Junk DNA and the long non-coding RNA twist in cancer genetics.

The central dogma of molecular biology states that the flow of genetic information moves from DNA to RNA to protein. However, in the last decade this dogma has been challenged by new findings on non-coding RNAs (ncRNAs) such as microRNAs (miRNAs). More recently, long non-coding RNAs (lncRNAs) have attracted much attention due to their large number and biological significance. Many lncRNAs have been identified as mapping to regulatory elements including gene promoters and enhancers, ultraconserved regions and intergenic regions of protein-coding genes. Yet, the biological function and molecular mechanisms of lncRNA in human diseases in general and cancer in particular remain largely unknown. Data from the literature suggest that lncRNA, often via interaction with proteins, functions in specific genomic loci or use their own transcription loci for regulatory activity. In this review, we summarize recent findings supporting the importance of DNA loci in lncRNA function and the underlying molecular mechanisms via cis or trans regulation, and discuss their implications in cancer. In addition, we use the 8q24 genomic locus, a region containing interactive SNPs, DNA regulatory elements and lncRNAs, as an example to illustrate how single-nucleotide polymorphism (SNP) located within lncRNAs may be functionally associated with the individual's susceptibility to cancer.

Use of aspirin postdiagnosis improves survival for colon cancer patients.

BACKGROUND: The preventive role of non-steroid anti-inflammatory drugs (NSAIDs) and aspirin, in particular, on colorectal cancer is well established. More recently, it has been suggested that aspirin may also have a therapeutic role. Aim of the present observational population-based study was to assess the therapeutic effect on overall survival of aspirin/NSAIDs as adjuvant treatment used after the diagnosis of colorectal cancer patients. METHODS: Data concerning prescriptions were obtained from PHARMO record linkage systems and all patients diagnosed with colorectal cancer (1998-2007) were selected from the Eindhoven Cancer Registry (population-based cancer registry). Aspirin/NSAID use was classified as none, prediagnosis and postdiagnosis and only postdiagnosis. Patients were defined as non-user of aspirin/NSAIDs from the date of diagnosis of the colorectal cancer to the date of first use of aspirin or NSAIDs and user from first use to the end of follow-up. Poisson regression was performed with user status as time-varying exposure. RESULTS: In total, 1176 (26%) patients were non-users, 2086 (47%) were prediagnosis and postdiagnosis users and 1219 (27%) were only postdiagnosis users (total n=4481). Compared with non-users, a survival gain was observed for aspirin users; the adjusted rate ratio (RR) was 0.77 (95% confidence interval (CI) 0.63-0.95; P=0.015). Stratified for colon and rectal, the survival gain was only present in colon cancer (adjusted RR 0.65 (95%CI 0.50-0.84; P=0.001)). For frequent users survival gain was larger (adjusted RR 0.61 (95%CI 0.46-0.81; P=0.001). In rectal cancer, aspirin use was not associated with survival (adjusted RR 1.10 (95%CI 0.79-1.54; P=0.6). The NSAIDs use was associated with decreased survival (adjusted RR 1.93 (95%CI 1.70-2.20; P<0.001). CONCLUSION: Aspirin use initiated or continued after diagnosis of colon cancer is associated with a lower risk of overall mortality. These findings strongly support initiation of a placebo-controlled trial that investigates the role of aspirin as adjuvant treatment in colon cancer patients.

Early morbidity encountered in the dietary-related mouse model of Barrett's esophagus: a question of zinc?

Recently, a mouse model for Barrett's esophagus based on a zinc-deficient diet supplemented with deoxycholic bile acids has been published. The aim of this study was to attempt to reproduce these data and extend them by employing genetically modified mice and intraperitoneal iron supplementation. The study design encompassed six experimental groups (wild type, Apc-mutant and Smad4-mutant mice, with or without iron injections), with all animals fed with the zinc-deficient diet supplemented with deoxycholic bile acids. All treatments were started at 3-5 weeks of age (the majority [78%] at 5 weeks). Animals were scheduled for euthanasia at two distinct time points, namely at 3 and 6 months of age. All mice showed signs of considerable distress already 4 weeks after the start of the modified diets, and had to be euthanized before the first evaluation time point (mean age 9.3 weeks, range 5-15 weeks). No differences were observed between wild type and genetically modified mice, or between animals with or without iron supplementation. On histological examination, we could not detect any lesions (Barrett's esophagus-like or tumors) other than esophagitis. In the currently presented experimental settings, we were not able to reproduce the mouse model according to which Barrett's-like lesions could be detected in animals fed with the zinc-deficient diet supplemented with deoxycholic bile acids.

