| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Meeting Reports |
Gastroenterology Division, Department of Medicine [C. N. J., L. M. C.] and Abramson Cancer Center and Family Cancer Research Institute [C. N. J., L. M. C.], University of Pennsylvania, Philadelphia, Pennsylvania 19104, and Ludwig Institute for Cancer Research, Melbourne Tumor Biology Branch, Royal Melbourne Hospital, Parkville 3050, Australia [M. E.]
Introduction
The American Association for Cancer Research meeting "Colon Cancer: Genetics to Prevention" was held in Philadelphia, March 7–10, 2002. This American Association for Cancer Research Special Conference in Cancer Research was organized and chaired by Dr. Anil K. Rustgi from the University of Pennsylvania and Dr. Raju Kucherlapati from the Harvard-Partners Center for Genetics and Genomics in Boston and brought together nearly 250 attendees and 30 speakers with a keen interest in learning the very latest news in colorectal cancer CRC2 from the leading researchers in the field. The conference began with the keynote address by world-renowned cancer researcher Dr. Kenneth Kinzler (Sidney Kimmel Cancer Center, Johns Hopkins University, Baltimore, MD). After this, the conference was organized into six distinct sessions each focusing on a particular area of CRC research spanning the breadth of basic, translational, and clinical research from cancer genetics and cell biology to chemoprevention. The six sessions were as follows: APC/ß-catenin interaction and downstream targets; TGF-ß receptor type II signaling and SMADs; DNA mismatch repair; mouse models of colorectal cancer; gene arrays and functional genomics; and finally, chemoprevention and genetic epidemiology. These sessions were interspersed with two poster sessions that provided a more informal environment for interaction among conference delegates.
Summary
Novel functions and new approaches to the study of the multifaceted colorectal tumor suppressor gene, APC, were presented along with increasingly sophisticated mouse models of CRC. The application of technologies such as SAGE and oligonucleotide and cDNA microarray platforms to colon cancer research is at an exciting stage as these experimental approaches begin to bear fruit. These techniques have revolutionized the ability of cancer researchers to analyze the global changes in gene expression specific to each stage of multistep tumorigenesis and to correlate particular expression signatures with distinct histopathological phenotypes. As recently demonstrated for diffuse B-cell lymphomas (1) , the capacity to molecularly profile histologically indistinguishable lesions may not only identify novel therapeutic targets and help predict disease progression and outcome, but perhaps more importantly, allow selection of the most appropriate treatment regimen. In addition, microarray analysis has the potential to provide insight into the fundamental mechanisms through which chemoprevention strategies exert their effects. New genes transcriptionally regulated during carcinogenesis and that may play a causal role in the growth and/or spread of colorectal tumor cells in vivo have already been identified using these methods. Many more await discovery as researchers optimize use of the new technologies in concert with improvements in gene clustering and statistical analysis software and apply them to new experimental systems. One area where global analysis of gene expression will be particularly useful is the study of the process of tumor invasion and metastasis, where our knowledge lags behind our understanding of the earlier events in CRC development. Secondly, expression profiling during experimental carcinogenesis in mice harboring defined genetic alterations on a defined genetic background should also be revealing. The ability to analyze many tumors from the same cancer-prone mouse will also result in fewer false positives.
Detailed Report
Kenneth Kinzler began by reminding the audience that the validity of expression profiling directly relates to the cellular composition of the analyzed biopsies, in particular when dealing with (colonic tumor) metastasis to the liver. Therefore, before using the SAGE profiling technology previously developed in the Kinzler-Vogelstein laboratory, efforts were undertaken to purify the metastatic cells from inflammatory and other host components using immunomagnetic separation techniques (2) . The expression of one gene, the Mr 22,000 protein tyrosine phosphatase PRL-3(also known as PTP4A3), was highly correlated with liver metastases because of amplification of a small region of chromosome 8q that excluded the c-myclocus (3) . For the moment, it is unclear how the PRL-3 protein promotes tumor metastasis. Given the essential requirement for angiogenesis to support tumor cell growth beyond a certain size as well as the relative stability of the nontumor cell genome, tumor blood vessels provide an attractive target for antitumor therapy. Thus, Kinzler, Vogelstein, and colleagues have applied SAGE technology to identify cell surface marker genes exclusively expressed by tumor endothelium. Expression of four of these TEMs (TEM1, TEM5, TEM7, and TEM8) was restricted to CRC-derived blood vessels (4) , and there was no evidence for additional expression of the mouse orthologs in normal adult tissues. Indeed, TEM distribution during mouse embryonic development showed expression of TEM1, TEM5, and TEM8 at sites undergoing normal angiogenesis in liver and brain. Considering the specificity of TEM mRNA to endothelial cells and the predicted cell-surface localization of the corresponding proteins, these molecules present as attractive targets for various types of anticancer therapy. The stage is now set for the production of reagents capable of targeting one or more TEMs in tumor endothelium and for the generation of TEM knockout mice.
