Search

Cyclooxygenases (COXs) are key enzymes that convert arachidonic acid to prostaglandins. Overexpression of one of the COX isozymes, COX2, has been shown to play an important role in colorectal cancer progression. Recently, however, low expression of COX2 has been reported in a subset of colorectal and gastric cancers. Aberrant CpG island methylation and associated transcriptional silencing are common in colorectal cancer, and we therefore investigated the potential role of methylation in the transcriptional silencing of COX2. We examined the methylation status of the COX2 5' CpG island in a series of tumor cell lines. Among the 33 cell lines examined, dense methylation (>70%) of COX2 was detected in 5 cell lines, and partial methylation was detected in 10 cell lines. Detailed methylation mapping using bisulfite genomic sequencing revealed that loss of expression of COX2 mRNA was closely correlated with methylation of a region upstream of exon 1, and expression could be restored by demethylation using the DNA methyltransferase inhibitor 5-aza-deoxycytidine. Aberrant methylation of COX2 was also detected in 12 of 92 (13%) unselected sporadic primary colorectal cancers and 7 of 50 (14%) colorectal adenomas. COX2 methylation was strongly associated with the presence of the CpG island methylator phenotype (P<0.01), inversely related to p53 gene mutation (P<0.01), and unrelated to microsatellite instability status. We propose that COX2 expression in colorectal tumors is modulated by functional factors that favor high expression and by the CpG island methylator phenotype that favors silencing in a subset of cases. These results raise the possibility that tumors with COX2 methylation may be less sensitive to treatment using specific COX2 inhibitors.

The Ki-ras protooncogene frequently is mutated in colorectal adenocarcinomas, but the etiology of this molecular event is uncertain. We investigated the association between variables known or suspected to be related to risk for colorectal cancer and the occurrence of Ki-ras mutations in colorectal adenomas. This study was conducted among 678 male and female participants, 40-80 years of age, enrolled in a phase III trial testing the effects of a wheat bran fiber supplement on adenoma recurrence. Exposure information on the risk factors of interest was assessed through self-administered questionnaires. Mutations in codons 12 and 13 of the Ki-ras protooncogene were analyzed in baseline adenomas 0.5 cm or larger by PCR amplification followed by direct sequencing. Eighteen percent (120 of 678) of the participants had one or more adenoma(s) with Ki-ras mutations. A higher risk of Ki-ras mutations was associated with increasing age and a lower intake of total folate. The odds ratio (OR) for Ki-ras mutations for individuals >72 years of age was 1.98 [95% confidence interval (CI) = 1.19-3.27; P for trend = 0.008] compared with those less than 65 years of age. Compared with individuals in the lower tertile of total folate, those in the upper tertile had an approximately 50% lower risk of having Ki-ras mutation-positive adenomas (OR = 0.52; 95% CI = 0.30-0.88; P for trend = 0.02). There was a suggestion of a stronger inverse association of total folate with G-->T transversions (OR = 0.41; 95% CI = 0.20-0.87) than G-->A transitions (OR = 0.61; 95% CI = 0.31-1.21), although the CIs for the associations overlap. The results of these analyses suggest that the protective effect of folate in colon cancer observed in published studies may be mediated through folate's effect on Ki-ras mutations.

The DNA repair protein O6-methylguanine DNA methyltransferase (MGMT) removes alkyl adducts from the O6 position of guanine. MGMT expression is decreased in some tumor tissues, and lack of activity has been observed in some cell lines. Loss of expression is rarely due to deletion, mutation, or rearrangement of the MGMT gene, but methylation of discrete regions of the CpG island of MGMT has been associated with the silencing of the gene in cell lines. We used methylation-specific PCR to study the promoter methylation of the MGMT gene. All normal tissues and expressing cancer cell lines were unmethylated, whereas nonexpressing cancer cell lines were methylated. Among the more than 500 primary human tumors examined, MGMT hypermethylation was present in a subset of specific types of cancer. In gliomas and colorectal carcinomas, aberrant methylation was detected in 40% of the tumors, whereas in non-small cell lung carcinomas, lymphomas, and head and neck carcinomas, this alteration was found in 25% of the tumors. MGMT methylation was found rarely or not at all in other tumor types. We also analyzed MGMT expression by immunohistochemistry in relation to the methylation status in 31 primary tumors. The presence of aberrant hypermethylation was associated with loss of MGMT protein, in contrast to retention of protein in the majority of tumors without aberrant hypermethylation. Our results suggest that epigenetic inactivation of MGMT plays an important role in primary human neoplasia.