Tumour-stroma interactions in colorectal cancer: converging on beta-catenin activation and cancer stemness.

Sporadic cases of colorectal cancer are primarily initiated by gene mutations in members of the canonical Wnt pathway, ultimately resulting in beta-catenin stabilisation. Nevertheless, cells displaying nuclear beta-catenin accumulation are nonrandomly distributed throughout the tumour mass and preferentially localise along the invasive front where parenchymal cells are in direct contact with the stromal microenvironment. Here, we discuss the putative role played by stromal cell types in regulating beta-catenin intracellular accumulation in a paracrine fashion. As such, the tumour microenvironment is likely to maintain the cancer stem cell phenotype in a subset of cells, thus mediating invasion and metastasis.

Wnt pathway-related gene expression during malignant progression in ulcerative colitis.

Long-standing ulcerative colitis (UC) has been associated with a high risk of developing colonic adenocarcinoma. Importantly, both low- and high-grade dysplasia are strongly related to the presence or development of malignancy. The canonical Wnt/beta-catenin signaling pathway is of crucial importance in cancer development and progression, but its role in UC-related carcinogenesis remains to be determined. We evaluated the immunolabeling patterns of beta-catenin, as well as the products of Wnt-associated cancer genes E-cadherin, cyclin D1 and c-myc, along the dysplasia-carcinoma pathway in UC. For this purpose, immunohistochemistry (IHC) was performed on 18 adenocarcinomas and 17 dysplasias, derived from 21 patients. We found that intracellular beta-catenin accumulation, the hallmark of Wnt signaling activation, is observed in dysplasia, together with enhanced labeling of nuclear protein cyclin D1 and reduction of membranous labeling of E-cadherin. c-myc displayed moderate immunolabeling in the (pre)malignant lesions. Thus, the Wnt pathway is activated in early stages of malignant progression in UC. Furthermore, upregulation of the oncogene cyclin D1 and downregulation of tumor suppressor E-cadherin also occurs in the (pre)neoplastic state. This may contribute to the high potential for malignant degeneration of dysplasia in UC-related colitis.

Expression and genomic profiling of colorectal cancer.

Colorectal cancer still represents a paradigm for the elucidation of the cellular, genetic and molecular mechanisms that underly solid tumor initiation, progression to malignancy, and metastasis to distal organ sites. The relative ease with which pathological specimens can be obtained by either surgery or endoscopy from different stages of tumor progression has facilitated the application of omics technologies to allow the genome-wide analysis both at the RNA (gene expression) and DNA (aneuploidy) levels. Here, we have reviewed the multiplicity of studies appeared to date in the scientific literature on the expression and genomic analysis of colorectal cancer, and attempted an integration of the profiling data generated and made available in the public domain. This approach is likely to pinpoint specific chromosomal loci and the corresponding genes which (i) play rate-limiting roles in colorectal cancer, (ii) represent putative diagnostic and prognostic markers for the accurate prediction of clinical outcome and response to treatment, and (iii) encompass potential therapeutic targets. Moreover, cross-species data mining and integration of the human colorectal cancer profiles with those obtained from mouse models of intestinal tumorigenesis will even more contribute to the elucidation of highly conserved pathways and cellular functions underlying malignancy in the GI tract. Notwithstanding the above promises, tumor heterogeneity, limited cohort sizes, and methodological differences among experimental and bioinformatic approaches still poses main obstacles towards the optimal utilization and integration of omics profiles.

Smad4 haploinsufficiency in mouse models for intestinal cancer.

The Smad4(+/E6sad) mouse carries a null mutation in the endogenous Smad4 gene resulting in serrated adenomas and mixed polyposis of the upper gastrointestinal (GI) tract with 100% penetrance. Here, we show by loss of heterozygosity (LOH) analysis and immunohistochemistry (IHC) that, although the majority of the tumors appear at 9 months of age, somatic loss of the wild-type Smad4 allele occurs only at later stages of tumor progression. Hence, haploinsufficiency underlies Smad4-driven tumor initiation in the GI tract. As both the Apc and Smad4 tumor suppressor genes map to mouse chromosome 18, we have bred Smad4(+/E6sad) with the Apc(+/1638N) model to generate two distinct compound heterozygous lines carrying both mutations either in cis (CAS) or in trans (TAS). Strikingly, both models show increased tumor multiplicities when compared with the single mutant littermates, although CAS mice are more severely affected and became moribund at only 5-6 weeks of age. Phenotypic and molecular analyses indicate that Smad4 haploinsufficiency is sufficient to significantly affect tumor initiation and progression both prior to and upon loss of Apc function. Moreover, complete loss of Smad4 strongly enhances Apc-driven tumor formation.