APC/ß-Catenin Interaction and Downstream Targets.
Over the past 5 years, much attention has focused on the central role of APC as a negative regulator of the wnt-signaling pathway through its cytoplasmic association with ß-catenin (Fig. 1)
. APC participates in a large oligomeric destruction complex (comprising at least APC, axin/conductin, and the serine-threonine kinase glycogen synthase kinase-3ß) that ultimately promotes the ubiquitination and subsequent degradation of ß-catenin by the proteasome (Ref. 5
and the references within). However, emerging evidence suggests additional activity of APC aside from modulation of the wnt-pathway. Work from Joanna Groden’s laboratory (University of Cincinnati, Cincinnati, OH) demonstrated that in SW480 CRC cells that normally express truncated APC, delivery of full-length APC resulted in cell cycle arrest in G1 and an associated increase in apoptosis (6)
. Whereas APC-mediated growth arrest was overcome by simultaneous overexpression of mutant ß-catenin or the combined expression of the ß-catenin-TCF4 target genes cyclin D1 and c-myc, the proapoptotic activity of APC appeared to be independent of transcription because full-length APC protein accelerated apoptosis in a cell-free system consisting of liver nuclei. Similarly, full-length APC, but not mutant APC, inhibited DNA replication in a cell-free system using Xenopus oocyte extract. Whether APC regulates these processes in vivo remains to be determined.
|
). Similarly ES cells derived from homozygous embryos bearing a nonsense mutation at codon 1638 (Apc1638T/1638T) also showed chromosomal instability but to a lesser extent than in Apcmin/min cells. Because APC localizes to microtubules bound to the kinetochore during mitosis, improper chromosome segregation in APC mutant cells might contribute to tumor progression. This is supported by data showing that those adenomas containing APC inactivation progress to carcinoma more frequently than those adenomas harboring an activating mutation in ß-catenin (11)
.
Kristi Neufeld (University of Utah, Salt Lake City, UT) summarized data investigating the nuclear localization and nuclear export sequences of the APC protein. Two NLSs have been identified in the COOH-terminal region, and several nuclear export sequences have been discovered throughout the molecule, including two at the NH2 terminus (12, 13, 14)
. Although it has been known for some time that APC can shuttle between the nucleus and cytoplasm, the physiological relevance of this phenomenon remains unclear. Neufeld has shown that endogenous APC and ß-catenin interact within the nucleus, and thus APC may export ß-catenin form the nucleus to terminate wnt-signaling. How the opposing actions of APC-ß-catenin and TCF4-ß-catenin complexes are regulated within nuclei is unclear at present. Intriguingly, the subcellular localization of APC, at least in vitro, appears to be modulated by cell density rather than the cell cycle (15)
. Exponentially growing MDCK and IEC6 cells feature nuclear localization of ectopically expressed and endogenous APC, whereas APC localizes to both the cytoplasm and the nucleus in superconfluent, growth-arrested cells. This nucleocytoplasmic shuttling was mediated by the second NLS (NLS2, amino acids 2048–2053), probably through association with 
-importin. It was regulated by the phosphorylation status of each of two closely spaced and evolutionarily conserved consensus CKII and PKA phosphorylation sites adjacent to NLS2. Mutational analyses and pharmacological studies showed that phosphorylation of the CKII site enhanced nuclear localization of an NLS2-ß-galactosidase fusion protein, whereas phosphorylation of the PKA site promoted its cytoplasmic accumulation. One can only speculate as to what the function of this density-mediated shuttling process might be in vivo. We clearly don’t understand all of the mechanisms regulating subcellular localization of APC because cells derived from Apc1638T/1638T mice with NLS1 and NLS2 both lacking in Apc still displayed nuclear staining of Apc (16)
.
