This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Reichling, T. Articles by Groden, J. Search for Related Content PubMed PubMed Citation Articles by Reichling, T. Articles by Groden, J.

Transcriptional Profiles of Intestinal Tumors in Apc^Min Mice are Unique from those of Embryonic Intestine and Identify Novel Gene Targets Dysregulated in Human Colorectal Tumors

¹ Department of Molecular Genetics, Biochemistry and Microbiology, and ² Howard Hughes Medical Institute, University of Cincinnati College of Medicine and Divisions of ³ Developmental Biology and Pediatric Informatics and ⁴ Pathology, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio

Requests for reprints: Joanna Groden, Department of Molecular Genetics, Biochemistry and Microbiology, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0524. Phone: 513-558-0088; Fax: 513-558-2794; E-mail: joanna.groden{at}uc.edu

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
The adenomatous polyposis coli (APC) tumor suppressor is a major regulator of the Wnt signaling pathway in normal intestinal epithelium. APC, in conjunction with AXIN and GSK-3ß, forms a complex necessary for the degradation of ß-catenin, thereby preventing ß-catenin/T-cell factor interaction and alteration of growth-controlling genes such as c-MYC and cyclin D1. Inappropriate activation of the Wnt pathway, via Apc/APC mutation, leads to gastrointestinal tumor formation in both the mouse and human. In order to discover novel genes that may contribute to tumor progression in the gastrointestinal tract, we used cDNA microarrays to identify 114 genes with altered levels of expression in Apc^Min mouse adenomas from the duodenum, jejunum, and colon. Changes in the expression of 24 of these 114 genes were not observed during mouse development at embryonic day 16.5, postnatal day 1, or postnatal day 14 (relative to normal adult intestine). These 24 genes are not previously known Wnt targets. Seven genes were validated by real-time reverse transcription-PCR analysis, whereas four genes were validated by in situ hybridization to mouse adenomas. Real-time reverse transcription-PCR analysis of human colorectal cancer cell lines and adenocarcinomas revealed that altered expression levels were also observed for six of the genes Igfbp5, Lcn2, Ly6d, N4wbp4 (PMEPA1), S100c, and Sox4.

Key Words: adenoma • APC • gastrointestinal cancer • mouse models of cancer • transcriptional profile

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Colorectal cancer is the third most common cancer in men andwomen and accounts for 11% of all cancer deaths. Whereas the 5-year survival rate is extremely favorable when detected at a localized stage (90%), most colorectal cancers are either locally or distantly invasive at diagnosis, limiting treatment options and lowering survival rates (1). Clearly, a more comprehensive view of the molecular events associated with colorectal tumorigenesis is needed to detect and evaluate tumors earlier in order to treat mid- to late-stage colorectal tumors more effectively.

The Wnt signaling pathway regulates cell fate and proliferation in normal cells, particularly during embryonic development. Activation of Wnt signaling occurs when secreted Wnt ligands activate transmembrane frizzled receptors through a mechanism known to involve disheveled proteins. This ligand-receptor engagement inhibits GSK-3ß–mediated phosphorylation of ß-catenin. Thus, ubiquitin-dependent destruction of ß-catenin ceases, and ß-catenin levels rise in the cytoplasm and nucleus. ß-catenin can interact with members of the T-cell factor (TCF) transcription factor family (TCF1, LEF1, TCF3, and TCF4) and be transported to the nucleus, where the complex alters the expression of target genes containing cis-acting, TCF-binding elements (2). Genes whose transcription is altered by TCF/ß-catenin include those that are growth promoting, such as c-MYC and cyclin D1, as well as the matrix-remodeling enzyme matrilysin (MMP7), the nuclear receptor PPAR, gastrin, connexin 43, WISP1, WISP2, fibronectin, Nr-CAM, AXIN2, fra-1, c-jun, and the recently discovered transcription factor ITF-2 (1, 3–10). In the absence of a Wnt signal, GSK-3ß, AXIN, and the APC tumor suppressor form a complex that can degrade ß-catenin via the ubiquitin-proteosome pathway (11).

Inappropriate activation of the Wnt pathway is associated with the development of cancer. Disruption of the Wnt regulatory complex by mutation of ß-catenin (CTNNB1), AXIN2, APC, or GSK-3ß leads to ß-catenin accumulation and TCF/ß-catenin–mediated transcriptional alteration of target genes (1). The NH₂ terminus of ß-catenin contains four highly conserved serine/threonine residues, S33, S37, T41, and S45 (12), that are phosphorylated by GSK-3ß and are required for recognition and degradation of ß-catenin by the E3 ubiquitin complex containing ß-TRCP. Missense mutations of these residues have been identified in numerous tumors leading to nuclear localization of ß-catenin, a marker for aberrant Wnt signaling (2, 11, 13). Approximately 85% of human colorectal tumors carry biallelic-inactivating mutations of APC; 7% to 15% of colorectal tumors lacking a mutation in APC carry mutations in ß-catenin (CTNNB1; ref. 11). Mutations in AXIN2 have been observed in some hepatocellular tumors which lack CTNNB1 and APC mutation (13), as well as in some colorectal cancers defective in DNA mismatch repair (14). Inactivating AXIN2 mutations prevent binding of AXIN to GSK-3ß or ß-catenin (13).

