Insulin-like Growth Factor (IGF)-binding Protein-4 Inhibits Colony Formation of Colorectal Cancer Cells by IGF-independent Mechanisms

Insulin-like growth factor-binding proteins (IGFBPs) bind IGFs with high affinity. This interaction affects the half-lives of IGFs in the circulation (1) as well as the availability of IGFs for their receptors (2). However, there is also evidence that the IGFBPs possess IGF-independent actions. For instance, IGF-independent stimulation of cell growth and migration or inhibition of apoptosis have been described for IGFBP-1 (3), IGFBP-2 (4), IGFBP-4 (5), or IGFBP-5 (6), and IGF-independent growth-inhibitory effects have been shown for IGFBP-3 (7). IGFBP-4 has been demonstrated to exert inhibitory effects on cell proliferation under a wide variety of experimental conditions in vivo and in vitro (8). To rule out the biological potential of IGFBP-4 and to address the question of whether IGFBP-4 also exerts inhibitory actions during malignant growth of cells that are independent of IGF, we stably transfected IGF-sensitive (LS1034 and Isreco-2) or IGF-resistant (Isreco-1) colon cancer cell lines (9, 10) with mouse IGFBP-4 (mIGFBP-4) expression vectors. Consequences of IGFBP-4 overexpression for cell proliferation, anchorage-independent colony formation (as a measure of escape from anoikis), and invasion, as defined hallmarks of malignant progression of colon cancer cells, were investigated.

The LS1034 cell line (obtained from Dr. L. Suardet, ISREC, Epalinges, Switzerland) was cultured in DMEM/NUT Mix F-12 (Ham’s-F12) (Invitrogen, Karlsruhe, Germany). Isreco-1 and Isreco-2 cells, derived from a surgical specimen of a primary human ascending colon cancer (Dukes’ C, Class III; Ref. 11) and its corresponding liver metastasis, respectively, were cultured in DMEM (Invitrogen). Both media were supplemented with 5% (LS1034 and Isreco-1) or 10% (Isreco-2) FCS (PAA, Linz, Austria) and incubated at 37°C in a humidified atmosphere of 5% CO₂ in air. LS1034 cells were used between passages 210 and 227, Isreco-1 cells between passages 104 and 139, and Isreco-2 cells between passages 50 and 71. Murine IGFBP-4 (mIGFBP-4) cDNA (a kind gift from Dr. S. Drop, Rotterdam, the Netherlands) was cloned in either direction into the pCI-neo vector to express mouse IGFBP-4 sense or antisense RNA under control of the cytomegalovirus promoter. LS1034, Isreco-1, and Isreco-2 cells were transfected with 10 μg of the mIGFBP-4-pCI-neo sense and antisense construct or with the pCI-neo plasmid alone (mock transfectants) using the Qiagen Superfect Reagent (Qiagen, Hilden, Germany). G418 (Invitrogen) was used for selection at concentrations of 1 mg/ml for LS1034 and 2 mg/ml for Isreco-1 and Isreco-2 cells. Selection of resistant clones was performed in culture medium containing G418 over 2 weeks. Surviving colonies were picked and subcultured for the selection of stably transfected clones.