Like APC, ß-catenin is also likely to have a role outside of the wnt-signaling pathway because 
90% of total ß-catenin is associated with E-cadherin at the inner side of the plasma membrane with the remainder localized in the cytoplasm and/or nucleus. Overexpression of E-cadherin in SW480 cells decreased proliferation possibly via sequestration of cytoplasmic ß-catenin (Barry Gumbiner, Memorial Sloan-Kettering Cancer Center, New York, NY). Conversely, ectopic expression of ß-catenin in breast and prostate cancer cell lines inhibited their invasive potential in Matrigel. These observations may complement those recently published by Kinzler et al. 1(17)
showing that the mutant ß-catenin found in HCT116 CRC cells (containing a 3-bp deletion removing the serine phosphorylation site at codon 45) had reduced affinity for E-cadherin compared with wild-type ß-catenin. It is therefore likely that mutant ß-catenin lacking the NH2-terminal phosphorylation sites not only promotes tumorigenesis via its resistance to the proteasomal degradation pathway but also by its inability to stably bind E-cadherin, thereby abrogating productive cell-cell adhesion.
TGF-ß Receptor Type II Signaling and SMADs.
There is a body of evidence suggesting that the TGF-ß signaling system suppresses tumorigenesis by inhibiting cell growth, stabilizing the genome, and stimulating apoptosis. However, local high levels can also favor tumor growth by suppression of immune surveillance and stimulation of connective tissue formation and angiogenesis. Through the use of transgenic mice expressing dominant interfering versions of the signaling (type II) TGF-ß receptor, Harold Moses (Vanderbilt-Ingram Cancer Center, Nashville, TN) provided compelling evidence that many of the tumor suppressive activities of TGF-ß may arise from intracellular signaling pathways different from those regulating, for instance, cellular motility. Intracellular signaling from TGF-ß receptors is transduced primarily through dimerization of particular SMAD protein combinations. Frequently, these signals are terminated through ubiquitination of the receptor-regulated SMADs (SMAD1, SMAD2, SMAD3, SMAD 5, and SMAD8) and antagonistic SMADs (SMAD6 and SMAD7) via association with the SMURF proteins possessing E3 ubiquitin-ligase activity (Jeffrey Wrana, Samuel Lunenfeld Research Institute, Toronto, Ontario, Canada). Furthermore, a SMAD2/SMURF2 complex can associate with the transcriptional corepressor SnoN, leading to its degradation in a TGF-ß-dependent fashion. Rik Derynck (University of California, San Francisco, CA) provided several examples of SMAD complexes acting as transcriptional coactivators such as in the presence of basal transcriptional activity provided by cEBP. Similarly, the cellular context together with the nucleotide sequence of SMAD target genes (for instance, the promoter of Runt-domain transcription factor Cbfa1) ultimately dictate whether TGF-ß signaling results in their transcriptional activation or repression.
DNA Mismatch Repair.
In the DNA mismatch repair session, Richard Fishel (Thomas Jefferson University, Philadelphia, PA) described the biochemical functions of the mammalian homologs of the bacterial mismatch repair proteins MutS and MutL on mismatched DNA sequences. Human homologues of MutS (hMSH) and MutL (hMLH) are mutated in HNPCC, and these cancers display MSI. hMSH2 and hMLH1 are the main mismatch repair genes mutated in HNPCC, whereas 15–20% of sporadic colorectal tumors are microsatellite unstable because of hypermethylation of the hMLH1 promoter (18)
. Fishel presented data to support the notion that all mammalian MSH proteins are ATPases, acting as molecular switches and showing defective nucleotide-exchange activity when mutated in CRC. The first step of the DNA mismatch repair process is recognition of the mismatched sequence by ADP-bound MSH dimers. Mismatch repair provokes exchange of ADP for ATP. MLH proteins then induce dissociation of ATP-bound MSH proteins on mismatched DNA. In this way, MLH proteins can be considered as downstream effectors of MSH proteins and act as secondary molecular switches.
Mouse Models of CRC.
The theme of DNA mismatch repair was continued in the mouse models session in which Winfried Edelman (Albert Einstein College of Medicine, Bronx, NY) summarized the phenotypes of mice containing null alleles in genes involved in mismatch repair pathways. Mouse strains with engineered null-mutations in Msh2, Msh3, Msh6, Mlh1, Pms2, and Pms1 are all cancer-prone (usually lymphomas and gastrointestinal tumors) and are often infertile because of meiotic defects. Importantly, Msh2, Msh6, and Mlh1 null mice also develop gastrointestinal tumors and therefore model the HNPCC disease. A description of the phenotype of Exonuclease-1 (Exo1) knockout mice was then given. Exonuclease-1 is a 5' to 3' exonuclease capable of acting on both single-strand and double-strand DNA breaks and is recruited to the repair complex via interaction with MSH2 and MLH1. Exo1-/- mice showed reduced survival and a frequency of MSI (14%) comparable with that observed in mMLH1-/- mice (16%). Five of 71 Exo1-/- mice examined developed tumors, although none of these tumors displayed MSI.