Although genetic alteration of CTNNB1, AXIN, or GSK-3ß result in aberrant Wnt signaling, the most common mechanism to activate the Wnt pathway involves APC. Most APC mutations are nonsense mutations which delete part or all of the central domain of APC containing the 15- and 20-amino-acid repeats necessary and sufficient to bind and down-regulate ß-catenin (11, 13). COOH-terminal truncation of five or more of the seven 20-amino-acid repeats in APC prevents ß-catenin degradation and results in TCF/ß-catenin–mediated transcription in intestinal epithelial cells. Although APC mutation is an early and common event in colorectal cancer development, the extensive molecular consequences of APC inactivation are not completely understood. One approach to this problem is to ask which genes are altered in the intestinal epithelium of a mouse model of gastrointestinal tumor formation in response to activated Wnt signaling in vivo. Such information would identify new biomarkers and therapeutic targets for human colorectal cancer.

In this study, gene expression profiles were examined in adenomatous gastrointestinal tumors of the Apc^Min/+ mouse in response to somatic Apc mutation and consequent activated Wnt signaling. Incyte Gem1 cDNA microarrays were used to obtain these transcription profiles and those of normal and embryonic tissues along the anterior-posterior axis of the gastrointestinal tract. Previously unidentified Wnt target genes were obtained from a screen of more than 8,600 genes and validated by other methods to confirm their changes in expression during tumor formation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Tissue Samples for Microarray Analysis. Apc^Min/+ mice were obtained from The Jackson Laboratory and maintained on the C57BL/6J background. Fourteen adenomas (duodenum, n = 5; proximal jejunum, n = 5; and distal colon, n = 4) were harvested from 10 Apc^Min/+ mice (male or female) between 5 and 7 months. Normal duodenum, jejunum, and colon tissues were also harvested. In addition, seven normal gastrointestinal tissues (stomach, duodenum, jejunum, ileum, cecum, proximal colon, and distal colon) from 6- to 8-week-old male C57BL/6J mice and three developing whole intestinal samples (embryonic day 16.5, postnatal day 1, and postnatal day 14.5) were included in our analyses as part of an institutional database profiling generalized mouse gene expression (15). This database included various control tissues and used a common reference RNA collected from C57BL/6J whole postnatal day 1 (P1) mice.

RNA Extraction. Animals were euthanized and tissues immediately dissected and flash frozen in liquid nitrogen. Total RNA was extracted from all tissues and polyadenylated RNA was isolated from 150 µg of total RNA using Oligotex (Qiagen, Chatsworth, CA). Each polyadenylated RNA was purified twice using a GenElute kit (Sigma-Aldrich, St. Louis, MO), whereas RNase degradation and genomic contamination were assessed using a Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA).

Fluorescent Probe Preparation and Microarray Hybridization. cDNA was synthesized using reverse transcription with an 18-mer oligo-dT primer in the presence of fluorescent dUTP using the CyScribe kit (Amersham Pharmacia Biotech, Bjorkgatan, Sweden). All cDNA targets generated from database test tissues were Cy5-dUTP labeled and competitively hybridized against Cy3-dUTP–labeled reference cDNA from whole mouse P1 to Incyte GEM1 cDNA microarray containing 8,638 sequence-verified Image EST clones, ranging in size from 500 to 5,000 nucleotides (Incyte Genomics, St. Louis, MO). Cy5-dUTP– and Cy3-dUTP–labeled cDNAs were mixed and purified with a QiaQuick column (Qiagen) before competitive hybridization. For each RNA sample, technical replicates were hybridized. Following initial Lowess dye intensity per chip normalization using the P1 reference, gene ratios were normalized across all samples included in the database or in the sample series for more restricted sample intercomparisons.

Microarray Data Analyses. Incyte GEM1 mouse microarrays contain 8,638 cDNA clones corresponding to identified genes. Less than a 2-fold variation across the 192 control genes permitted interarray comparisons. Target signal intensity and background noise were calculated using GenePix 3.0 software following scanning by the GenePix4000B scanner (AxonInstruments, Inc., Union City, CA). Individual spot quality was assessed with GEMTools 2.5.0 software (Incyte Genomics). For unannotated EST sequences included in the probe lists, the Celera Discovery System (Celera Genomics, Norwalk, CT) was used to identify genes in the mouse genome. Gene prediction was done using the Otto method (16). We were able to identify a known or predicted gene for 7 of 26 unknown ESTs.

Second-stage data analysis was done with GeneSpring 6.1 software (Silicon Genetics, Redwood City, CA). To make comparisons between multiple chips, normalizations were applied to raw data. To identify genes altered during tumorigenesis, the 8,638 probe hybridizations were filtered using two approaches: (1)a statistical filter (Welch ANOVA corrected for multiple testing using the Benjamini Hochberg False Discovery Rate, P<0.005) comparing mean expression levels between tumors and normal adult tissue, and (2) a filter method based on fold changes in gene intensity when comparing one sample to another. Only those genes showing a 1.8-fold increase or decrease in expression were selected for further analysis. All lists were then combined and sorted by probe name. This master database was then mined for genes expressed in at least 50% of the 14 tumors irrelevant of location, genes exhibiting site-specific expression in at least 60% of the duodenum and jejunum tumors and 75% of distal colon tumors.

