SY Chun et al

CDX2 transcriptionally represses IGFBP-3, Supplementary Data

SUPPLEMENTARY DATA

Methods

Gene Expression Analysis:

LOVO and two clones of LOVOCDX2-/- cells were grown over-agar (anchorage-independent conditions)(Dang et al., 2006). Cells were harvested and total RNA extracted. Gene expression analyses on the samples were performed at the University of Michigan Comprehensive Cancer Center Affymetrix Core Facility. To control for clonal variation, two clones of LOVOCDX2-/- were compared to LOVO cells and to each other. Commercial high-density oligonucleotide arrays (GeneChip Human Genome U133A; Affymetrix, Inc.) were used, following protocols and methods developed by the supplier. The human genome U133A chip consists of 22,283 probe sets of 25-base long single-stranded DNA sequences, each representing a transcript. Single stranded cDNA were synthesized and converted into double stranded cDNA. An in vitro transcription (IVT) reaction was then carried out in the presence of biotinylated UTP and CTP to produce biotin-labeled cRNA. cRNA were then fragmented and hybridized to the test array, washed and stained with streptavidin-phycoerythrin, then scanned using GeneArray scanner.Images were analyzed following quality control parameters provided by Affymetrix. Once the data obtained from the test array is satisfactory, the samples were hybridized to the standard array for 16 hours at 45oC, washed and stained using the fluidics station, then scanned. The images were analyzed using Microarray Suite Software and comparison analyses were carried out according to the instructions provided by Affymetrix. The first step in data analysis was to assure overall quality of the raw data. For each sample, the distribution of the perfect match (PM) probes were graphed (log intensity as a function of density), and RNA degradation plot was generated (Probe Number as a function of Mean Intensity:shifted and scaled) (Chaib et al., 2003; Ray et al., 2004) . The shapes and slopes of all curves appeared similar, so the raw data were of high quality. The expression values for all the genes on each chip were computed using a robust multi-array average (RMA) and genes that appear not to be expressed were filtered out (Chaib et al., 2003; Ray et al., 2004). Comparisons were then made using a 2-fold cutoff. Data in the form of expression values and fold change were output into HTML files. Microsoft Excel data spreadsheets were generated fordata analysis.

Real time RT-PCR

LOVO and two clones of LOVOCDX2-/- cells were grown over-agar, RNA was extracted, reverse transcribed, and subjected to real time RT-PCR using primers for IGFBP-3 Forward 5’-ctctgcgtcaacgctagtgc-3’ and Reverse 5’-cggtcttcctccgactcact-3’ (Dang et al., 2006).

Western Blot Analyses

LOVO and two clones of LOVOCDX2-/- cells were grown over-agar and whole cell lysates extracted for western blot as previously described (Dang et al., 2006). CDX2 is a 37 kD band, IGFBP-3 is a 41-44 kilodaltons (kD) doublet, -tubulin, a control for loading, is a 43 kD band. CDX2 antibody was purchased from Biogenex (San Ramon, CA), IGFBP-3 antibody from R & D Systems (Minneapolis, MN), and -tubulin antibody from Sigma (St. Louis, MO). Antibody dilutions were as recommended by the manufacturers.

ELISA for IGFBP3

LOVO and two clones of LOVOCDX2-/- cells were grown in serum-free media over 48 hours and conditioned media harvested as previously described (Kansra et al., 2000). The media was subjected to IGFBP-3 ELISA per manufacturer directions (R & D Systems).

Sequencing

The IGFBP-3 gene in LOVO cells was sequenced as previously described (Dang et al., 2006), using previously described primers (Zou et al., 1998).

CHIP Assays

Cells were grown over-agar and subjected to CHIP assays (Active Motif, Carlsbad, CA) (Solomon et al., 1988). Chromatin was isolated, sheared, and immunoprecipitated with anti-CDX2 or anti-RNAPII antibody. Input chromatin was similarly treated, but not immunoprecipitated. All chromatins were then reverse cross-linked and the eluted genomic DNA subjected to PCR with primers flanking the regions of interest. All primers are detailed in Supplementary Data, Table 1. CDX2 antibody was purchased from Biogenex, RNAPII antibody from Active Motif.