For measurement of proliferation, cells were seeded at 2 × 10³ cells/well (LS1034 and Isreco-1) and 1 × 10⁴ cells/well (Isreco-2) into 96-well flat-bottomed microtiter plates (Nunc, Wiesbaden, Germany) and then grown for 72 h in the absence or presence of recombinant human insulin-like growth factor II (Mediagnost, Tübingen, Germany). Proliferation was assessed by measuring the cleavage of the tetrazolium salt 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (Sigma, Munich, Germany) into the blue-colored formazan by mitochondrial enzyme activity during the last 2 h of culture. Formazan crystals were solubilized in DMSO, and the absorbance was measured using an ELISA reader at a test wavelength of 570 nm and a reference wavelength of 690 nm. For measurement of anchorage-independent growth, cells were trypsinized and resuspended in culture medium containing 0.9% methylcellulose (Fluka, Deisenhofen, Germany) and 5% FCS. Aliquots of 1 ml containing 2 × 10³ cells (LS1034 and Isreco-1) or 1 × 10⁴ cells (Isreco-2) were plated into 35-mm bacteriological dishes (Greiner, Nürtingen, Germany) in the absence or presence of human IGF-II. Colonies were counted under an inverted microscope after 14–16 days. Invasion of the tumor cells was assessed according to their migration through a 12-μm-pore size polycarbonate filter (Neuro Probe Inc., Gaithersburg, MD) coated with Matrigel (Collaborative Biomedical Products, Bedford, MA; 1 mg/ml), which was placed between the two compartments of an invasion chamber (Neuro Probe Inc.). The lower compartment was filled with Ham’s-F12 or DMEM including 10% FCS, and human IGF-II where indicated. Single-cell suspensions (1 × 10⁵ cells/well in a volume of 100 μl) of LS1034, Isreco-1, and their corresponding mIGFBP-4 transfectants were added on top of the filter in the upper compartment. After 16 h of incubation, the filters were removed, and the cells were fixed and stained with toluidine blue. Cells on the upper side of the filter were removed with a cotton swab (or were not removed for determination of total cell fraction). Filters were cut out, placed in 200 μl of 1% SDS solution, and incubated for 60 min at 37°C to fully solubilize the toluidine blue stain. The absorbance of the dye was determined at 620 nm. The fraction of invaded cells was calculated as follows:

Total RNA was prepared using the Tri-Pure isolation reagent (Roche Diagnostics, Mannheim, Germany). cDNA was produced from 2.5 μg of total RNA by incubation with 20 units of Moloney murine leukemia virus reverse transcriptase (Invitrogen) for 1 h at 37°C in a reaction volume of 20 μl. PCR analyses were carried out in a 20-μl reaction solution using the Qiagen Taq Polymerase kit (Qiagen) containing 1 μl of cDNA, 50 μm deoxynucleotide triphosphates, and 0.1 μm of both sense and antisense primers. The amplification was performed as follows: samples were heated at 94°C for 4 min, followed by 36 cycles at 94°C for 1 min, 60°C for 1 min, and 72°C for 2 min. After a final extension at 72°C for 10 min, samples were analyzed by agarose gel electrophoresis.

Protein extracts were analyzed by immunoblotting and Western ligand blotting to measure IGFBP-4 expression and functionality. Amounts of the conditioned media were adjusted to the protein amount of the corresponding cell lysate. Then proteins were precipitated from conditioned media of LS1034, Isreco-1, Isreco-2, and the transfectants using 10% ice-cold trichloroacetic acid. The whole, precipitated proteins were resuspended in buffer A (5% SDS, 0.1 m NaOH), diluted 1:5 with sample buffer [312.5 mm Tris (pH 6.8), 50% (w/v) glycerol, 5 mm EDTA (pH 8), 1% (w/v) SDS, and 0.02% bromphenol blue], separated by SDS-PAGE, and transferred to polyvinylidene difluoride membranes (Millipore, Schwalbach, Germany). The membranes for ligand blot analysis were blocked with 1% fish gelatin for 2 h and incubated with ¹²⁵I-labeled IGF-II (5 × 10⁵ cpm/ml; Amersham Biosciences). Proteins binding IGF-II were detected on a Storm 860 PhosphorImager (Amersham Biosciences). The membranes for immunoblot analysis were blocked with 3% milk powder (Töpfer, Dietmannsried, Germany) for 24 h and were incubated with a rabbit anti-IGFBP-4 antiserum (Upstate Biotechnology, Lake Placid, NY). A peroxidase-coupled polyclonal goat anti-rabbit IgG was used as secondary antibody (Dianova, Hamburg, Germany). Enhanced chemiluminescence (ECL; Amersham Biosciences) was used for detection of the IGFBP-4 protein. Bands were visualized on a Hyperfilm MP (Amersham Biosciences) with intensifying screens.