Real-time reverse transcription-PCR Analysis. Normal tissue from wild-type C57BL/6J and adenomas from Apc^Min/+ mice ages 5 to 7 months was harvested as described above. Human colorectal samples were collected from patients undergoing surgery at University Hospital in Cincinnati. Subsequent total RNA extraction was done using TriZol (Invitrogen, San Diego, CA). Using oligo-dT_12-18 cDNA was synthesized with Superscript II reverse transcriptase (Invitrogen). SYBR Green real-time PCR was done using an ABI7700 (Applied Biosystems, Foster City, CA). Each point of the standard curve that was generated represented the average of triplicates. All experimental samples from tumor and other normal tissues were expressed as an n-fold difference relative to the standard. Experimental target quantities were normalized to the endogenous GAPDH control or ß-actin.

In Situ Hybridization. Whole intestines from wild-type C57/BL6 and Apc^Min/+ mice ages 5 to 7 months were removed and incubated overnight in DEPC-treated 4% paraformaldehyde in PBS at 4°C. Intestines were transferred to PBS with 30% sucrose and incubated at 4°C prior to embedding in M-1 for subsequent frozen sectioning. ³⁵S-labeled sense and antisense riboprobes were synthesized by in vitro transcription from linearized templates containing cloned cDNA fragments for Expi, Lcn2, N4wbp4, and Sox4 in the pT7T3D-Pac plasmid (Incyte). In vitro transcription with T3 polymerase was done on EcoRI-digested plasmids to produce the antisense probes ranging in size from 500 to 800 bp; in vitro transcription with T7 polymerase and NotI-digested templates generated sense probes to assess nonspecific hybridization. In situ hybridization of the sections was done as previously described (17).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Mice that carry the Apc^Min/+ allele, a nonsense point mutation at nucleotide 2549 in the Apc gene, develop multiple intestinal adenomas and have been used extensively as a mouse model of the human colorectal cancer syndrome, familial adenomatous polyposis coli (18). Interestingly, adenomas in Apc^Min/+ mice occur primarily in the small intestine, whereas tumors in individuals with familial adenomatous polyposis coli are generally restricted to the colon and rectum. To determine the gene expression profiles of adenomas along the proximal-distal gastrointestinal tract axis, we harvested normal tissues and 14 tumors from multiple sites of tumor development in Apc^Min/+ mice from the duodenum, proximal jejunum, and distal colon. Polyadenylated mRNAs were prepared from these tissues, and gene expression profiles generated using Incyte Gem1 mouse cDNA microarrays. Gene expression profiles were then generated from normal stomach, duodenum, jejunum, ileum, cecum, proximal colon, and distal colon from pubescent male C57BL/6J mice, as well as 16.5 dpc embryonic, day 1 newborn, and day 14.5 juvenile whole intestine.

Our experiments were designed with the goal of identifying genes dysregulated in the early stages of tumor progression by comparing gene expression patterns in adenomas to those of normal adult and developing whole intestine. Two separate methods, the Welch ANOVA and a fold-change analysis, were used to compare the expression profiles of adenomas with normal tissue. The Welch ANOVA method, corrected for multiple testing using the Benjamini Hochberg False Discovery Rate, identified 428 genes with statistically significant changes in mean expression levels between normal adult tissue and adenomatous tissue from the intestine (P < 0.005). Inaddition, 237 genes were identified with statistically significant changes in mean expression levels between normal adult tissue (distal colon) and adenomatous tissue from the large intestine (P<0.005). Lastly, after comparing both large and small intestinal adenomas to normal gastrointestinal tissue, 51 genes displayed statistically significant expression changes. The statistical filter using the Welch ANOVA method detects differences <1.8-fold between normal tissue and tumors if variation in expression is low. However, the Welch ANOVA lacks the sensitivity to discover potentially interesting genes with high variation in expression among tumors or to extract site-specific expression patterns. Therefore, a more intensive analysis was used to identify genes based on fold change in expression when comparing each tumor to its appropriate normal adult tissue. Of the 8,638 probes on the array, a total of 2,273 showed a 1.8-fold difference in at least one of the 14 tumors when compared with normal tissue on a log₂ scale (1,001 overexpressed and 1,272 underexpressed); 89 genes were differentially expressed in 7 of 14 tumors (51 overexpressed and 38 underexpressed).

Using a Venn diagram, the lists generated from the two types of analyses were compared. Both the Welch ANOVA and the fold-change methods detected 26 genes differentially expressed in small and large intestinal tumors (17 up and 9 down). Unique lists were also detected by each analysis. The fold-change method detected 63 genes that were differentially expressed in at least 50% of the tumors, whereas 25 genes were detected by the Welch ANOVA method. These three lists, when combined, contained 114 genes which were subsequently hierarchically clustered (Fig. 1). Surprisingly, none of the known Wnt targets spotted on the GEM1 array were contained in this list of 114 genes. Known Wnt targets contained on the GEM1 chip, along with their tumor intensity values, were normalized to the average intensity of normal tissue (Table 1). Some Wnt targets, such as cyclin D1 and c-MYC known to be slightly overexpressed in colon cancer, were excluded from our list of 114 due to the stringency of the filter methods requiring genes to be differentially expressed up or down in at least 50% of the tumors by 1.8-fold.

View larger version (78K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Hierarchical cluster shows 114 genes differentially expressed in 50% of ApcMin gastrointestinal adenomas relative to normal small and large intestine tissues. Red, genes are overexpressed; blue, genes are underexpressed; yellow, genes are unchanged.