Riccardo Fodde (Leiden University Medical Center) explained the phenotypes of the mouse models for the inherited autosomal dominant syndrome FAP in which truncating mutations in the mouse Apc gene result in varying degrees of intestinal polyposis. The different mutant strains essentially represent an allelic series for the function of the 2843 amino acid wild-type Apc protein and, in particular, the extent of wnt-signaling (Fig. 1)
. Nonfunctional Apc protein is encoded by the most severely truncating alleles (Apc
716 and Apcmin, possessing nonsense mutations at codons 716 and 850, respectively), whereas the expression level of the longer protein in Apc1638N/1638N ES cells is only 2% of that of wild-type Apc (19
, 20)
. The protein encoded by the Apc1572T allele lacks all three axin-binding SAMP repeats, whereas one SAMP motif is retained in Apc1638T mutants. Accordingly, enhanced CRT is most pronounced in cells harboring the Apcmin and Apc1638N mutations, is less so in Apc1572T cells, and is close to wild-type levels in Apc1638T cells (16)
. These observations directly correlate with the ability of these germ-line mutations to induce both polyposis in the (small) intestine (when the mutation is heterozygous) and neonatal lethality (when the mutation is homozygous). The latter observation is also mirrored by the inverse correlation between wnt-signaling throughput and the capacity of s.c.-injected homozygous Apc-mutant ES cells to differentiate into teratomas.3
Surprisingly, retention of only the most NH2-terminal SAMP repeat (Apc1638T) is sufficient for down-regulation of CRT similar to wild-type levels, and both Apc1638T/wt and Apc1638T/1638T mice are free of intestinal tumors and any of the extraintestinal complications observed in other Apc-mutant mice. Collectively, these observations demonstrate a tight correlation between the capacity of Apc to act as a gatekeeper (for wnt-signaling) and the extent of intestinal polyposis suggesting that small variations in CRT are sufficient to modulate tumor formation in the gastrointestinal tract in mice. A similar observation was presented by Niall Tebbutt (Ludwig Institute for Cancer Research, Melbourne, Australia), who found increased colonic polyposis in mice misexpressing a weak oncogenic allele of ß-catenin and also enhanced susceptibility to colonic tumorigenesis after administration of the organotrophic carcinogen azoxymethane (21)
. Genetic cooperation between the wnt- and ras-signaling pathway was elegantly demonstrated in compound mutant Apc1638N/wt mice expressing oncogenic Kirsten-Ras (K-RasG12V) from the villin promoter in intestinal epithelial cells (the latter strain was developed by Klaus-Peter Janssen, Sylvie Robine, and Daniel Louvard at the Institut Pasteur, Paris, France). Villin-K-RasG12V mice developed intestinal lesions with an incidence of 75 and 40% of the tumors containing mutation or loss of heterozygosity at the p53 locus (22)
. However, mutations in Apc were not found. Tumors from Apc1638N/wt x villin-K-Ras mice were highly invasive, although they contained a greater number of near diploid cells and a decrease in the percentage of nuclei with broken chromosomes.4
Fodde hypothesized that a proapoptotic signal emanating from activated K-Ras may allow clearance of excessively unstable cells within the tumor.
Mark Taketo (Kyoto University, Tokyo, Japan) has taken a pharmacogenetic approach to elucidating the role of nonsteroidal anti-inflammatory drugs in CRC prevention. A major target for nonsteroidal anti-inflammatory drugs in CRCs is the inducible enzyme COX2, which coverts arachidonic acid to prostaglandin H2, ultimately leading to prostaglandin E2 production. COX2 is induced in colorectal tumor epithelial cells as well as in tumor stromal cells. Crossing the Apc716/wt mice (Fig. 1)
to mice lacking the gene encoding COX2 (Ptgs2) resulted in a reduction in the number and size of intestinal polyps. A similar phenotype was also observed in compound Apc716/wt x EP2-/- mice lacking the prostaglandin receptor EP2 (23
, 24)
. Surprisingly, the angiogenic factors vascular endothelial growth factor and basic fibroblast growth factor, which are highly expressed in the larger polyps in Apc716/wt mice, were absent in the lesions of Apc716/wt x EP2-/- compound mutant mice (25)
. Hence, enhanced COX2 activity in stromal cells may, in a cell autonomous fashion, activate EP2 receptor-dependent cyclic AMP release via synthesis of prostaglandin E2. In turn, this results not only in a further, auto-stimulatory induction of COX2 but also in the induction of angiogenic factors (vascular endothelial growth factor and basic fibroblast growth factor) and the basement membrane component laminin 2 (25)
. A supportive environment is thus created for the further growth of the adenomatous epithelium.