Data are given as the mean values ± SE and were analyzed using Student’s t test (Prism; Graph Pad, San Diego, CA).

Endogenous IGFBP-4 mRNA was detected in all cell lines tested (not shown), whereas mIGFBP-4 mRNA was present only in cytomegalovirus-mIGFBP-4 transfectants (Fig. 1,A). At the protein level, IGFBP-4 was found in the conditioned media of LS1034, Isreco-1, and Isreco-2 cells but was present at higher amounts in mIGFBP-4 transfectants (Fig. 1,B). The functionality of overexpressed mIGFBP-4 was demonstrated by its capacity to bind ¹²⁵I-labeled IGF-II in Western ligand blots (Fig. 1 C).

Proliferation of LS1034 cells was significantly reduced by overexpression of mIGFBP-4 (Fig. 2,A, left panel). The proliferation of all sense clones was significantly stimulated by the addition of exogenous IGF-II (P < 0.05), although to a lower extent if compared with parental, mock-, or antisense-transfected cells (Fig. 2,A, left panel). Expression of IGFBP-4 antisense RNA increased proliferation of LS1034 transfectants (Fig. 2,A, left panel). Proliferation of IGF-independent Isreco-1 cells was not affected by overexpression of mIGFBP-4 nor by addition of exogenous IGF-II (Fig. 2,A, right panel). Exogenous IGF-II significantly stimulated proliferation of Isreco-2 cells, whereas the growth rate of mIGFBP-4-overexpressing clones was comparable with that of parental cells (Table 1A). However, the response to exogenous IGF-II was reduced in mIGFBP-4 transfectants (Table 1A), which is consistent with an IGF-dependent mechanism.

Anchorage-independent colony formation was consistently decreased in mIGFBP-4 clones of all three cell lines. The cloning efficiency of these clones was between 20 and 50% when compared with the values of the parental cell lines. In contrast, antisense and mock-transfected clones produced colony numbers similar to untransfected parental cells (Fig. 2,B; Table 1B). Interestingly, the addition of exogenous IGF-II did not revert the inhibitory effect of mIGFBP-4 in all three cell lines (Fig. 2,B; Table 1B), indicating that this effect of mIGFBP-4 is independent of IGF.

The fraction of LS1034 cells that migrated through Matrigel-coated membranes was significantly reduced in the mIGFBP-4-overexpressing clones (Fig. 2,C, left panel). Exogenous IGF-II was able to stimulate invasion dose dependently in all transfectants and in the parental cell line (Fig. 2,C, left panel). The invasive capacity of Isreco-1 cells was not affected by IGFBP-4 transfection nor by addition of exogenous IGF-II (Fig. 2 C, right panel). Invasion could not be determined in Isreco-2 because of the very low migration rates of these cells through the Matrigel barrier (data not shown).

