Induction of vascular endothelial growth factor by insulin-like growth factor 1 in colorectal carcinoma.

Vascular endothelial growth factor (VEGF) is an angiogenic hormone that is produced by and supports the growth of many types of malignancies. The present study shows that insulin-like growth factor 1 (IGF-I), a mitogen that promotes the propagation of cancers through autocrine and paracrine mechanisms, increases the expression of mRNA for VEGF and production of VEGF protein by COLO 205 colon carcinoma cells. IGF-I also induces expression of VEGF mRNA in SW620, LSLiM6, and HCT15 colon carcinoma cells showing that this is a common response to IGF-I. Whereas IGF-I induced VEGF mRNA in each cell line examined (2.3-12-fold), it induced proliferation only in COLO 205 and LSLiM6 cells. Thus, the proliferative response induced by IGF-I and its ability to induce VEGF occur through distinguishable mechanisms. IGF-I increases the cellular content of VEGF mRNA by increasing the rate of transcription (5-fold after 4 h) and also by increasing the half-life of VEGF mRNA (0.6 ± 0.07 h in control cells to 2.0 ± 0.37 h in IGF-I-treated cells). Monoclonal antibody (αIR3) directed against the type 1 IGF receptor significantly attenuated the ability of IGF-I to promote expression of VEGF mRNA. Interestingly, by itself αIR3 acted as a weak agonist and induced a modest increase in the cellular content of VEGF mRNA. αIR3 also promoted tyrosine phosphorylation of the β subunit of the IGF-I receptor, and the magnitude of this response was comparable with that induced by IGF-I. These observations point to a nonlinear relationship between activation of the IGF-I receptor and induction of VEGF mRNA. Thus, in addition to its direct, growth stimulatory effect on transformed cells, IGF-I induces the expression of VEGF which can promote the progression of cancer by regulating the development of new blood vessels.

Angiogenesis, the development of new blood vessels by sprouting from pre-existing endothelium, is a significant component of a wide variety of biological processes including embryonic vascular development and differentiation, wound healing, and organ regeneration (1,2). Angiogenesis also contributes to the progression of pathologies that depend upon neovascularization, including tumor growth, diabetes mellitus, ischemic ocular diseases, and rheumatoid arthritis (1,2). A variety of growth factors are associated with angiogenesis including tumor necrosis factor, transforming growth factor b, and prostaglandin E 2 . However, these factors are believed to induce angiogenesis indirectly (1,2). Other growth factors important to angiogenesis, such as basic fibroblast growth factor and platelet-derived growth factor, are mitogens for a large number of cell types. Vascular endothelial growth factor (VEGF) 1 is unique, as it acts as a direct, specific mitogen for endothelial cells (3,4). In addition to its mitogenic activity, VEGF also stimulates glucose uptake and the production of tissue factor, collagenase, and plasminogen activators and their inhibitors by endothelial cells (3,4 and references therein). VEGF is also known as vascular permeability factor by virtue of its ability to enhance blood vessel permeability (5).
The temporal and spatial expression of VEGF and VEGF receptors in embryonic development and in the female reproductive system provide strong evidence that this mitogen plays an important role in developmental and hormonally regulated angiogenesis (6,9). Excessive production of VEGF by epidermal keratinocytes is associated with the vascular permeability and angiogenesis of healing wounds (10,11). VEGF levels in the vitreous are substantially elevated in diseases associated with ocular neovascularization, such as retinal vein occlusion or proliferative diabetic retinopathy (12,13). The vascular hyperpermeability and angioproliferation that characterize psoriatic skin (14) and the destructive synovial pannus in rheumatoid arthritis (15) are also accompanied by overexpression of VEGF. Thus, VEGF plays an important role in normal and restorative angiogenesis, as well as angiogenesis that accompanies and promotes pathologies.