View larger version (114K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Expi, Lcn2, N4wbp4, and Sox4 are overexpressed in gastrointestinal adenomas from murine small and large intestine (A). In situ hybridizations using riboprobes for Expi (A-D), Lcn2 (E and F), N4wbp4 (G and H), and Sox4 (I-L). Right, positive signal identified by bright white grains in the dark field–illuminated images; some are highlighted by white arrows. Expression of Expi, Lcn2, N4wbp4, and Sox4 is confined to the atypical epithelium of the adenomas. No expression of these genes is observed in the normal glandular epithelium. A, B, I and J, sections from adenomas of the small intestine. C, D, G, H, K and L, sections from adenomas of the large intestine. A-L, magnification x100.

Tumor-specific changes in the expression of six genes of interest were validated using real-time RT-PCR analysis with cDNAs from aset of human colorectal adenomas, adenocarcinomas, and established cancer cell lines. Human homologues of Lcn2, N4wbp4, S100c, and Sox4 from the list of 24 were evaluated to determine if expression was similarly altered in human tumors. IGFBP5, from the list of 114, was also included in this study due to conflicting reports in the literature about its role in cancer as either growth promoting or growth inhibitory (19). Expi had no known human homologue and therefore was not included in these analyses. Another gene, LY6D, from the list of 24 was examined in its place.

Expression of LCN2, PMEPA1, and S100C was increased in one human adenoma and all nine human adenocarcinomas regardless of tumor stage; expression of PMEPA1 was increased in only one cell line each (Table 5). Expression of SOX4 was unchanged in 1/2 human adenomas and unchanged or decreased in all of the human adenocarcinomas relative to patient-matched normal tissue (Table5). Expression of SOX4 in human colorectal cancer cell lines was increased in DLD-1, HCT 116, HT-29, SW48, and SW837 when compared with normal human fibroblasts (Table 5). Expression of IGFBP5 in human adenomas and adenocarcinomas relative to normal matched tissue is correlated with tumor stage (Table 5). IGFBP5 levels were decreased in tumors classified as benign and returned to near normal levels when invasion through the muscularis propria was evident. IGFBP5 levels were decreased greater than 10-fold in DLD-1, HCT 116, HT-29, SW48, and SW837 cell lines (Table 5). The breast cancer cell line Hs 578T, included as a control, exhibited a 2-fold increase relative to normal human fibroblasts. Similar to IGFBP5, LY6D changed its expression pattern as tumors penetrated through the muscularis propria (Table 5). Four tumors that showed signs of invasion but not penetration through the muscularis propria (all negative for metastasis) exhibited low levels of LY6D expression. All four tumors positive for invasion through the muscularis propria exhibited high levels of expression relative to patient matched normal colon (two were also positive for metastasis). All colorectal cell lines examined expressed high levels of LY6D compared with normal fibroblasts, whereas the breast cancer cell line expressed LY6D at low levels when compared with normal fibroblasts (Table 5).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have identified 114 genes associated with early intestinal tumor development, regardless of longitudinal position along the gastrointestinal tract, that have not previously been described as Wnt targets. Twenty-four of these genes are specific to gastrointestinal tumors, as they are not highly expressed in embryonic or postnatal whole intestine. Two other reports examining expression profiles of gastrointestinal adenomas in Apc^Min/+ mice have been published recently (20, 21), as well as expression data from human colorectal adenomas and adenocarcinomas (22). The mouse studies profiled intestinal expression only, whereas our study examined expression in the duodenum, jejunum, colon, and developing gastrointestinal separately, and without RNA amplification. Our study also included array validation via in situ and real-time RT-PCR in additional mouse adenomas not used for microarray analysis. Additionally, we used real-time RT-PCR validation in human samples. Several genes with altered expression characteristics (Expi, Caspase-1, Cidea, Igfbp5, S100c, and Secreted Phosphoprotein I) were identified in both developing tissue and tumors whereas Lcn2, Ly6d, N4wbp4, and Sox4 were identified in tumors only.

This study has identified an increase in expression of the Expi gene in murine gastrointestinal adenomas. Expi encodes an extracellular proteinase inhibitor that may be important for extracellular matrix remodeling and prevention of metastasis (23). Increased expression of Expi was evident in all 14 murine intestinal adenomas as well as in involuting mammary gland tissues, as shown by our institutional database of murine gene expression profiles. The rat homologue of Expi, Wdnm1, is a putative metastasis suppressor gene based on its loss of expression in metastatic cells relative to nonmetastatic cells (24). Overexpression of Expi in adenomas along with Serpine2, another member of the serine protease inhibitors which was overexpressed in 9 of 14 of the gastrointestinal adenomas profiled (Fig. 1), suggests a potential role of protease inhibitors in prevention of cellular invasion in early tumor formation.