Lynn Matrisian (Vanderbilt University, Nashville, TN) presented data supporting a role for MMPs in CRC cell growth. MMPs are cleaved at the cell surface releasing the active enzyme and may promote tumor cell invasion and metastasis by breakdown of the extracellular matrix. As for COX2, most MMPs up-regulated during tumor formation are produced by reactive stromal cells (26) . However, matrilysin (MMP7) was up-regulated in the epithelial cells of human colorectal adenomas and carcinomas as well as in polyps from Apcmin/wt mice. This may be attributable to increased gene transcription achieved via a complex of ß-catenin/p300 and the ETS family member, PEA3 (27) . Accordingly, matrilysin is expressed at the adenoma stage, whereas other MMPs such as Gelatinase-A, Gelatinase-B, Stromelysin-1, and Stromelysin-3 are induced at the carcinoma stage. Compound Apcmin/wt mice lacking matrilysin expression showed a 58% reduction in tumor number when compared with simple Apcmin/wt mice (28) , and similarly, administration of the MMP inhibitor BB-94 reduced tumor burden by 50%. Because matrilysin is expressed by the tumor epithelium and COX2 by stromal cells, coadministration of MMP7- and COX2-specific inhibitors showed an additive effect with respect to reducing polyp multiplicity, although an additive effect on apoptosis induction and inhibition of proliferation was not observed. As pointed out by Matrisian, results of clinical trials using MMP inhibitors for the treatment of various cancer types have been largely disappointing to date, however, MMP inhibitors may provide improved patient survival if used in combination with other targeted therapies.
Robert Coffey (Vanderbilt University) demonstrated the importance of EGFR signaling in colorectal tumor development using mouse genetics. When waved2 mice, expressing an EGFR variant defective in tyrosine kinase activity, were crossed to Apcmin/wt mice, polyposis was almost completely abrogated in the resulting compound mutant animals, although polyp size and the occurrence of microadenomas was unaffected (29) . These results agree with a previous report showing a reduction in polyp number after administration of EGFR inhibitors (30) and implicate EGFR signaling in expansion of Apcmin/wt microadenomas (formed by loss of heterozygosity of the wild-type Apc allele) to polyps. These results also suggest that EGFR kinase inhibitors may be beneficial for the treatment of CRC. The combined power of microdissection and expression profiling has lead to the identification of genes with an expression pattern either confined to the differentiated or proliferating compartments in the colonic epithelium. In work performed in collaboration with Robert Whitehead, intact crypts were isolated, divided in two, and expression profiling performed. Twenty-five genes were expressed exclusively in the top half, and 50 genes in the bottom half of the crypt. Interestingly, N-Ras was found to be restricted to proliferating cells in the lower crypt region. Labeling studies demonstrated that N-Ras-positive cells do not migrate toward the crypt surface and rather remain localized to the proliferative zone. It remains to be established whether N-Ras may label a progenitor cell population anchored at the base of the crypt.
Gene Arrays and Functional Genomics.
The theme of functional genomics and gene arrays was continued by Jian Yu (Sidney Kimmel Cancer Center, Johns Hopkins University) describing a screen for novel p53-inducible genes in DLD-1 CRC cells using SAGE technology. Genes were prioritized on the basis of early induction, induction in multiple cell lines, and p53-dependent up-regulation by chemotherapeutic drugs. This process of elimination identified PUMA as a downstream target of p53 (31)
. PUMA is a BH3-domain only protein and interacts with the apoptotic regulators Bcl-2 and Bcl-XL. Expression of PUMA suppressed growth of CRC cells via induction of apoptosis that occurred more quickly than that stimulated by induction of p53. HCT116 CRC cells deficient for PUMA are now being generated and should provide a good model system to study the requirement for PUMA in p53-dependent apoptotic pathways.