The role of the APC tumor suppressor in chromosomal instability.

Colorectal cancer (CRC) still represents the model of choice to study the mechanisms underlying tumor initiation and progression. Accordingly, CRC has been central in the analysis of the role played by chromosomal instability (CIN) in tumor initiation and progression. Although loss of APC tumor suppressor function initiates the adenoma-carcinoma sequence in the vast majority of CRCs through constitutive activation of Wnt/beta-catenin signaling, the APC gene also represents a candidate CIN gene in CRC. Accordingly, two studies published in 2001 showed that truncating Apc mutations can lead to both quantitative and qualitative ploidy changes in primary mouse cell lines, mainly due to kinetochore and centrosome abnormalities. Here, we review and discuss the more recent literature on APC's functional activities possibly related to its role in eliciting CIN in tumor initiation and progression. We propose a model where loss and/or truncation of APC cause mitotic spindle defects that, upon somatic inactivation of other putative CIN genes (e.g. spindle and cell cycle checkpoint genes, DNA repair, telomere maintenance, etc.) underlie aneuploidy as observed in the majority of CRCs.

Multiplicity in polyp count and extracolonic manifestations in 40 Dutch patients with MYH associated polyposis coli (MAP).

OBJECTIVE: To investigate the contribution of MYH associated polyposis coli (MAP) among polyposis families in the Netherlands, and the prevalence of colonic and extracolonic manifestations in MAP patients. METHODS: 170 patients with polyposis coli, who previously tested negative for APC mutations, were screened by denaturing gradient gel electrophoresis and direct sequencing to identify MYH germline mutations. RESULTS: Homozygous and compound heterozygous MYH mutations were identified in 40 patients (24%). No difference was found in the percentage of biallelic mutation carriers between patients with 10-99 polyps or 100-1000 polyps (29% in both groups). Colorectal cancer was found in 26 of the 40 patients with MAP (65%) within the age range 21 to 67 years (median 45). Complete endoscopic reports were available for 16 MAP patients and revealed five cases with gastro-duodenal polyps (31%), one of whom also presented with a duodenal carcinoma. Breast cancer occurred in 18% of female MAP patients, significantly more than expected from national statistics (standardised morbidity ratio = 3.75). CONCLUSIONS: Polyp numbers in MAP patients were equally associated with the attenuated and classical polyposis coli phenotypes. Two thirds of the MAP patients had colorectal cancer, 95% of whom were older than 35 years, and one third of a subset of patients had upper gastrointestinal lesions. Endoscopic screening of the whole intestine should be carried out every two years for all MAP patients, starting from age 25-30 years. The frequent occurrence of additional extraintestinal manifestations, such as breast cancer among female MAP patients, should be thoroughly investigated.

Radiation induces different changes in expression profiles of normal rectal tissue compared with rectal carcinoma.

PURPOSE: Radiotherapy is a very effective adjuvant treatment for rectal cancer with little side effects. Its killing effect on tumor cells seems to be more profound than the effect on normal tissue. The molecular events caused by irradiation are mainly analyzed in in vitro and animal models; investigations on human material are rare. In the current study, we analyzed the effects of irradiation on gene expression in normal and tumor tissue of rectal cancer patients. METHODS AND MATERIALS: Normal and carcinoma tissue of patients from a randomized clinical trial of the benefits of preoperative radiotherapy were analyzed using the Affymetrix Human Cancer Gene Chip. Preoperative radiotherapy was given within 5 days prior to surgery. Results for normal tissue and tumor were compared to investigate the radiation-related differences between normal and tumor cells. We clustered the differentially expressed genes based on their functional annotation. Results were compared with immunohistochemical and literature data. RESULTS: The majority of the investigated cancer-related genes remained unchanged by irradiation (92% in tumor tissue and 93% in normal tissue). The differentially expressed genes varied between tumor and normal tissue except for maspin and IL-8. Both in tumor and normal tissue, differentially expressed genes were present related to cell signaling and cycle control, apoptosis and cell survival and tissue response and repair. However, the spectrum of affected genes was totally different. CONCLUSION: Pre-existing differences in gene expression between normal tissue and tumor tissue might explain the differences in their responses to radiation. This change in response may explain the clinical beneficial effect of radiotherapy on tumor cells (low local recurrence rate) and the less severe effects on normal tissue (minor side effects).