Another gene identified in our study encodes IGFBP5, which stimulates the growth of prostate cancer cells, improves the survival of breast cancer cells (25), yet inhibits the proliferation of both cervical carcinoma (26) and osteosarcoma cell lines (27). In the human cancer cell lines evaluated here, IGFBP5 was elevated 2-fold in Hs 578T breast carcinoma cells but was decreased more than 10-fold in five colorectal cell lines (DLD-1, HCT 116, SW837, HT-29, and SW48) relative to normal human fibroblasts. In addition, IGFBP5 expression was also decreased in 6 of 9 human tumors (Table 5). Given that IGFBP5 expression is decreased in all of the human colorectal carcinoma cell lines and in human adenocarcinomas, it may function as either an inhibitor of epithelial cell growth or a proapoptotic factor in the colon. The increased expression of Igfbp5 in gastrointestinal adenomas of Apc^Min/+ mice, however, may suggest a role in apoptosis rather than in growth inhibition. Other rodent studies have associated increased levels of Igfbp5 with apoptosis in the rat ovary, as well as with tissue remodeling and apoptosis during rat mammary gland involution (28). Because the human and murine data are contradictory, the function of IGFBP5 in gastrointestinal tumor formation remains unclear.

Lcn2 and N4wbp4 are genes highly expressed in the neoplastic epithelial cells of murine intestinal adenomas and in human adenocarcinomas. Lcn2 may protect cells during inflammatory responses, as its expression is associated with apoptosis in neutrophils following IL-3 deprivation (29), whereas N4wbp4 may also participate in inflammatory processes (16). N4wbp4 is 86% homologous to the human gene PMEPA1, which has been shown previously to be overexpressed in renal, stomach, and prostate cancer cells, as well as primary and metastatic colon cancer (16, 30, 31). PMEPA1 expression is induced by androgens and transforming growth factor-â1, a known tumor suppressor in intestinal epithelial cells (16, 31). In situ hybridization studies have shown increased PMEPA1 expression in liver metastases from colon adenocarcinomas but not in matched normal liver samples (31). In 9 of 9 human colon tumors examined in this study, high PMEPA1 expression levels were also observed.

Ly6d was highly expressed in all 14 murine gastrointestinal adenomas (Table 3). In our panel of human colorectal adenocarcinomas (Table 5), expression of LY6D, also known as E48, was strongly correlated with tumor invasion. Four of the five tumors characterized by LY6D overexpression were described as invading through the muscularis propria into the subserosa or into nonperitonealized pericolic or perirectal tissues. LY6D/E48 encodes a GPI-anchored cell surface protein expressed on normal and malignant squamous cell epithelia and is already used as a marker for head-and-neck squamous cell carcinoma (32, 33).

S100c is overexpressed in 9 of 14 Apc^Min adenomas profiled by cDNA arrays and 22 of 22 Apc^Min gastrointestinal tumors evaluated with real-time RT-PCR. Real-time results also show that S100C isoverexpressed 1.8-fold in 8 of 10 human tumors. S100C, also known in humans as CALGIZZARIN and S100A11, has been reported previously as up-regulated in various human cancer cell lines (34). S100C is a nuclear phosphoprotein linked to suppression of DNA synthesis in normal confluent human fibroblasts, although in neoplastic cells such as HeLa and Saos-2 cells, loss of nuclear S100C localization has been observed (17). Increased immunohistochemical staining of S100C has been observed in the cytoplasm of cells from small intestinal adenocarcinomas (35).

Sox4 gene expression was increased in 11 of 14 Apc^Min adenomas analyzed by cDNA arrays and in 16 of 16 additional adenomas analyzed by real-time PCR. The SOX4 gene is highly expressed inhuman breast cancer cell lines, colon cancer cell lines, hepatocarcinoma, medulloblastomas, and adenoid cystic carcinomas (36–40). Consistent with these reports, our analysis of human colorectal carcinoma cell lines demonstrated strong increases in the expression of SOX4 relative to levels in normal human fibroblasts. Sox-4 levels increased by 27-fold in DLD-1, 8-fold in HCT 116, 13-fold in SW837, 16-fold in HT-29, and 42-fold in SW48 . SOX4 levels increased 11-fold in the breast cancer cell line Hs 578T. In contrast, SOX4 expression decreased more than 1.8-fold in 9 of 16 primary human colorectal tumor samples. This result is surprising given the significant up-regulation of Sox4/SOX4 in Apc^Min adenomas and human tumor cell lines. However, it is possible that Sox4 overexpression only occurs in a subset of human colorectal cancers and our sample size is too small to detect this class of tumors. Alternatively, this expression profile may be indicative of biological differences between mice and human tumor progression. Transcription of the Sox4/SOX4 gene may be induced by activated Wnt signaling since the gene contains four potential TCF/LEF cis elements within 10 kb of the promoter; SOX4 is also a homologue of LEF1 (Celera: Score, 116; Expect, 1 x 10^–5; Identities = 23/72; Positives = 41/72), which encodes an architectural transcription factor that binds to ß-catenin. The Xenopus laevis high mobility group box proteins, XSox17/ß and XSox3, interact directly with ß-catenin and repress Wnt/ß-catenin–responsive gene expression in a concentration-dependent manner (1). These data suggest that SOX genes may also play a role in a negative feedback loop. The significance of SOX4 expression in intestinal tumorigenesis is intriguing.

Finally, high levels of Lcn2, Ly6d, N4wbp4, and S100c expression in mouse adenomas were also detected in primary human tumors at increased levels suggesting a similar role in tumor formation in both species. High levels of Igfbp5 expression in mouse adenomas were not recapitulated in primary human tumors. Two tumor characteristics that are important to this consideration of the differences in expression patterns are that most of the mouse adenomas were from the small intestine (10 of 14) and that the majority of human samples were colorectal adenocarcinomas. Preliminary results analyzing expression levels of mouse adenomas (n = 6) and mouse adenocarcinomas (n = 5) from Blm^Cin/+, Apc^Min/+ mice show increased Igfbp5, Ly6d, S100c, and Sox4 expression in all murine tumors relative to normal age-matched tissue.