Microarray technology is likely to identify additional pathways involved in tumor formation, provide new diagnostic and/or prognostic markers, and possibly highlight new molecular targets for cancer treatment. Daniel Notterman (Princeton University, Princeton, NJ) compared preparations of normal colon, adenoma, and adenocarcinoma using the Affymetrix GeneChip oligonucleotide arrays (32)
. This study identified GRO/melanocyte growth stimulatory activity, a CXC chemokine often overexpressed by melanoma cells, as a gene up-regulated during the adenoma to adenocarcinoma transition. Enforced expression of GRO/melanocyte growth stimulatory activity in CRC cells resulted in a reduction of cell-cell adhesion. An expanded four-way comparison of normal epithelium with aberrant crypt foci, adenoma, and adenocarcinoma is now underway. Filippo Randazzo (Chiron Corporation, Emeryville, CA) compared the expression of 30,000 genes between laser-capture microdissected biopsies of normal and neoplastic colon. One of the hits, the transmembrane protein Net4, was up-regulated at least 2-fold in 80% of tumors and at least 4-fold in 60% of CRCs. Also, expression of antisense Net4 reduced the growth rate of a CRC cell line.
Chemoprevention and Genetic Epidemiology.
To start the final session, Leonard Augenlicht (Albert Einstein College of Medicine) reviewed the early approaches to gene expression profiling in the 1980s (33)
, which identified a role for mitochondrial function in regulation and coordination of proliferation and apoptotic pathways in colon cells (reviewed in Ref. 34
) and risk for tumor formation (35)
. Recent work has used microarray methods to demonstrate the complexity of altered gene expression profiles that characterize and distinguish responses to chemopreventive agents (36)
and, when the data are coupled to use of an Apc1638T, p21-/- mouse model, demonstrated the necessity of p21 for the antitumorigenic effects of sulindac (37)
. Finally, Augenlicht described the phenotype of a mouse model in which the Muc2 gene, which encodes the major gastrointestinal mucin synthesized and secreted by goblet cells, had been targeted for inactivation (38)
. Muc2-/- mice lacked discernible goblet cells in the colon and small intestine and displayed elongated crypts, increased proliferation, and reduced apoptosis in the mucosa. Moreover, these mice also developed tumors in the small intestine, colon, and rectum. The phenotype may reflect a lack of cellular protection against dietary carcinogens, or more likely, adds Muc2 to the ever-growing list of molecules that directly regulate epithelial cell proliferation in the mammalian intestine. This presentation was followed by prominent leaders in CRC epidemiology, Martin Lipkin (Cornell University, New York, NY), Dean Brenner (University of Michigan, Ann Arbor, MI), and Ernst Hawk (National Cancer Institute, Bethesda, MD), who delineated current progress and challenges in dietary modulation of CRC risk and chemopreventive strategies.
FOOTNOTES
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.
1 To whom requests for reprints should be addressed. E-mail: cameronj{at}mail.med.upenn.edu 
2 The abbreviations used are: CRC, colorectal cancer; APC, adenomatous polyposis coli; TGF, transforming growth factor; SAGE, serial analysis of gene expression; TEM, tumor endothelial marker; Min, multiple intestinal neoplasia; MSI, microsatellite instability; HNPCC, hereditary nonpolyposis colorectal cancer; hMLH, human MutL homologue; FAP, familial adenomatous polyposis; ES, embryonic stem; NLS, nuclear localization sequence; CKII, casein kinase II; PKA, protein kinase A; hMSH, human MutS homologue; SAMP, Ser-Ala-Met-Pro; CRT, ß-catenin regulated transcription; COX, cyclooxygenase; MMP, matrix metalloprotease; EGFR, epidermal growth factor receptor; PUMA, p53 up-regulated modulator of apoptosis.
3 M. Kielman et al., submitted for publication.
4 K-P. Janssen, S. Robine, and R. Fodde, manuscript in preparation.
Received 7/ 8/02. Accepted 9/19/02.
REFERENCES
716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell, 87: 803-809, 1996.[Medline]
716) knockout mice. Nat. Med., 7: 1048-1051, 2001.[Medline]
716) mouse intestinal polyps. Cancer Res., 62: 506-511, 2002.This article has been cited by other articles:
|
|
 |
|
  R.C. Sills, H.L. Hong, G. Flake, C. Moomaw, N. Clayton, G.A. Boorman, J. Dunnick, and T.R. Devereux o-Nitrotoluene-induced large intestinal tumors in B6C3F1 mice model human colon cancer in their molecular pathogenesis Carcinogenesis, April 1, 2004; 25(4): 605 - 612. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