Carrier risk status changes resulting from mutation testing in hereditary non-polyposis colorectal cancer and hereditary breast-ovarian cancer.

CONTEXT: In hereditary cancer syndrome families with an identified cancer associated mutation, mutation testing changes the carrier risk status of the tested person and may change the carrier risk status of relatives. OBJECTIVE: This study aimed to describe the change in the distribution of carrier risk status resulting from testing in hereditary breast-ovarian cancer (HBOC) and hereditary non-polyposis colorectal cancer (HNPCC) families. DESIGN: This was an observational cohort study. PATIENTS: The cohort included members of 75 HBOC and 47 HNPCC families. Of the 10 910 cohort members, 1408 were tested for a mutation and learned their test results. OUTCOME MEASURE: Carrier risk for all cohort members was assessed before and after mutation testing. RESULTS: There was a change in carrier risk status in 2906 subjects after testing of 1408 family members. The most common type of carrier risk change, from at risk to non-carrier status, accounted for 77% of the risk changes; 12% were a change to known carrier status from a lower risk. Sixty percent of persons with a carrier risk status change were not themselves tested; their risk status changed because of a relative's test result. CONCLUSIONS: Carrier risk status changes from uncertainty to certainty (that is, to carrier or to non-carrier) account for 89% of risk changes resulting from testing. These risk changes affect cancer prevention recommendations, most commonly reducing their burden. Current practices do not ensure that untested family members are informed about changes in their carrier risk status which result from mutation testing of their relatives.

Rapidly progressive adenomatous polyposis in a patient with germline mutations in both the APC and MLH1 genes: the worst of two worlds.

The two most common inherited forms of colorectal cancer are familial adenomatous polyposis and hereditary non-polyposis colorectal cancer. Simultaneous inheritance of both an APC gene mutation and a mismatch repair gene (for example, MLH1) mutation has never been described. In the present case report, we report rapidly progressive adenomatous polyposis in a 10 year old boy with a germline frame shift mutation in the APC gene and a germline splice site mutation in the MLH1 gene. Immunohistochemical investigations showed abnormal expression of beta-catenin in early adenomas with low grade dysplasia, attributed to the APC gene mutation. Subsequent loss of function of the MLH1 gene, as shown by absent immunostaining of its protein in adenomas with high grade dysplasia, may well have caused the rapid progression to high grade dysplasia in many of the adenomas.

The APC gene in colorectal cancer.

Mutations in the adenomatous polyposis coli (APC) gene are not only responsible for familial adenomatous polyposis (FAP), but also play a rate-limiting role in the majority of sporadic colorectal cancers. Colorectal tumours are known to arise through a gradual series of histological changes, the so-called 'adenoma-carcinoma' sequence, each accompanied by a genetic alteration in a specific oncogene or tumour suppressor gene. Loss of APC function triggers this chain of molecular and histological changes. In general, an intestinal cell needs to comply with two essential requirements to develop into a cancer: it must acquire selective advantage to allow for the initial clonal expansion, and genetic instability to allow for multiple hits at other genes responsible for tumour progression and malignant transformation. Inactivation of APC seems to fulfill both requirements. In this short review, I will discuss the role played by APC in providing, when mutated, selective advantage, through constitutional activation of the Wnt signal transduction pathway, and chromosomal instability to the nascent intestinal tumor cell.

Dynamic expression and nuclear accumulation of beta-catenin during the development of hair follicle-derived structures.