VEGF also plays a major role in tumor-induced angiogenesis and can facilitate the transition from a dormant, small tissue mass into a rapidly expanding tumor (16). VEGF is produced by many types of tumors and accumulates in nearby blood vessels, its target of action (17,23). Interference with VEGF action in vivo either by administration of a neutralizing monoclonal antibody directed against VEGF or through the introduction of a dominant negative mutant VEGF receptor can effectively inhibit tumorigenesis in animal models of colon cancer or glioblastoma (18,24,25). These observations make it important to understand how the expression of VEGF by transformed cells is regulated. The present investigation was initiated to test the possibility that mitogens that directly support the proliferation of transformed cells through autocrine or paracrine mechanisms might additionally induce the elaboration of factors such as VEGF, which promote tumor outgrowth through the devel-opment of a blood supply. Here we report that insulin-like growth factor 1 (rhIGF-I), a mitogen that induces the growth of various malignancies and promotes the transformed phenotype (26,27), also induces the expression of an angiogenic factor, VEGF, in colon carcinoma cell lines. This identifies a new activity for this mitogen and shows that IGF-I can support the progression of cancer through direct and indirect mechanisms. We also show that the ability of IGF-I to promote colon carcinoma cell proliferation and expression of VEGF occur through distinguishable mechanisms.
Cell Proliferation Assay-To determine the effect of rhIGF-I on the proliferation of colon carcinoma lines, 1 ϫ 10 4 cells/ml were seeded into flat-bottomed 96-well microtiter plates. After overnight culture at 37°C, the media was aspirated, and fresh, serum-free medium containing 0 or 100 ng/ml rhIGF-I was added to five replicate wells. Cells were cultured for 72 h, and cell number was determined using the 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay (28). Data are expressed as the fold increase in cell number versus untreated serum-free cultures. The cultures were washed three times with ice-cold medium and lysed into 0.1 N NaOH for 30 min incubation at room temperature. Released radioactivity was measured in a gamma counter (LKB Diagnostics, Inc.). For saturation binding, triplicate, confluent 3.8-cm 2 cultures in binding buffer (Hank's balanced salt solution, HEPES, pH 7.4, 1 mg/ml bovine serum albumin) were incubated with various concentrations of 125 I-IGF-I (50 -500 pM) in the absence (total binding) or presence (nonspecific binding) of a 500-fold excess of IGF-I for 4 h at room temperature. Cultures were washed with binding buffer and solubilized into 0.1 N NaOH. The released radioactivity was then quantitated in a gamma counter. Specific binding was calculated as total binding minus nonspecific binding, and the data were analyzed by the method of Scatchard (29).
RNA Isolation and Northern Analysis-RNAzol (Tel-Test, Inc., Friendswood, TX) was used to extract total cellular RNA from confluent colon carcinoma cells grown in 9-cm 2 tissue culture plates. RNA samples (20 g) were resolved by electrophoresis through 1% agarose gels containing 2.2 M formaldehyde in PIPES buffer, transferred to nylon membranes (Amersham Corp.), and cross-linked to the membranes by UV irradiation. Blots were hybridized to the full-length human VEGF cDNA (30) for 1 h at 68°C in QuikHyb (Stratagene, La Jolla, CA). The DNA probes were labeled with [␣-32 P]dCTP to a specific activity of 1-2 ϫ 10 8 cpm/mg DNA using the random hexamer labeling method as described previously (17). Typically, 1-2 ϫ 10 7 cpm of 32 P-labeled probe was used for an 80-cm 2 filter in 7 ml of solution. Final washes were in 0.1 ϫ SSC, 0.1% SDS at 40°C. The blots were exposed to Kodak (Rochester, NY) XAR-2 film with an intensifying screen at Ϫ70°C for 24 h. To control for total RNA content, the blots were stripped and subsequently hybridized with a human GAPDH probe (1.3-kilobase human GAPDH cDNA, ATCC, Rockville, MD).
Ribonuclease Protection Analysis for IGF-I-A 143-base pair fragment of human IGF-I cDNA corresponding to nucleotides 330 -473 of the protein coding sequence (31) was subcloned into pBluescript KS (Stratagene, La Jolla, CA). A single-stranded 32 P-labeled cRNA riboprobe was synthesized and purified by gel electrophoresis. Total cellular RNA from colon carcinoma lines was extracted into buffer containing 4 M guanidinium thiocyanate. The integrity of each sample was validated by gel electrophoresis, and the quantity of RNA was determined spectrophotometrically. Ten ng of total RNA was hybridized with 1.5 ϫ 10 5 cpm probe at 42°C during an overnight incubation in 50 mM PIPES, 500 mM NaCl, 1 mM EDTA, and 80% formamide. The hybridization solution was treated with RNase A (10 mg/ml) ϩ RNase T 1 (2 mg/ml) in 10 mM Tris-Cl, pH 7.5, 300 mM NaCl, 5 mM EDTA at 37°C for 30 min and electrophoresed under denaturing conditions (6 M urea) in 6% polyacrylamide. The gel was dried and subjected to autoradiography for 4 days at Ϫ70°C. Molecular size markers consisted of an MspI digest of pBR322 end-labeled with [␣ 32 P]ATP by polynucleotide kinase according to the directions of the manufacturer (New England BioLabs).