Peutz-Jeghers syndrome (PJS) is an autosomal dominant condition characterized by intestinal hamartomatous polyps, mucocutaneous melanin deposition, and increased risk of cancer. Families with PJS from the Johns Hopkins Polyposis Registry were studied to identify the molecular basis of this syndrome and to characterize the pathogenesis of gastrointestinal hamartomas and adenocarcinomas in PJS patients. Linkage analysis in the family originally described by Jeghers in 1949 and five other families confirmed linkage to 19p13.3 near a recently identified gene responsible for PJS. Germ-line mutations in this gene, STK11, were identified in all six families by sequencing genomic DNA. Analysis of hamartomas and adenocarcinomas from patients with PJS identified loss of heterozygosity (LOH) of 19p markers near STK11 in 70% of tumors. Haplotype analysis indicated that the retained allele carried a germ-line mutation, confirming that STK11 is a tumor suppressor gene. LOH of 17p and 18q was identified in an adenocarcinoma but not in hamartomas, implying that allelic loss of these two regions corresponds to late molecular events in the pathogenesis of cancer in PJS. The adenocarcinomas showing 17p LOH also demonstrated altered p53 by immunohistochemistry. None of the 18 PJS tumors showed microsatellite instability, LOH on 5q near APC, or mutations in codons 12 or 13 of the K-ras proto-oncogene. These data provide evidence that STK11 is a tumor suppressor gene that acts as an early gatekeeper regulating the development of hamartomas in PJS and suggest that hamartomas may be pathogenetic precursors of adenocarcinoma. Additional somatic mutational events underlie the progression of hamartomas to adenocarcinomas, and some of these somatic mutations are common to the later stages of tumor progression seen in the majority of colorectal carcinomas.

In December 1997, the National Cancer Institute sponsored "The International Workshop on Microsatellite Instability and RER Phenotypes in Cancer Detection and Familial Predisposition," to review and unify the field. The following recommendations were endorsed at the workshop. (a) The form of genomic instability associated with defective DNA mismatch repair in tumors is to be called microsatellite instability (MSI). (b) A panel of five microsatellites has been validated and is recommended as a reference panel for future research in the field. Tumors may be characterized on the basis of: high-frequency MSI (MSI-H), if two or more of the five markers show instability (i.e., have insertion/deletion mutations), and low-frequency MSI (MSI-L), if only one of the five markers shows instability. The distinction between microsatellite stable (MSS) and low frequency MSI (MSI-L) can only be accomplished if a greater number of markers is utilized. (c) A unique clinical and pathological phenotype is identified for the MSI-H tumors, which comprise approximately 15% of colorectal cancers, whereas MSI-L and MSS tumors appear to be phenotypically similar. MSI-H colorectal tumors are found predominantly in the proximal colon, have unique histopathological features, and are associated with a less aggressive clinical course than are stage-matched MSI-L or MSS tumors. Preclinical models suggest the possibility that these tumors may be resistant to the cytotoxicity induced by certain chemotherapeutic agents. The implications for MSI-L are not yet clear. (d) MSI can be measured in fresh or fixed tumor specimens equally well; microdissection of pathological specimens is recommended to enrich for neoplastic tissue; and normal tissue is required to document the presence of MSI. (e) The "Bethesda guidelines," which were developed in 1996 to assist in the selection of tumors for microsatellite analysis, are endorsed. (f) The spectrum of microsatellite alterations in noncolonic tumors was reviewed, and it was concluded that the above recommendations apply only to colorectal neoplasms. (g) A research agenda was recommended.