Beta-catenin has a dual role in the cell. At the membrane, it connects E-cadherin to the actin cytoskeleton, while in the nucleus, it controls gene expression in concert with Tcf-like transcription factors. Nuclear translocation of beta-catenin is induced by the Wnt signal transduction pathway. Control of this process is essential since elevated beta-catenin levels interfere with differentiation and development, and can initiate cancer in many tissues. An important role for beta-catenin during hair follicle related development and tumorigenesis has recently been established, though little is known of its endogenous expression during the development of these structures. Here, we have investigated the expression of beta-catenin in relation to markers for proliferation, differentiation and Wnt signaling during the development of three hair follicle related structures, i.e. whiskers, normal body hair and the preputial gland, and a hair follicle-derived tumor, the epidermal cyst. We observed nuclear accumulation of beta-catenin, the hallmark of Wnt signaling, in the upper matrix, the dermal papilla, the developing ringwulst of the whisker and in the tumor, though it was never in association with proliferation or terminal differentiation. Co-localization of nuclear beta-catenin with Tcf-3/4 was found only in the dermal papilla and the developing ringwulst of the whisker, but not in the upper matrix or in the tumor. These results further elucidate the role of the Wnt signal transduction pathway during hair follicle related development and tumorigenesis and illustrate the dynamic role of beta-catenin in signal transduction and cell-adhesion.

A nonsense mutation in MLH1 causes exon skipping in three unrelated HNPCC families.

Germline mutations in the DNA mismatch repair genes MSH2 and MLH1 are responsible for the majority of hereditary nonpolyposis colorectal cancer (HNPCC) families. A common mutation mechanism is to disrupt MLH1 and MSH2 mRNA splicing. The disruption creates aberrant mRNAs lacking specific coding exons (exon skipping). Here, we report a novel skipping of MLH1 exon 12 caused by an AAG to TAG nonsense mutation at codon 461 in three HNPCC families of North American origins. The nonsense codon was found in a conserved haplotype in the three unrelated families and seems to represent a founder mutation. The skipping created an aberrant MLH1 mRNA transcript lacking exon 12. The effect of the codon 461 nonsense mutation on exon 12 skipping is evident even though it was placed in a minigene construct containing entirely different coding sequences. Notably, the effect of the nonsense mutation on exon skipping is incomplete. Accordingly, a second aberrant MLH1 transcript encompassing the nonsense codon is also produced. Whereas the latter transcript is unstable, presumably because of nonsense-mediated mRNA decay, neither of the aberrant transcripts seems to affect the stability of wild-type MLH1 mRNA. This study demonstrates that the germ-line nonsense mutation at codon 461 of MLH1 disrupts normal MLH1 mRNA processing, and that exon skipping underlies pathogenesis in these HNPCC families.

Disease model: familial adenomatous polyposis.

Mutations in the APC gene are responsible for familial adenomatous polyposis (FAP) and for the majority of sporadic colorectal cancers. The establishment of genotype-phenotype correlations in FAP is often complicated by the great clinical variability observed among carriers of the same APC mutation even within the same kindred. This variability is likely to arise from the interaction of genetic and environmental modifying factors, the dissection of which ideally requires the employment of mouse models where the effects of specific Apc mutations are analyzed in an inbred, homogeneous genetic background and a controlled environment. The availability of different Apc mouse models allows not only the establishment of more precise genotype-phenotype correlations but has also provided very important clues for the understanding of the function of APC in homeostasis and tumorigenesis. Also, the close phenotypic resemblance to the human disease makes these mice unique preclinical models to test chemopreventive and therapeutic interventions.

A targeted mouse Brca1 mutation removing the last BRCT repeat results in apoptosis and embryonic lethality at the headfold stage.

A mouse model with a targeted mutation in the 3' end of the endogenous Brca1 gene, Brca1(1700T), was generated to compare the phenotypic consequences of truncated Brca1 proteins with other mutant Brca1 models reported in the literature to date. Mice heterozygous for the Brca1(1700T) mutation do not show any predisposition to tumorigenesis. Treatment of these mice with ionizing radiation or breeding with Apc, Msh-2 or Tp53 mutant mouse models did not show any change in the tumor phenotype. Like other Brca1 mouse models, the Brca1(1700T) mutation is embryonic lethal in homozygous state. However, homozygous Brca1(1700T) embryos reach the headfold stage but are delayed in their development and fail to turn. Thus, in contrast to Brca1(null) models, the mutant embryos do not undergo growth arrest leading to a developmental block at 6.5 dpc, but continue to proliferate and differentiate until 9.5 dpc. Homozygous embryos die between 9.5-10.5 dpc due to massive apoptosis throughout the embryo. These results indicate that a C-terminal truncating Brca1 mutation removing the last BRCT repeat has a different effect on normal cell function than does the complete absence of Brca1.