Nuclear Run-on Transcription Assays-Nuclei were prepared from LSLiM6 and COLO 205 cells by the method of Greenberg and Ziff (32). Briefly, the cells were lysed in ice-cold nuclear extraction buffer (10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 , and 0.5% Nonidet P-40) and allowed to swell on ice for 5 min. The nuclei were pelleted at 500 ϫ g at 4°C for 5 min, washed once in nuclear extraction buffer, and repelleted. Washed nuclei were resuspended at 10 7 nuclei/200 nl in 40% glycerol, 5 mM MgCl 2 , 0.1 mM EDTA, and 50 mM Tris-HCl, pH 8.3.
The transcription assay was a modification of the method of Howard and Ortlepp (33). An equal volume of transcription buffer (10 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 300 mM KCl, 0.5 mM dithiothreitol) containing 1 mM each of ATP, CTP, and GTP and 125 mCi [␣-32 P]UTP was added to each reaction and incubated at 32°C for 40 min. Nuclear RNA was then isolated by the addition of 1 ml of RNAzol and 100 ml of chloroform. The mixture was placed on ice for 5 min and centrifuged at 12,000 ϫ g at 4°C for 15 min. The aqueous layer was removed and combined with an equal volume of isopropyl alcohol and incubated at Ϫ70°C overnight. Samples were centrifuged as above for 30 min. The pellet was washed with 100% ethanol, dissolved in sterile water, and used as a probe for hybridization. Prior to hybridization, RNA was heated at 65°C for 15 min.
Nitrocellulose membranes containing 2 ng of VEGF and GAPDH cDNAs were prepared. Two micrograms of DNA in 200 ml of TE buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA) were made 0.3 N in NaOH, incubated at 65°C for 30 min, cooled to 22°C, and then 1 volume of 2 M ammonium acetate was added. The DNA was slot blotted using a Schleicher & Schuell Minifold II manifold, and the membranes were baked at 80°C for 1 h. Filters were prehybridized in 50% formamide, 5 ϫ SSC, 0.2% SDS, 1 ϫ Denhardt's solution (0.02% Ficoll, 0.02% bovine serum albumin, 0.02% polyvinylpyrrolidone), and 20 mg/ml salmon sperm DNA at 42°C for 20 min. Filters were hybridized to the run-on products (10 7 cpm/ml) in 1 ml of hybridization solution at 42°C for 72 h. Standardization was achieved by adding the same amount of radioactivity to all hybridizations in a single experiment. The filters were washed in 2 ϫ SSC, 0.1% SDS at 22°C, again at 42°C, and finally in 0.1 ϫ SSC, 0.1% SDS at 42°C before autoradiography. Autoradiographs were analyzed by laser densitometry, and values are reported as relative increases in transcriptional rate of VEGF after normalizing to GAPDH transcriptional rates.
VEGF mRNA Stability-COLO 205 and LSLiM6 cells were grown to 90% confluence in 60-mm tissue culture dishes. At "0 time" actinomycin D was added with or without IGF-I at a final concentration of 5 mg/ml and 100 ng/ml, respectively. After varying times of incubation, total cellular RNA was isolated from the cells, and Northern blot analyses were performed. Densitometric scanning was performed with a Bio-Rad model GS-700 computing densitometer. Relative VEGF mRNA levels represent arbitrary units normalized to GAPDH mRNA levels.

RESULTS
In order to determine whether IGF-I affects the production of VEGF by human colorectal carcinoma, COLO 205 cells were incubated with IGF-I for various times. Expression of VEGF mRNA was then determined by Northern blotting. As shown in Fig. 1, A and B, IGF-I increased expression of VEGF mRNA such that it attained a level above the basal after 8 h. An enzyme-linked immunosorbent assay was used to determine whether the augmented expression of VEGF mRNA by IGF-Istimulated COLO 205 cells was succeeded by increased production of VEGF protein. As illustrated in Fig. 1C, after cellular stimulation with IGF-I, VEGF secretion increased progres-sively for 20 h, rising 2-fold above the basal level of elaboration. Subsequently, the amount of VEGF in cellular supernatants diminished somewhat, perhaps due to degradation.