Recent studies indicate that nonsteroidal anti-inflammatory drugs have a chemopreventive effect against colorectal neoplasia. Nonsteroidal anti-inflammatory drugs inhibit cyclooxygenases, principal enzymes that mediate the formation of prostanoids. To determine whether prostanoids are involved in the pathogenesis of colorectal adenomas, we compared the levels of five major stable metabolic products of the cyclooxygenase pathway in the normal-appearing mucosa and in adenomas of patients with familial adenomatosis polyposis. Of 12 patients tested, 6 had elevated levels of at least one prostanoid in the adenomas. More importantly, the relative levels of three prostanoids [prostaglandin (PG)D2, PGE2, and 6-keto-PGF1alpha] were elevated in adenomas compared to normal-appearing mucosa from the same patients, and the resulting ratios were correlated with the size of the adenoma. These results suggest a role for prostanoids in progression of colorectal polyposis in familial adenomatosis polyposis patients.

We have searched for the presence of genetic alterations in serum DNA obtained from 44 colorectal cancer patients. Microsatellite analysis using highly polymorphic markers revealed loss of heterozygosity and/or microsatellite instability in 35 of 44 (80%) primary tumors. No alterations were detected in the paired serum DNA. We next used an oligonucleotide-mediated mismatch ligation assay to detect tumor specific gene mutations in the serum. Among the 16 cases with a K-ras gene mutation in the tumor, the same mutation was detected in three paired serum samples. In the 10 cases with a p53 mutation in the tumor, the identical mutation was detected in seven corresponding serum samples. Comparison of the molecular analysis with clinical diagnosis of these patients revealed that none of the seven Dukes' stage B patients with a K-ras mutation in their tumors demonstrated a mutation in the serum. In contrast, five of seven stage B patients with a p53 mutation in the tumor demonstrated a mutation in the paired serum (P = 0.01, Fisher's exact test). Taken together, either a K-ras or p53 mutation was detected in the serum in 40% of the 25 patients (95% confidence interval, 21-61%), whose primary tumors contained a mutation and in 23% of the 44 patients (95% confidence interval, 12-38%) with colorectal cancer. The frequent detection of p53 mutation in the serum of patients with early stage tumors suggests a possible use of this approach for clinical prognosis and cancer monitoring of colorectal cancer patients.

To begin to identify new tumor markers, we recently performed a systematic study of gene expression in cancers of the colon and pancreas. Of the 45,000 genes identified, 183 were found to be expressed at significantly elevated levels in pancreatic cancer. One of the genes was tissue inhibitor of metalloproteinase type I (TIMP-1), which encodes a secreted protein. Analysis of TIMP-1 serum levels revealed significant increases in pancreatic cancer patients, but TIMP-1 by itself was inadequate as a serum marker for cancer. However, a combination of individually suboptimal markers (TIMP-1, CA19-9, and carcinoembryonic antigen) detected 60% of 85 patients with pancreatic cancers in a highly specific manner. These results suggest that a systematic analysis of gene expression can reveal novel serum markers and that individually suboptimal markers can be combined to yield higher sensitivity and specificity.

Allelic loss of chromosome 18q has been noted in intestinal type gastric adenocarcinomas. Smad4 is a gene located at 18q that was recently cloned in humans and found to be significantly altered in pancreatic cancers. We sought to determine whether Smad4 genetic alterations played a significant role in gastric tumorigenesis by studying 35 gastric adenocarcinomas of all histopathological types and pathological stages. Microdissected specimens were used for mutational analysis of Smad4 at the nucleotide level, including the entire coding region and intron/exon boundaries. Allelic imbalance was also analyzed at the Smad4 locus using two nearby microsatellite markers. One case of apparent biallelic inactivation of Smad4 was found in our study of 35 gastric carcinomas. A nonsense point mutation at codon 334 was demonstrated, which, similar to other Smad4 mutations, is predicted to truncate the conserved COOH-terminal domain of this protein. This Smad4 C to T transition mutation was proven to be somatically acquired. Allelic loss was also noted on chromosome 18q at a marker near Smad4 in this mutated gastric cancer, apparently producing complete inactivation of Smad4 in this tumor. Significant 18q allelic loss (56% of 34 informative cases) was noted in our gastric carcinomas using microsatellite markers near the Smad4 locus, regardless of histological subtype or pathological stage. Additionally, three cases of microsatellite instability were observed. Thus, Smad4 inactivation was noted in our gastric carcinomas; however, this event was rare. The frequent loss of chromosomal arm 18q observed in gastric cancers suggests the presence of other tumor suppressor genes in this region that are involved in gastric tumorigenesis. Further studies are needed to identify these other targets of inactivation during gastric cancer development.

