour, and for a long time, it was thought to be a tumour promoter . However, this may not be a universal rule since IGF2 was found to induce apoptosis and necrosis in Wilms' tumour cells .
BWS as a link between overgrowth and tumourigenesis provides an unusually useful concept and has given unique insights into how epigenetics links diseases together. Human chromosome 11p15.5 contains two imprinted domains. In one of these domains, simply referred to as domain 1, IGF2 and H19 genes are normally expressed. The IGF2 gene is parentally imprinted, meaning that the IGF2 gene is only expressed if inherited from the father. H19, on the other hand, is only expressed if inherited from the mother. The prevailing theory is that the imprinting is regulated by methylation of the IC1 region. Methylation controls the activity of IC1. On the maternal chromosome, IC1 is unmethylated. This enables zinc finger proteins to bind, most importantly CTCF, which prevents IGF2 promoters and enhancers from being activated. H19 is then expressed while IGF2 is silent (fig. 6).
On the paternal chromosome, IC1 is methylated, so that CTCF cannot bind; thus, IGF2 promoters and enhancers can be activated. This leads to IGF2 being expressed, while H19 is silent. In most cases of BWS, this mechanism is dysfunctional, and there is a dysregulation of gene expression. In 5% of BWS patients, the molecular defects are limited to a gain of methylation at IC1 . This means that IC1 is methylated on both the maternal and paternal chromosome, resulting in biallelic activation of IGF2 and biallelic silencing of H19. This leads to a higher propensity for developing Wilms' tumours as well as a variety of other cancers.
However, there are also other genes involved in the etiology of BWS, such as KCNQ1OT1 and CDKN1C. They involve a region named IC2, which is responsible for most of the molecular defects in BWS. Paternal uniparental disomy of chromosome 11p15 is another possible cause of BWS, meaning that the affected individual inherits two copies of the chromosome from the mother and none from the father [26,27].
Even though data are not always consistent, it is nowadays believed that IGFs in general and IGF2 in particular can promote tumour growth in situ in an autocrine or paracrine fashion once the tumour has been established. However, the efficacy depends on the tissue of origin. Increased expression of IGF1, IGF2 and IGF1R has been determined in a variety of neoplasias, including important geriatric malignancies such as brain tumours, mammary carcinoma, gastrointestinal cancer including pancreatic carcinoma and ovarian carcinoma .
Imprinting of the IGF2 gene has been shown to be relaxed in a variety of human neoplastic tissues . Biallelic expression of the IGF2 gene has been reported in Ewing sarcoma, rhabdomyosarcoma, Wilms' tumour, clear cell sarcoma, renal cell carcinoma, malignant glioma and a variety of gynaecological tumours and testicular neoplasms . Moreover, for some tumours, loss of imprinting seems to be a stage-specific event during carcinogenesis. Earlier data on experimental tumour formation in rodents suggest that loss of imprinting may be an important route to increased IGF2 expression in many types of tumours.
However, it is important to demonstrate that the transcription of a particular gene is complemented by subsequent translation to yield increased levels of bioactive protein. Squaring epidemiological and laboratory data has resulted in a notion that high levels of IGF1 or IGF2 protein and/or low levels of IGFBPs increase the risk of tumour development . This has been shown for a variety of cancers including mammary carcinoma, prostate carcinoma, lung cancer, colorectal carcinoma, endometrial carcinoma and urinary bladder cancer, thereby supporting the idea of a potential paracrine role for IGF2 in tumourigenesis. However, whereas a correlation between IGF1 or IGF2 protein levels and tumour progression could be consistently documented in some malignancies (e.g. colorectal, hepatocellular and pancreatic carcinoma), no consistent correlation was seen in others (e.g. mammary carcinoma) .
In the most clear-cut example , loss of imprinting of the IGF2 gene in the colonic mucosa remains an individual risk factor for developing colon carcinoma. Moreover, the same genetic alteration was also found in the peripheral lymphocytes of colon cancer patients. Viewed from a different standpoint, an increased incidence of precancerous colon carcinoma has been linked to increased levels of circulating IGF1, which suggests that the IGF system may play a role in the early stages of transformation and carcinogenesis.
Loss of imprinting in the IGF2 gene results in biallelic expression of IGF2. Adjacent to the promoter of IGF2, there is a differently methylated region (DMR). Biallelic expression has a strong correlation with hypomethylation of the DMR. Recently, the relationship between overexpression of IGF2, IGF2 loss of imprinting and the DMR was investigated, and the results suggested that two forms of abnormal IGF2 gene expression are associated with colorectal cancer .
In a recent case-control study where serum levels of IGF1, IGF2 and IGFBP-3 were measured and compared to the clinical stage of advanced colorectal adenoma , elevated serum levels of IGF2 had a significant association, but only in the highest quartile, and when further adjustment for IGF1 and/or IGFBP-3 levels was made, the association was weakened and no longer significant. However, the molar ratio of IGF2/IGFBP-3 was still associated with risk even after adjustment for IGF1/IGFBP-3 or IGF1. These results point to a relatively weak association between IGF2 and advanced colorectal adenoma [30,31].
The IGF1-growth hormone axis has long been suggested to affect cardiac structure and performance. In a general population, low levels of serum IGF1 were associated with higher prevalence of ischaemic heart disease and mortality .
There is also some evidence for IGF2 being involved in the development of cardiovascular disease. Large genetic studies have shown that the IGF2 genomic region is implicated in various common disorders such as the metabolic syndrome, type 2 diabetes and coronary heart disease . Animal models have shown that IGFs can delay infarction and generally improve postinfarction healing. However, IGF2 overexpression will result in gross abnormalities in the cardiac architecture including cardiomegaly, enlarged left ventricle, bradycardia and hypotension .
Atherosclerotic lesions consist of smooth muscle cells (SMCs), inflammatory cells, lipids and extracellular matrix. Inflammatory cells migrate from the blood stream and SMCs from adjacent tissue, forming a lesion. Mice genetically predisposed to atherosclerosis, in combination with homozygosity for a disrupted IGF2 allele, produced aortic lesions that were 80% smaller and contained 50% fewer proliferating cells in comparison with mice possessing non-disrupted IGF2 alleles .
This study showed that IGF2 clearly contributed to lesion formation by promoting cell differentiation via autocrine and paracrine signaling. The circulating levels of IGF2 did not affect the formation of atherosclerotic lesions, but increased local expression of IGF2 in SMCs resulted in focal intimal thickenings per se. The results in regards to migration of SMCs and lipid circulation were inconclusive. It is assumed then that IGF2 mainly acts locally by autocrine and paracrine actions in atherosclerosis .
This article has aimed to discuss the molecular characteristics of the IGFs in general, and IGF2 in particular, their receptors and binding proteins as well as their potential role in clinical medicine. A complex picture has emerged suggesting that their role in the pathogenesis and progression of disease is complicated as well as multifactorial. Although there is a rapidly increasing bulk of data from cellular and animal models, the ultimate evidence for a functional involvement of IGF2 overexpression and/or the IGF receptors in large disease groups such as cancer and cardiovascular disease is not overwhelmingly convincing.
The literature clearly indicates some obvious limitations of cellular or animal-based experimental systems when it comes to understanding human disease. However, it is also possible that the IGF system is far more complex than hitherto understood. There may well be alternative pathways that come into play that may bypass the ligand-receptor-binding protein system and may be equally powerful in driving human pathogenesis. However, the IGF system still offers a wide variety of interesting developments that may contribute to the future development of growth factor-based therapies.