The concentration dependence with which IGF-I increases expression of VEGF mRNA and the production of VEGF protein were next characterized. At concentrations of IGF-I as low as 10 ng/ml discernible increases in VEGF mRNA (Fig. 2, A and  B) and protein (Fig. 2C) were detectable. Higher concentrations of IGF-I further increased mRNA expression and protein production until these processes saturated after treatment of COLO 205 cells with 50 -100 ng/ml IGF-I.
The observations presented in Figs. 1 and 2 show that IGF-I regulates the expression of VEGF by COLO 205 cells. To test whether the effects of IGF-I were unique to this particular carcinoma cell line, three other lines were studied and compared with COLO 205. Observations illustrated in Fig. 3 showed that increased expression of VEGF mRNA is a common response of various colon carcinoma cell lines to IGF-I. Although the basal and IGF-I-stimulated levels of VEGF mRNA expression varied from one cell line to another, invariably the cellular content of this message was increased significantly.
Additional experiments were undertaken to characterize other properties of the cell lines used in these investigations. (Table  I). Most interesting is the observation that two of the cell lines proliferate in response to IGF-I and two do not; however, IGF-I was able to elevate expression of VEGF mRNA in all of the cell lines. Thus, the effects of IGF-I on these two cellular responses appear to be mediated through distinct mechanisms. Also apparent from Table I is that the responses of the carcinoma cell lines to IGF-I were not determined by the number or affinity of the expressed IGF-I receptors or by autocrine production of IGF-I, which was not detected in any of the cell lines (data not shown).
The mechanisms by which IGF-I increases expression of VEGF mRNA were then defined. Nuclear run-off assays were used to determine whether IGF-I increases the rate of transcription for VEGF mRNA. Fig. 4 shows that in COLO 205 cells IGF-I increased this process by about 5-fold within 4 h. Thereafter, the rate of transcription diminished although it remained FIG. 1. Induction of VEGF by IGF-I. COLO 205 cells were incubated with 100 ng/ml rhIGF-I for various times at 37°C. A, expression of VEGF mRNA was assayed by Northern analysis. Filters were also hybridized with a GAPDH probe to assess loading differences. Kb, kilobase pairs. B, the data in A were quantitated using a densitometer and are reported in arbitrary units relative to VEGF mRNA expression at time 0. Densities are corrected for loading differences by normalization for variances in GAPDH expression. C, the secretion of VEGF protein in triplicate cellular incubates was quantitated using an enzyme-linked immunosorbent assay.

FIG. 2. Concentration dependence of VEGF induction.
COLO 205 cells were incubated with various concentrations of IGF-I for 16 h at 37°C. A, expression of VEGF mRNA was assayed by Northern analysis. Filters were also hybridized with a GAPDH probe to assess loading differences. Kb, kilobase pairs. B, the data in A were quantitated using a densitometer and are reported in arbitrary units relative to VEGF mRNA expression at time 0. Densities are corrected for loading differences by normalization for variances in GAPDH expression. C, VEGF protein secretion in triplicate cellular incubates was quantitated using an enzyme-linked immunosorbent assay. significantly above the control level 8 h after initiation of cellular treatment with IGF-I. Additional nuclear run-off assays were conducted with the LSLiM6 cell line. Treatment of these cells with 100 ng/ml IGF-I increased expression of VEGF mRNA 1.8-fold after 16 h (data not shown). The higher basal level of VEGF mRNA expression in LSLiM6 relative to COLO 205 cells (see Fig. 3) made it more difficult to quantitate the stimulatory effect of IGF-I on VEGF transcription rates in the former relative to the latter cell line. For this reason, the more extensive experiments conducted with the COLO 205 cell line have been presented.