We have identified 114 genes differentially expressed in gastrointestinal adenomas of Apc^Min/+ mice using cDNA arrays to assess gene expression profiles. These genes have not been previously described as Wnt targets but may be important in tumor initiation or progression. Further experiments will determine which genes are rate-limiting in gastrointestinal tumor formation in the mouse and/or human.

	Acknowledgments

 
Grant support: Grants CA 63507 (J. Groden), CA 84291 (J. Groden and B. Aronow), CA 88460 (K.H. Goss), T32-DK-64581-1 (D. Carson), and NIH grant 2M01-RR08084-11 (GCRC Tumor Bank).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Sheila Blanck and Bob Cohen for assistance with the GCRC Tumor Bank and Danielle Halbleib, Saikumar Karyala, and Craig Tomlinson of the University of Cincinnati Microarray Core for technical assistance and providing cloned cDNA fragments for in situ analyses.

	Footnotes

 
Note: K.H. Goss is currently at the Division of Epithelial Pathobiology, Department of Surgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0558. J. Groden is an investigator with the Howard Hughes Medical Institute.

Received 6/22/04. Revised 8/19/04. Accepted 11/ 3/04.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Bienz M, Clevers H. Linking colorectal cancer to Wnt signaling. Cell 2000;103:311–20.[CrossRef][Medline]
2. Smalley MJ, Dale TC. Wnt signaling in mammalian development and tumorigenesis. Cancer Met Rev 1999; 18:215–30.[CrossRef][Medline]
3. Koh TJ, Bulitta CJ, Fleming JV, Dockray GJ, Varro A, Wang TC. Gastrin is a target of the ß-catenin/TCF-4 growth-signaling pathway in a model of intestinal polyposis. J Clin Invest 2000;106:533–9.[Medline]
4. Ai Z, Fischer A, Spray DC, Brown AM, Fishman GI. Wnt-1 regulation of connexin 43 in cardiac myocytes. J Clin Invest 2000;105:161–71.[Medline]
5. Pennica D, Swanson TA, Welsh JW, et al. WISP genes are members of the connective tissue growth factor family that are up-regulated in Wnt-1 transformed cells and aberrantly expressed in human colon tumors. Proc Natl Acad Sci U S A 1998;95:14717–22.[Abstract/Free Full Text]
6. Gradl D, Kuhl M, Wedlich D. The Wnt/Wg signal transducer ß-catenin controls fibronectin expression. Mol Cell Biol 1999;19:5576–87.[Abstract/Free Full Text]
7. Conacci-Sorrell ME, Ben-Yedidia T, Shtutman M, Feinstein E, Einat P, Ben-Ze'ev A. Nr-CAM is a target gene of the b-catenin/LEF-1 pathway in melanoma and colon cancer and its expression enhances motility and confers tumorigenesis. Cell Mol Biol Lett 2002;792–5.
8. Yan D, Weismann M, Rohan M, et al. Elevated expression of AXIN2 and HNKD mRNA provides evidence that Wnt/b-catenin signaling is activated in human colon tumors. Proc Natl Acad Sci U S A 2001;98:14973–80.[Abstract/Free Full Text]
9. Mann B, Gelos M, Siedow A, et al. Target genes of ß-catenin-T cell-factor/lymphoid-enhancer-factor signaling in human colorectal carcinomas. Proc Natl Acad Sci U S A 1999;96:1603–8.[Abstract/Free Full Text]
10. Kolligs FT, Nieman MT, Winer I, et al. ITF-2, a downstream target of the Wnt/TCF pathway, is activated in human cancers with b-catenin defects and promotes neoplastic transformation. Cancer Cell 2002;1:145–55.[CrossRef][Medline]
11. Oving IM, Clevers HC. Molecular causes of colon cancer. Eur J of Clin Invest 2002;32:448–57.
12. Liu C, Li Y, Semenov M, et al. Control of ß-catenin phosphorylation/degradation by a dual-kinase mechanism. Cell 2002;108:837–47.[CrossRef][Medline]
13. Polakis P. Wnt signaling and cancer. Genes and Dev 2000;14:1837–51.[Free Full Text]
14. Liu W, Dong X, Mai M, et al. Mutations in AXIN2 cause colorectal cancer with defective mismatch repair by activating b-catenin/TCF signaling. Nat Genet 2000;26:146–7.[CrossRef][Medline]
15. Bates MD, Erwin CR, Sanford LP, et al. Novel genes and functional relationships in the adult mouse gastrointestinal tract identified by microarray analysis. Gastroenterology 2002;122:1467–82.[CrossRef][Medline]
16. Xu LL, Shi Y, Petrovics G, et al. PMEPA1, an androgen-regulated NEDD4-binding protein, exhibits cell growth inhibitory function and decreased expression during prostate cancer progression. Cancer Res 2003;63:4299–304.[Abstract/Free Full Text]
17. Sakaguchi M, Miyazaki M, Inoue Y, et al. Relationship between contact inhibition and intranuclear S100C of normal human fibroblasts. J Cell Biol 2000;149:1193–206.[Abstract/Free Full Text]
18. Moser AR, Pitot HC, Dove WF. A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science 1990;247:322–4.[Abstract/Free Full Text]
19. Schneider MR, Wolf E, Hoeflich A, Lahm H. IGF-binding protein-5: Flexible player in the IGF system and effector on its own. J Endocrinol 2002;172:423–40.[Abstract]