De novo methylation of promoter region CpG islands has been increasingly associated with transcriptional inactivation of important genes in neoplasia. To study the potential mechanisms underlying aberrant methylation in cancer, we have determined the methylation patterns of selected genes in colorectal cancers with and without microsatellite instability (MI), which results from defects in one of several base mismatch repair genes. A total of 47 colorectal cancers were analyzed, of which 15 were MI+ (32%). We now report that both the frequency and the extent of de novo methylation are strikingly increased in MI+ cancers. Hypermethylation of the p16 gene was found in 60% of MI+ cancers, compared to only 22% in MI- cancers (P = 0.02). Similarly, hypermethylation of the thrombospondin-1 (TSP-1) gene, an angiogenesis inhibitor, was increased in MI+ cancers (27% versus 0%; P = 0.008). Extensive methylation of insulin-like growth factor II (IGF2) and hypermethylated in cancer-1 (HIC-1) genes was observed in 60 and 80% of MI+ cancers, respectively, as contrasted with 6 and 38% of MI- cancers (P = 0.0002 and 0.01, respectively). Furthermore, 60% of the MI+ cancers displayed the hypermethylation events at two or more loci in a concordant manner compared to only 9% of the MI- cancers (P < 0.001). These results demonstrate a strong link between promoter hypermethylation and genetic instability due to deficient DNA repair.

Familial adenomatous polyposis (FAP), due to germ-line mutation of the adenomatous polyposis coli (APC) gene, is characterized by development of colorectal adenomas and ultimately colorectal cancer. The usefulness of ornithine decarboxylase (ODC) activity and polyamine levels in normal-appearing colorectal mucosa to stratify risk for colorectal neoplasia by discriminating presymptomatic individuals with germ-line APC mutation (genotype-positive) from genotype-negative family controls was evaluated in 36 at-risk subjects undergoing endoscopic and genetic screening for FAP. ODC activity and levels of putrescine, spermidine, and spermine were significantly higher in presymptomatic genotype-positive patients compared to genotype-negative persons (P = 0.029, <0.001, 0.002, and <0.001, respectively). Moreover, a putrescine level with a cutoff point of 1.5 nmol/mg protein was the most accurate single discriminator of risk status. ODC activity and polyamine levels are significantly elevated in gene carriers of FAP before the development of polyposis, suggesting a role for these compounds in tumorigenesis of FAP. These assays may be useful in evaluating at-risk members of FAP families in which mutation of the APC gene cannot be found.

Hereditary nonpolyposis colorectal cancer is caused by inherited mutations of mismatch repair genes. We developed monoclonal antibodies to the prototype human mismatch repair gene hMSH2 and used them to detect an immunoreactive protein of M(r) 100,000 in mismatch-proficient cell lines. In addition, a M(r) 150,000 protein coimmunoprecipitated with the hMSH2 gene product in cell lines expressing hMSH2. Immunohistochemistry demonstrated that the hMSH2 protein was exclusively nuclear. Whereas the hMSH2 protein was expressed in a variety of tissues, the most striking pattern was observed in esophageal and intestinal epithelia, where expression was limited to the replicating compartment. Neoplastic cells within benign and malignant mismatch repair-proficient tumors expressed the protein, but no hMSH2 immunoreactivity was observed in the colorectal tumors of patients with germline hMSH2 mutation. These results have implications for tumorigenic mechanisms and, potentially, for diagnosis.