We next determined whether IGF-I affected the stability of VEGF mRNA. To accomplish this we first studied LSLiM6 cells in which the higher basal expression of VEGF mRNA would facilitate more reliable determinations of mRNA decay. Thus, LSLiM6 cells were stimulated with IGF-I for 16 h. Actinomycin D was then added, and cells were harvested at various times for Northern analysis of VEGF mRNA. As shown in Fig. 5 Finally, since insulin can stimulate VEGF mRNA expression (34) and IGF-I can activate both insulin as well as IGF receptors, experiments were conducted to determine whether IGF-I augments expression of VEGF mRNA in colon carcinoma cells via the type I IGFR. Observations presented in Fig. 6A show that a monoclonal antibody to the ␣ subunit of the type I IGF receptor (␣IR3), but not a control antibody, significantly attenuated the ability of IGF-I to promote expression of VEGF mRNA, indicating that the type I IGF receptor mediates at least part of the induction of VEGF by rhIGF-I. The control antibody alone had no effect on basal VEGF expression (data not shown). ␣IR3 proved to be a weak agonist and by itself induced a modest increase in the cellular content of VEGF mRNA (Fig. 6A). Its ability to induce increased expression of VEGF mRNA led us to test whether ␣IR3 would also promote phosphorylation of the ␤ subunit of its receptor. Indeed, this antibody did induce IGF-I receptor tyrosine phosphorylation and with a magnitude comparable with that of IGF-I (Fig. 6B). This observation suggests a disjunction between the ability of IGF-I and antibody to the IGF-I receptor to induce phosphorylation of the IGF-I receptor and increased expression of VEGF mRNA. These results appear consistent with the observations presented in Table I showing that the abilities of IGF-I to induce proliferation in carcinoma cell lines and VEGF mRNA are not coordinate properties. DISCUSSION IGF-I is a single chain polypeptide that is structurally homologous to proinsulin (35). While produced predominantly by the liver, from which it is released into the circulation, IGF-I is also elaborated by many other normal and transformed cell types. For this reason, IGF-I can act systemically as well as locally via endocrine, paracrine, and autocrine mechanisms. The primary receptor to which IGF-I binds is a heterotetrameric structure composed of two ␣and two ␤ subunits (36). The ␣ subunits contain the IGF-I binding domain and the ␤ subunits encompass an intrinsic protein tyrosine kinase that is important to signal transmission. While it is clear that IGF-I is important to the growth, differentiation, and metabolism of diverse normal cell types and various tissues, its role in the development and progression of cancer is still being defined. Recent observations support the view that IGF-I and the IGF-I receptor contribute to neoplastic transformation and the development of tumorigenicity (26,27). Consistent with this are observations showing that overexpression of the IGF-I receptor produces ligand-dependent malignant transformation of fibroblasts (37), but the presence of a dominant negative IGF-I receptor mutant inhibits tumorigenicity (38). Additionally, BALB/c3T3 cells overexpressing IGF-I and the IGF-I receptor proliferate in the absence of the added factors required by the parental cells (39), whereas experiments with cells from mice in which the gene for the IGF-I receptor has been disrupted by homologous recombination show that this receptor plays an obligatory role in transformation by the SV40 large T antigen (40,41). Also, rat glioblastoma cells expressing antisense RNA to IGF-I or the IGF-I receptor became non-tumorigenic (42)(43)(44). IGF-I may also play a role in promoting metastasis. In cultured human melanoma cells, the IGF-I receptor mediates motility that could enhance the potential for local and distant spread (45). The insulin-like growth factors also stimulate motility in human breast, bladder, and ovarian cancer cell lines, and the concentrations required for optimal migration are lower than those required for maximal growth stimulation (46). In addition to being important to transformation, tumorigenicity, and motility, IGF-I is an important mitogen for many tumor types. Thus, IGF-I is expressed by a variety of malignancies in vivo and in vitro and can support the proliferation of  some, but not all, transformed cell lines (47)(48)(49). Tumor cells expressing IGF-I receptors may enhance their own growth by synthesis of endogenous IGF-I, and such autocrine secretion of IGF-I probably contributes to the partial autonomy and rapid growth characteristic of malignant cells (50). The liver is frequently the first and the most common site of spread of colorectal cancer, and a paracrine role for hepatocyte-derived IGF-I in liver-specific metastasis has been proposed by Long and co-workers (51). These authors also demonstrated that by down-regulating expression of the type I IGF receptor, using an antisense construct, the ability of H-59 lung cancer cells to form liver metastases was lost, whereas the capacity of these cells to form primary tumors in the subcutis was unaffected (52). The present investigation adds a new and important activity to the IGF-I repertoire. IGF-I induces the expression of VEGF mRNA and augments the level of VEGF protein in superna-tants of colon carcinoma cell lines. This effect is mediated by the ability of IGF-I to increase the rate of transcription of VEGF mRNA and by stabilizing VEGF mRNA, thereby increasing its cellular half-life. Increased expression of VEGF mRNA was detected in carcinoma cell lines in which IGF-I induced a proliferative response and from carcinoma cell lines in which it did not. Thus, IGF-I-promoted proliferation and induction of VEGF mRNA occur through distinguishable mechanisms. These observations are of significance in suggesting that in addition to its ability to induce tumor growth directly, IGF-I may induce expression of angiogenic factors, such as VEGF, which provide the blood supply essential to the progressive growth of primary malignancies and to the development of metastatic disease.