Atypical HNPCC owing to MSH6 germline mutations: analysis of a large Dutch pedigree.

Hereditary non-polyposis colorectal cancer (HNPCC) is the most common genetic susceptibility syndrome for colorectal cancer. HNPCC is most frequently caused by germline mutations in the DNA mismatch repair (MMR) genes MSH2 and MLH1. Recently, mutations in another MMR gene, MSH6 (also known as GTBP), have also been shown to result in HNPCC. Preliminary data indicate that the phenotype related to MSH6 mutations may differ from the classical HNPCC caused by defects in MSH2 and MLH1. Here, we describe an extended Dutch HNPCC family not fulfilling the Amsterdam criteria II and resulting from a MSH6 mutation. Overall, the penetrance of colorectal cancer appears to be significantly decreased (p<0.001) among the MSH6 mutation carriers in this family when compared with MSH2 and MLH1 carriers (32% by the age of 80 v >80%). Endometrial cancer is a frequent manifestation among female carriers (six out of 13 malignant tumours). Transitional cell carcinoma of the urinary tract is also relatively common in both male and female carriers (10% of the carriers). Moreover, the mean age of onset of both colorectal cancer (MSH6 v MSH2/MLH1 = 55 years v 44/41 years) and endometrial carcinomas (MSH6 v MSH2/MLH1 = 55 years v 49/48 years) is delayed. As previously reported, we confirm that the pattern of microsatellite instability, in combination with immunohistochemical analysis, can predict the presence of a MSH6 germline defect. The detailed characterisation of the clinical phenotype of this kindred contributes to the establishment of genotype-phenotype correlations in HNPCC owing to mutations in specific mismatch repair genes.

Mutations in the APC tumour suppressor gene cause chromosomal instability.

Two forms of genetic instability have been described in colorectal cancer: microsatellite instability and chromosomal instability. Microsatellite instability results from mutations in mismatch repair genes; chromosomal instability is the hallmark of many colorectal cancers, although it is not completely understood at the molecular level. As truncations of the Adenomatous Polyposis Coli (APC) gene are found in most colorectal tumours, we thought that mutations in APC might be responsible for chromosomal instability. To test this hypothesis, we examined mouse embryonic stem (ES) cells homozygous for Min (multiple intestinal neoplasia) or Apc1638T alleles. Here we show that Apc mutant ES cells display extensive chromosome and spindle aberrations, providing genetic evidence for a role of APC in chromosome segregation. Consistent with this, APC accumulates at the kinetochore during mitosis. Apc mutant cells form mitotic spindles with an abundance of microtubules that inefficiently connect with kinetochores. This phenotype is recapitulated by the induced expression of a 253-amino-acid carboxy-terminal fragment of APC in microsatellite unstable colorectal cancer cells. We conclude that loss of APC sequences that lie C-terminal to the beta-catenin regulatory domain contributes to chromosomal instability in colorectal cancer.

Cyclooxygenase-two (COX-2) modulates proliferation in aggressive fibromatosis (desmoid tumor).

Aggressive fibromatosis is a locally invasive soft tissue lesion. Seventy-five per cent of cases harbor a somatic mutation in either the APC or beta-catenin genes, resulting in beta-catenin protein stabilization. Cyclooxygenase-2 (COX-2) is an enzyme involved in prostaglandin synthesis that modulates the formation of colonic neoplasia, especially in cases due to mutations resulting in beta-catenin stabilization. Human aggressive fibromatoses and lesions from the Apc+/Apc1638N mouse (a murine model for Apc-driven fibromatosis) demonstrated elevated COX-2 levels. COX-2 blockade either by the selective agent DFU or by non-selective COX blocking agents results in reduced proliferation in human tumor cell cultures. Breeding mice with Cox-2-/- mice resulted in no difference in number of aggressive fibromatoses formed, but in a smaller tumor size, while there was a decrease in number of GI lesions by 50%. Mice fed various COX blocking agents also showed a decline in tumor size. COX-2 expression was regulated by tcf-dependent transcription in this lesion. COX-2 partially regulates proliferation due to beta-catenin stabilization in aggressive fibromatosis. Although COX blockade alone does not cause tumor regression, this data suggests that it may have a role as an adjuvant therapy to slow tumor growth in this lesion.