The p53-regulated gene product p21WAF1/CIP1 is the prototype of a family of small proteins that negatively regulate the cell cycle. To learn more about p21WAF1/CIP1 regulation in vivo, monoclonal antibodies were developed for immunohistochemistry. These revealed that p21WAF1/CIP1 expression followed radiation-induced DNA damage in human skin in a pattern consistent with its regulation by p53. A detailed comparison of the human, rat, and mouse p21WAF1/CIP1 promoter sequences revealed that this induction was probably mediated by conserved p53-binding sites upstream of the transcription start site. In unirradiated tissues, p21WAF1/CIP1 expression was apparently independent of p53 and was observed in a variety of cell types. Moreover, there was a striking compartmentalization of p21WAF1/CIP1 expression throughout the gastrointestinal tract that correlated with proliferation rather than differentiation. As epithelial cells migrated up the crypts, the Ki67-expressing proliferating compartment near the crypt base ended abruptly, with the coincident appearance of a nonproliferating compartment expressing p21WAF1/CIP1. In colonic neoplasms, this distinct compartmentalization was largely abrogated. Cell cycle inhibitors are thus subject to precise topological control, and escape from this regulation may be a critical feature of neoplastic transformation.

Colorectal tumorigenesis proceeds through an accumulation of specific genetic alterations. Studies of the mechanism by which these genetic changes effect malignant transformation have focused on the deregulation of cell proliferation. However, colorectal epithelial homeostasis is dependent not only on the rate of cell production but also on apoptosis, a genetically programmed process of autonomous cell death. We investigated whether colorectal tumorigenesis involved an altered susceptibility to apoptosis by examining colorectal epithelium from normal mucosa, adenomas from familial adenomatous polyposis, sporadic adenomas, and carcinomas. The transformation of colorectal epithelium to carcinomas was associated with a progressive inhibition of apoptosis. The inhibition of apoptosis in colorectal cancers may contribute to tumor growth, promote neoplastic progression, and confer resistance to cytotoxic anticancer agents.

Germline mutations of the adenomatous polyposis coli (APC) gene lead to multiple intestinal tumors in familial adenomatous polyposis patients and in multiple intestinal neoplasia (Min) mice. Current models predict that inactivation of the remaining normal allele of a tumor suppressor gene is rate limiting for tumor formation, but this has been difficult to prove. While examination of colorectal adenomas from familial adenomatous polyposis patients identified somatic inactivating mutations of the second allele in the majority of tumors (19 of 24), the absolute requirement for an early inactivating event could not be demonstrated. In contrast, inactivation of the remaining allele of the murine APC (Apc) could be demonstrated in 100% (30 of 30) of tumors from Min mice. Moreover, inactivation was observed in the earliest recognizable phase of tumors, including some lesions containing as few as two dysplastic crypts. These results suggest that the mutation of the second APC allele is an early event in Min and familial adenomatous polyposis tumorigenesis, supporting Knudson's hypothesis.

One hallmark of malignant potential is dysplasia, the disruption of normal morphology. While it is generally recognized that cancer is the result of a series of genetic changes, the relationship of these alterations and their timing to the advent of dysplasia remains obscure. To address this issue, 54 small benign colorectal lesions of various malignant potential were analyzed for APC and K-RAS mutations, two alterations which have been implicated in the early stages of colorectal tumorigenesis. APC mutations were closely associated with dysplasia. In contrast, K-RAS mutations were found to be remarkably common in small nondysplastic lesions which apparently have a limited potential to progress to larger tumors. These results provide evidence that the nature and order of genetic changes can have a specific impact on both tumor morphology (e.g., dysplasia) and the likelihood of tumor progression.

It has recently been shown that hereditary nonpolyposis colorectal cancer (HNPCC) is caused by hereditable defects in DNA mismatch repair genes. However, the fraction of HNPCC due to defects in any one repair gene and the nature of these mutations are not known. We analyzed 29 HNPCC kindreds for mutations in the prototype DNA mismatch repair gene hMSH2 by a combination of linkage analysis, polymerase chain reaction-based screening, and sequencing of the coding region. The complete intron/exon structure of the gene was ascertained to facilitate this analysis. The results suggest that at least 40% of classic HNPCC kindreds are associated with germline mutations in hMSH2 and that most of these mutations produce drastic alterations in the predicted protein product.