Dual-Luciferase Assays

Four fragments of the IGFBP-3 promoter were PCR amplified and linked to the pGL3-Firefly luciferase reporter (Promega, Madison, Wisconsin). The –732 construct contains the putative CDX2 binding region D, which includes 3 putative CDX2 binding sites: site 1 at –694 (TTAAA), site 2 at –676 (GTAAATTAAA), and site 3 at –593 (ATAAA). These putative sites are denoted as diamonds on figure 2A of the manuscript. The –680 construct contains only sites 2 and 3, the –605 construct contains site 3 alone, and the –560 construct does not contain any of the sites. The –732 IGFBP-3-pGL3 construct was also subjected to site-directed mutagenesis of site 1(-7321), site 2 (7322), or both sites 1 and 2 (73212). For site-directed mutagenesis, Site 1 was changed from ATAAA to AGCGC; and site 2 was changed from GTAAATTAAA to GGCGCTGCGC, using the Quickchange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The pGL3-Firefly luciferase constructs were co-transfected with a CMV-Renilla construct into cells using Lipofectamine as previously described (Dang et al., 2001). Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega). These studies were performed on cells grown on plates at post-confluence, rather than over-agar. We were unable to recover viable cells under over-agar growth conditions after transient transfections. We suspect that the over-agar condition is too severe for newly transfected cells to grow in. Thus, transient transfections were performed on cells grown on plates at post-confluence. This approach is based on observations that under post-confluent culture conditions, CDX2 expression increases (Boulanger et al., 2005; Qualtrough et al., 2002), and is associated with increased proliferation (Suh & Traber, 1996) and decreased IGFBP-3 expression (figure S1, supplementary data).

Colony formation of cells treated with IGFBP-3 and IGF-1 analogues

Cells were grown in soft agar and treated with recombinant human IGFBP-3 (rh IGFBP-3 at 100 ng/ml), IGFBP-3 monoclonal antibody (mAb IGFBP-3 at 5 ug/ml), insulin-like growth factor 1 (IGF-1 at 50 ng/ml), and Long insulin-like growth factor 1 (LIGF-1 at 50 ng/ml). rh IGFBP-3 and mAb IGFBP-3 were from R & D Systems and IGF-1 and LIGF-1 were from GroPrep (Adelaide, Australia). The doses used were as previously published (Chan et al., 2005; Francis et al., 1992; Haluska et al., 2006; Kirman et al., 2004; Martin et al., 2003; Ricort & Binoux, 2004). Photographs were taken of 5 different areas from each triplicate plate. Image Pro Plus software was used to determine the plate density, which is the percent of the photographed area taken up by colonies, as previously described (Dang et al., 2006).

Table 1: Primers used for CHIP Assays in Figure 2
Primer Name / Genomic Locus / Forward / Reverse
Region A / -2011 to –1784 from initiation of IGFBP-3 promoter (227 bp).1 / 5’-gccgaacccagctctaattt-3’ / 5’-caggctgtggctacactcaa
Region B / -1723 to -1474 of IGFBP-3 promoter (249 bp). 1 / 5’-actccacccccacaaactct-3’ / 5’-ctgtattctcgatgcgttgg
Region C / -1041 to -808 of IGFBP-3 promoter (233 bp). 1 / 5’-tacgaagtcgccttcaaaaa-3’ / 5’-attaagagtggaccggcaag-3’
Region D / -705 to –555 of IGFBP-3 promoter (150 bp). 1 / 5’-cccctggggatataaacagc-3’ / 5’-gcattcgtgtgtacctcgtg-3’
MUC2+ / -238 to –91 of MUC2 promoter (includes CDX2 binding site) (147 bp). 2 / 5’-tgttggcattcaggctacag-3’ / 5’-tgccagatgatcaagaagaca-3’
MUC2- / -937 to –790 of MUC2 promoter (Non-CDX2 binding site) (147 bp). 2 / 5’-ctcgtagtctgggtcggtgt-3’ / 5’-aggtctccccagaactacctc-3’
GAPDH+ / RNA polymerase II binding site on the GAPDH promoter (166 bp). 3 / 5’-tactagcggttttacgggcg-3’ / 5’-tcgaacaggaggagcagagagcga-3’
1 Figure 2A.
2(Gum et al., 1997).
3 Active Motif, Carlsbad, CA.

Figure S1: CDX2 and IGFBP-3 expression under pre-confluent, post-confluent, and anchorage-independent conditions. Cells were grown under anchorage-dependent (pre-confluent and post-confluent) and anchorage-independent conditions. CDX2 and IGFBP-3 expression were measured by real time RT-PCR. Consistent with the literature, CDX2 expression was modestly increased in post-confluent compared to pre-confluent cells (Boulanger et al., 2005; Qualtrough et al., 2002); and further increased in anchorage-independent cells. This increase in CDX2 was associated with decreased IGFBP-3 (BP3) expression. Thus, CDX2 and IGFBP-3 expression are inversely correlated in the examined subset of colorectal cancer cell lines. Furthermore, when CDX2 was disrupted (LOVOCDX2-/-), IGFBP-3 expression was not only increased ~2-3-fold, but also failed to be down-regulated as cells were grown in post-confluent and anchorage-independent conditions. Taken together, these data would suggest that as colorectal cancer cells are challenged under post-confluent or anchorage-independent growth conditions, CDX2 is up-regulated and corresponds to down-regulation of IGFBP-3. This notion is supported by previous reports that associated CDX2 induction and enhanced proliferation in post-confluent intestinal epithelial cells (Boulanger et al., 2005; Qualtrough et al., 2002; Suh & Traber, 1996). Based on these data showing an inverse relationship between CDX2 and IGFBP-3 under post-confluent conditions, and our inability to grow transiently transfected cells over-agar, the dual-luciferase reporter assays in Figure 3 were performed on post-confluent cells. Pre-confluent cells were seeded on plates and grown until 50% confluence. Post-confluent cells were seeded on plates and grown until one day post-confluence. Over-agar cells were grown in liquid media over-agar as previously described (Dang et al., 2006). Primers for CDX2 were Forward 5’-aaggacgtgagcatgtaccc-3’ and Reverse 5’-gtggtaaccgccgtagtcc-3’. Primers for IGFBP-3 were Forward 5’-ctctgcgtcaacgctagtgc-3’ and Reverse 5’-cggtcttcctccgactcact-3’.