Antimitogenic effects of IGFBP-4 have been demonstrated in different cellular systems, such as human fibroblasts (12), osteoblasts (13), neuronal cells (14), and in human prostate cancer cells (15). Notably, both IGF-dependent and -independent mechanisms have been suggested for the antimitogenic effects of IGFBP-4 (16). In the present study, we investigated whether colorectal cancer cells, besides cell proliferation and as well as later stages of the malignant progression, such as escape from anoikis or invasion, are affected by IGFBP-4 and whether potential alterations of these parameters are dependent on IGF binding. Thus, we stably transfected IGF-sensitive (LS1034 and Isreco-2) and -resistant (Isreco-1) colon carcinoma cells and assessed proliferative activity, colony formation, and invasive capacity. IGFBP-4 was detected constitutively at the protein level in LS1034, Isreco-1, and Isreco-2 colon cancer cells, although at a low level. Transfection with mIGFBP-4 expression vectors increased the amount of the 24-kDa IGFBP-4 in all three cell lines markedly. Western blot analyses revealed a single 24-kDa band in Isreco-1 and Isreco-2 cell lines, which represents the non-glycosylated protein (17). In contrast, in conditioned media from LS1034 mIGFBP-4 transfectants, additional 28- and 18-kDa bands appeared, corresponding to the glycosylated IGFBP-4 (17) and most likely to an IGFBP-4 cleavage product, respectively. Western ligand blot analysis revealed the functionality of overexpressed IGFBP-4 and its glycosylated form, whereas the cleavage product was unable to bind IGF-II. The growth-inhibitory activities appear to be mediated predominantly if not exclusively by the 24-kDa form. This form was the only one secreted by Isreco-1 and Isreco-2 cells. Nevertheless, colony formation in both cell types was also dramatically reduced. In addition, this isoform effectively abrogated growth stimulation by exogenous IGF-II in Isreco-2. The rather weak signal at 28 kDa detected in Western ligand blots in LS1034 suggests that this isoform may contribute, to some extent, to the effects seen in this cell line.

At a functional level, overexpression of mIGFBP-4 strongly inhibited proliferation of LS1034 cells in which IGF-II functions as an autocrine growth factor (9). In addition, IGF-II strongly counteracted the inhibitory effect of mIGFBP-4 in all transfectants of LS1034, although to a lower extent than in mock-transfected, antisense-transfected, or parental LS1034 cells, similar to results obtained with prostate cancer cells (15). In Isreco-2 cells that do not express endogenous IGF-II (10), mIGFBP-4 did not affect the proliferation rate. However, the response to exogenous IGF-II was significantly inhibited. These observations strongly suggest that the inhibitory effect of mIGFBP-4 on proliferation is IGF dependent. Because of the not-absolutely-complete reversion seen in LS1034 transfectants, we cannot totally exclude the possibility that a minor part of the observed response may be contributed in an IGF-independent manner.

Anchorage-independent colony formation, a marker for progressed cellular transformation, was shown previously to be affected by the IGF-I receptor pathway, because IGF-I receptor blocking antibodies severely inhibited colony formation in LS1034 cells (9). Nevertheless, addition of exogenous IGF-II failed to stimulate colony formation of mIGFBP-4-overexpressing clones of all three cell lines, despite the fact that both LS1034 and Isreco-2 cells are IGF sensitive. Therefore, inhibition of colony formation of colorectal carcinoma cells by IGFBP-4 appears to be clearly mediated by an IGF- independent mechanism.

Notably, if compared with controls, expression of antisense IGFBP-4 mRNA dramatically increased colony formation in Isreco-1 cells, further supporting a potent inhibitory and IGF-independent function of IGFBP-4 for colony formation.

Cell migration, which is affected through IGF-dependent and -independent mechanisms by, e.g., IGFBP-1 (18), was severely impaired by overexpression of mIGFBP-4 in LS1034 but not in Isreco-1 cells. Our data thus indicate IGF-dependent effects of IGFBP-4 on the migration of colon cancer cells, because IGFBP-4 had no significant effects on IGF-insensitive Isreco-1 cells, whereas exogenous IGF-II stimulated cell migration and partially antagonized the negative effects of IGFBP-4 overexpression in LS1034 cells.

In conclusion, our studies demonstrate that later stages of malignant progression in colorectal cancer cells are markedly influenced by IGFBP-4. Anchorage-independent colony formation as a marker for the escape from anoikis after loss of contact to the underlying surface was significantly reduced by IGFBP-4 via mechanisms independent of the functionality of the IGF/IGF receptor pathway and independent of IGF-II binding. In contrast, inhibitory activities of IGFBP-4 on cell proliferation and invasion of colon cancer cells also depend on its IGF-scavenging activities. This is the first study dissecting IGF-dependent and IGF-independent mechanisms of IGFBP-4 in colon tumorigenesis.

Special thanks go to Anneliese Helfrich for assistance in screening the clones from the different cell lines.