20. Paoni NF, Feldman MW, Gutierrez LS, Ploplis VA, Castellino FJ. Transcriptional profiling of the transition from normal intestinal epithelia to adenomas and carcinomas in the ApcMin/+ mouse. Physiol Genomics 2003;15:228–35.[Abstract/Free Full Text]
21. Chen T, Yang I, Irby R, et al. Regulation of caspase expression and apoptosis by adenomatous polyposis coli. Cancer Res 2003;63:4368–74.[Abstract/Free Full Text]
22. Notterman DA, Alon U, Sierk AJ, Levine AJ. Transcriptional gene expression profiles of colorectal adenoma, adenocarcinoma, and normal tissue examined by oligonucleotide arrays. Cancer Res 2001;61:3124–30.[Abstract/Free Full Text]
23. Dear TN, Kefford RF. The WDNM1 gene product is a novel member of the "four-disulphide core" family ofproteins. Biochem Biophys Res Commun 1991;176:247–54.[CrossRef][Medline]
24. Dear TN, Ramshaw IA, Kefford RF. Differentialexpression of a novel gene, Wdnm1, in nonmetastatic ratmammary adenocarcinoma cells. Cancer Res 1988;48:5203–9.[Abstract/Free Full Text]
25. Perks CM, Bowen S, Gill ZP, Newcomb PV, Holly JM. Differential IGF-independent effects of insulin-like growth factor binding proteins (1-6) on apoptosis of breast epithelial cells. J Cell Biochem 1999;75:652–64.[CrossRef][Medline]
26. Higo H, Duan C, Clemmons DR, Herman B. Retinoic acid inhibits cell growth in HPV negative cervical carcinoma cells by induction of insulin-like growth factor binding protein-5 (IGFBP-5) secretion. Biochem Biophys Res Commun 1997;239:706–9.[CrossRef][Medline]
27. Conover CA, Kiefer MC. Regulation and biological effect of endogenous insulin-like growth factor binding protein-5 in human osteoblastic cells. J Clin Endocrinol Metab 1993;76:1153–9.[Abstract]
28. Tonner E, Allan GJ, Flint DJ. Hormonal control of plasmin and tissue-type plasminogen activator activity in rat milk during involution of the mammary gland. J Endocrinol 2000;167:265–73.[Abstract]
29. Devireddy LR, Teodoro JG, Richard FA, Green MR. Induction of apoptosis by a secreted lipocalin that is transcriptionally regulated by IL-3 deprivation. Science 2001;293:829–34.[Abstract/Free Full Text]
30. Rae FK, Hooper JD, Nicol DL, Clements JA. Characterization of a novel gene, STAG1/PMEPA1, upregulated in renal cell carcinoma and other solid tumors. Mol Carcinog 2001;32:44–53.[CrossRef][Medline]
31. Brunschwig EB, Wilson K, Mack D, et al. PMEPA1, atransforming growth factor-ß-induced marker of terminal colonocyte differentiation whose expression ismaintained in primary and metastatic colon cancer. Cancer Res 2003;63:1568–75.[Abstract/Free Full Text]
32. Quak JJ, Balm AJ, van Dongen GA, et al. A 22-kd surface antigen detected by monoclonal antibody E 48 is exclusively expressed in stratified squamous and transitional epithelia. Am J Pathol 1990;136:191–7.[Abstract]
33. De Bree R, Roos JC, Quak JJ, Den Hollander W, SnowGB, Van Dongen GA. Clinical screening of monoclonal antibodies 323/A3, cSF-25 and K928 for suitability of targeting tumors in the upper aerodigestive and respiratory tract. Nucl Med Commun1994;15:613–27.
34. Tanaka M, Adzuma K, Iwami M, Yoshimoto K, Monden Y, Itakura M. Human calgizzarin; one colorectal cancer-related gene selected by large scale random cDNA sequencing and northern blot analysis. Can Lett 1995;89:195–200.
35. Kondo A, Sakaguchi M, Makino E, Namba M, Okada S, Huh NH. Localization of S100C immunoreactivity in various human tissues. Acta Med Okayama 2002;56:31–4.
36. McGowan EM, Clarke CL. Effect of overexpression of progesterone receptor A on endogenous progestin-sensitive endpoints in breast cancer cells.Endocrinology 1999;13:1657–71.
37. Ahn SG, Cho GH, Jeong SY, Rhim H, Choi JY, Kim IK. Identification of cDNAs for Sox-4, an HMG-Box protein, and a novel human homolog of yeast splicing factor SSF-1 differentially regulated during apoptosis induced by prostaglandin A2/d12-PGJ2 in Hep3B cells. Biochem Biophys Res Commun 1999;260:216–21.[CrossRef][Medline]
38. McCracken S, Kim CS, Xu Y, Minden M, Miyamoto NG. An alternative pathway for expression of p56lck from type I promoter transcripts in colon carcinoma. Oncogene 1997;15:2929–37.[CrossRef][Medline]
39. Lee CJ, Appleby VJ, Orme AT, Chan WI, Scotting PJ. Differential expression of SOX4 and SOX11 in medulloblastoma. J Neurooncol 2002;57:201–14.[CrossRef][Medline]
40. Frierson HFJ, El-Naggar AK, Welsh JB, et al. Large scale molecular analysis identifies genes with altered expression in salivary adenoid cystic carcinoma. Am J Pathol 2002;161:1315–23.[Abstract/Free Full Text]
41. Zorn AM, Barish GD, Williams BO, Lavender P, Klymkowsky MW, Varmus HE. Regulation of Wnt signaling by Sox proteins: XSox17 a/ß and XSox3physically interact with ß-catenin. Mol Cell 1999;4:487–98.[CrossRef][Medline]