Supplementary Data References:

Boulanger, J, Vezina, A, Mongrain, S, Boudreau, F, Perreault, N, Auclair, BA et al. (2005). Cdk2-dependent phosphorylation of homeobox transcription factor CDX2 regulates its nuclear translocation and proteasome-mediated degradation in human intestinal epithelial cells. J Biol Chem280: 18095-107.

Chaib, H, MacDonald, JW, Vessella, RL, Washburn, JG, Quinn, JE, Odman, A et al. (2003). Haploinsufficiency and reduced expression of genes localized to the 8p chromosomal region in human prostate tumors. Genes Chromosomes Cancer37: 306-13.

Chan, SS, Twigg, SM, Firth, SM,Baxter, RC. (2005). Insulin-like growth factor binding protein-3 leads to insulin resistance in adipocytes. J Clin Endocrinol Metab90: 6588-95.

Dang, DT, Mahatan, CS, Dang, LH, Agboola, IA,Yang, VW. (2001). Expression of the gut-enriched Kruppel-like factor (Kruppel-like factor 4) gene in the human colon cancer cell line RKO is dependent on CDX2. Oncogene20: 4884-90.

Dang, LH, Chen, F, Ying, C, Chun, SY, Knock, SA, Appelman, HD et al.(2006). CDX2 has tumorigenic potential in the human colon cancer cell lines LOVO and SW48. Oncogene25: 2264-72.

Francis, GL, Ross, M, Ballard, FJ, Milner, SJ, Senn, C, McNeil, KA et al. (1992). Novel recombinant fusion protein analogues of insulin-like growth factor (IGF)-I indicate the relative importance of IGF-binding protein and receptor binding for enhanced biological potency. J Mol Endocrinol8: 213-23.

Gum, JR, Hicks, JW,Kim, YS. (1997). Identification and characterization of the MUC2 (human intestinal mucin) gene 5'-flanking region: promoter activity in cultured cells. Biochem J325 ( Pt 1): 259-67.

Haluska, P, Carboni, JM, Loegering, DA, Lee, FY, Wittman, M, Saulnier, MG et al. (2006). In vitro and in vivo antitumor effects of the dual insulin-like growth factor-I/insulin receptor inhibitor, BMS-554417. Cancer Res66: 362-71.

Kansra, S, Ewton, DZ, Wang, J,Friedman, E. (2000). IGFBP-3 mediates TGF beta 1 proliferative response in colon cancer cells. Int J Cancer87: 373-8.

Kirman, I, Poltoratskaia, N, Sylla, P,Whelan, RL. (2004). Insulin-like growth factor-binding protein 3 inhibits growth of experimental colocarcinoma. Surgery136: 205-9.

Martin, JL, Weenink, SM,Baxter, RC. (2003). Insulin-like growth factor-binding protein-3 potentiates epidermal growth factor action in MCF-10A mammary epithelial cells. Involvement of p44/42 and p38 mitogen-activated protein kinases. J Biol Chem278: 2969-76.

Qualtrough, D, Hinoi, T, Fearon, E,Paraskeva, C. (2002). Expression of CDX2 in normal and neoplastic human colon tissue and during differentiation of an in vitro model system. Gut51: 184-90.

Ray, ME, Yang, ZQ, Albertson, D, Kleer, CG, Washburn, JG, Macoska, JA et al. (2004). Genomic and expression analysis of the 8p11-12 amplicon in human breast cancer cell lines. Cancer Res64: 40-7.

Ricort, JM,Binoux, M. (2004). Insulin-like growth factor binding protein-3 stimulates phosphatidylinositol 3-kinase in MCF-7 breast carcinoma cells. Biochem Biophys Res Commun314: 1044-9.

Solomon, MJ, Larsen, PL,Varshavsky, A. (1988). Mapping protein-DNA interactions in vivo with formaldehyde: evidence that histone H4 is retained on a highly transcribed gene. Cell53: 937-47.

Suh, E,Traber, PG. (1996). An intestine-specific homeobox gene regulates proliferation and differentiation. Mol Cell Biol16: 619-25.

Zou, T, Fleisher, AS, Kong, D, Yin, J, Souza, RF, Wang, S et al. (1998). Sequence alterations of insulin-like growth factor binding protein 3 in neoplastic and normal gastrointestinal tissues. Cancer Res58: 4802-4.

1