Since the ability of IGF-I to induce expression of VEGF is unlikely to be restricted to colorectal carcinoma cells, our observations provide a basis for believing that such activity is likely to play a significant role in pathologies other than cancer. Elevated vitreous levels of IGF-I are associated with increased retinopathy during late-onset diabetes (53,54) and, based on the observations provided in the present report, may account in part for the concomitant presence of VEGF, which is believed to play a major role in mediating intraocular neovascularization in patients with ischemic retinal diseases (55,56). Such intraocular neovascularization which occurs in diabetic retinopathy, ischemic retinal-vein occlusion, and the retinopathy of prematurity can produce vitreous hemorrhage, retinal detachment, and neovascular glaucoma with subsequent loss of vision (13).
Cells at sites of wounding actively synthesize and release IGF-I which plays a significant role in the process of tissue repair including re-epithelization (57). New tissue formation begins with granulation tissue formation, which includes macrophage accumulation, fibroblast ingrowth, matrix deposition, and angiogenesis. The ability of IGF-I to induce expres- FIG. 4. Transcriptional regulation of VEGF production. Colon carcinoma cells (COLO 205) were treated with 100 ng/ml rhIGF-I for various times at 37°C. Transcription was then determined by nuclear run-on assays as described under "Experimental Procedures." Results are reported as the fold increase in VEGF mRNA relative to the control level in untreated cells. A representative result of three replicate experiments is shown.
FIG. 5. Stability of VEGF mRNA. Colon carcinoma cells were exposed to 100 ng/ml IGF-I or vehicle for 16 h at 37°C before addition of actinomycin D (5 mg/ml). Total RNA was extracted from the cells at the indicated times after actinomycin administration, fractionated by electrophoresis, and transferred to nitrocellulose. The Northern blots were hybridized to a cDNA probe for VEGF. Results were quantitated using a densitometer. To correct for differences in loading, the signal density of each RNA sample hybridized to the VEGF probe was divided by that hybridized to the GAPDH probe. The corrected density is plotted as a percentage of the zero time value against time. One representative of four replicate experiments is shown.
FIG. 6. Activities of a monoclonal antibody to the type I IGF receptor. A, cells were incubated with 100 ng/ml rhIGF-I in the absence or presence of ␣IR3 or a control antibody for 16 h at 37°C. Expression of VEGF mRNA was then determined by Northern blot analysis. B, the effect of ␣IR3 on phosphorylation of the IGF-I receptor was determined. Cells were incubated in the absence or presence of rhIGF-I and ␣IR3 for 10 or 30 min. IGF-I receptors were immunoprecipitated from cell lysates, fractionated by SDS-polyacrylamide gel electrophoresis, and transferred to nitrocellulose. The Western blots were probed with anti-phosphotyrosine antibody to assess phosphorylation of the IGF-I receptor.
sion of VEGF suggests that it may promote new tissue formation through its direct growth stimulatory activity and through induction of VEGF by keratinocytes (10), leading to new blood vessel development during wound healing.
A weak agonist monoclonal antibody to the IGF-I receptor (␣IR3) was able to significantly attenuate the ability of IGF-I to induce expression of VEGF. The demonstration that the extent of receptor phosphorylation induced by ␣IR3 was comparable with that induced by IGF-I reinforces the conclusion that various functions of IGF-I, such as induction of cell proliferation and VEGF expression, can occur through distinguishable processes. Since the type 1 IGF receptor contains numerous sites at which tyrosine phosphorylation may occur, it is possible that IGF-I and the agonist receptor antiserum induce discrete phosphorylations and thereby the activation of discrete signaling mechanisms. Consistent with this conclusion is the recent demonstration that mutation of tyrosine 1251 in the IGF-I receptor to phenylalanine has no effect on the ability of the receptor to transmit a mitogenic signal but abrogates the ability to induce cellular transformation (58).
In summary, we have shown that IGF-I induces VEGF mRNA and protein in colorectal carcinoma cell lines. The induction is mediated by a combined increase in the transcriptional rate of the VEGF gene and in the stability of the mRNA and is mediated at least in part through the type 1 IGF receptor. IGF-I-induced expression of VEGF by colorectal carcinoma may have a role in promoting neovascularization and the progression of cancer.