This article has been cited by other articles:

				X. Pan, J. Zhao, W.-N. Zhang, H.-Y. Li, R. Mu, T. Zhou, H.-Y. Zhang, W.-L. Gong, M. Yu, J.-H. Man, et al. Induction of SOX4 by DNA damage is critical for p53 stabilization and function PNAS, March 10, 2009; 106(10): 3788 - 3793. [Abstract] [Full Text] [PDF]

				C.-C. Huang, S. Gadd, N. Breslow, C. Cutcliffe, S. T. Sredni, I. B. Helenowski, J. S. Dome, P. E. Grundy, D. M. Green, M. K. Fritsch, et al. Predicting Relapse in Favorable Histology Wilms Tumor Using Gene Expression Analysis: A Report from the Renal Tumor Committee of the Children's Oncology Group Clin. Cancer Res., March 1, 2009; 15(5): 1770 - 1778. [Abstract] [Full Text] [PDF]

				S. Segditsas, O. Sieber, M. Deheragoda, P. East, A. Rowan, R. Jeffery, E. Nye, S. Clark, B. Spencer-Dene, G. Stamp, et al. Putative direct and indirect Wnt targets identified through consistent gene expression changes in APC-mutant intestinal adenomas from humans and mice Hum. Mol. Genet., December 15, 2008; 17(24): 3864 - 3875. [Abstract] [Full Text] [PDF]

				C. Gaspar, J. Cardoso, P. Franken, L. Molenaar, H. Morreau, G. Moslein, J. Sampson, J. M. Boer, R. X. de Menezes, and R. Fodde Cross-Species Comparison of Human and Mouse Intestinal Polyps Reveals Conserved Mechanisms in Adenomatous Polyposis Coli (APC)-Driven Tumorigenesis Am. J. Pathol., May 1, 2008; 172(5): 1363 - 1380. [Abstract] [Full Text] [PDF]

				D. Sinner, J. J. Kordich, J. R. Spence, R. Opoka, S. Rankin, S.-C. J. Lin, D. Jonatan, A. M. Zorn, and J. M. Wells Sox17 and Sox4 Differentially Regulate {beta}-Catenin/T-Cell Factor Activity and Proliferation of Colon Carcinoma Cells Mol. Cell. Biol., November 15, 2007; 27(22): 7802 - 7815. [Abstract] [Full Text] [PDF]

				T. Fevr, S. Robine, D. Louvard, and J. Huelsken Wnt/{beta}-Catenin Is Essential for Intestinal Homeostasis and Maintenance of Intestinal Stem Cells Mol. Cell. Biol., November 1, 2007; 27(21): 7551 - 7559. [Abstract] [Full Text] [PDF]

				K. L.M. Boylan, M. A. Gosse, S. E. Staggs, S. Janz, S. Grindle, G. S. Kansas, and B. G. Van Ness A Transgenic Mouse Model of Plasma Cell Malignancy Shows Phenotypic, Cytogenetic, and Gene Expression Heterogeneity Similar to Human Multiple Myeloma Cancer Res., May 1, 2007; 67(9): 4069 - 4078. [Abstract] [Full Text] [PDF]

				P. Liu, S. Ramachandran, M. Ali Seyed, C. D. Scharer, N. Laycock, W. B. Dalton, H. Williams, S. Karanam, M. W. Datta, D. L. Jaye, et al. Sex-determining region y box 4 is a transforming oncogene in human prostate cancer cells. Cancer Res., April 15, 2006; 66(8): 4011 - 4019. [Abstract] [Full Text] [PDF]

				J. Harper, J. L. Burns, E. J. Foulstone, M. Pignatelli, S. Zaina, and A. B. Hassan Soluble IGF2 Receptor Rescues ApcMin/+ Intestinal Adenoma Progression Induced by Igf2 Loss of Imprinting Cancer Res., February 15, 2006; 66(4): 1940 - 1948. [Abstract] [Full Text] [PDF]

				P. Andreu, S. Colnot, C. Godard, P. Laurent-Puig, D. Lamarque, A. Kahn, C. Perret, and B. Romagnolo Identification of the IFITM Family as a New Molecular Marker in Human Colorectal Tumors Cancer Res., February 15, 2006; 66(4): 1949 - 1955. [Abstract] [Full Text] [PDF]

				Y. Du, S. E. Spence, N. A. Jenkins, and N. G. Copeland Cooperating cancer-gene identification through oncogenic-retrovirus-induced insertional mutagenesis Blood, October 1, 2005; 106(7): 2498 - 2505. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Reichling, T. Articles by Groden, J. Search for Related Content PubMed PubMed Citation Articles by Reichling, T. Articles by Groden, J.