Hereditary nonpolyposis colorectal carcinoma (HNPCC) is the most common form of hereditary colon cancer. Autosomal dominant inheritance is evident from pedigrees but the genetic basis of the disorder is otherwise unknown. Recently, two genes in 5q21 involved in colon carcinogenesis, APC and MCC, were identified, and APC was shown to be the gene predisposing to familial adenomatous polyposis. To determine if these genes also confer susceptibility to HNPCC we performed linkage analyses in nine affected families. The MCC-APC region could be formally excluded as the locus for HNPCC in seven families. In one family the results were suggestive of exclusion, although they were not conclusive. The remaining family was uninformative. We used two alternative definitions of affected status. Based on haplotypes for MCC and APC the added pairwise logarithm-of-odds score for all nine families was -22.57 at the recombination fraction of 0.00 using more stringent criteria for the HNPCC phenotype and -22.67 for less stringent criteria. In addition to blood DNA samples from living family members, DNA from formaldehyde-fixed archival pathology specimens from decreased individuals contributed to these linkage results.

Epidemiological and experimental studies suggest that dehydroepiandrosterone (DHEA), an adrenal cortical steroid, has chemoprotective properties. Rat colonic epithelium which had been induced to a premalignant state by the colonic carcinogen azoxymethane was used as a model for patients at high risk of colorectal carcinoma, and the efficacy of dietary DHEA for chemoprotection against tumorigenesis was evaluated. Ten-week-old male F344 rats (n = 100) were given 10 weekly s.c. injections of azoxymethane at a dose of 10 mg/kg/week. One day after the final dose of carcinogen, DHEA was added to the diet of 50 rats (0.5% DHEA chow), and the other rats were used as pair-fed controls. DHEA-fed rats lost body weight throughout the 17-week study, in contrast to their pair-fed controls. Serum DHEA in DHEA-fed rats at the end of the study was 6 times that of controls (120 +/- 30 versus 18 +/- 14 pmol/ml), and serum DHEA sulfate was 23 times that of controls (1311 +/- 13 versus 55 +/- 13 pmol/ml). Addition of DHEA to the diet produced no significant chemoprotection in our model. Tumor-related mortality was somewhat increased in DHEA-fed rats (20% versus 6% in week 16 of DHEA feeding, P not significant). The cumulative prevalence of left colonic tumors, identified by weekly colonoscopic examinations, was somewhat lower in DHEA-fed rats than in controls during weeks 10 through 13 (17% versus 33% in week 12, P not significant), but in week 14 the prevalence in DHEA-fed rats became similar to that in controls (39% versus 41%). Growth curves of autochthonous left colonic tumors, as assessed for 8 weeks by computerized image analysis of colonoscopic photographs, were similar for DHEA-fed and control rats. Prevalence, mean frequency, multiplicity, and diameter of colonic tumors at necropsy of colonoscopically negative rats in week 17 were somewhat lower in the DHEA-fed rats (e.g., prevalence of 47% versus 67%), but the differences from controls were not significant. Parameters of colonic epithelial proliferation after tritiated thymidine incorporation in DHEA-fed rats were similar to those in control rats (labeling index of 8.3 +/- 0.7% versus 8.4 +/- 0.6% in week 17), despite higher serum DHEA and DHEA sulfate levels. Our findings indicate that DHEA did not have significant postinduction chemoprotective activity against azoxymethane-induced colonic tumorigenesis in this model utilizing pair-fed controls. Further preclinical studies appear to be needed before dietary DHEA can be recommended for chemoprotection trials in patients with premalignant colorectal epithelium.

We compared the metabolism of azoxymethane (AOM) and of N-nitrosodimethylamine (NDMA) by liver microsomes obtained from male F344 rats pair-fed for 3 weeks either a control liquid diet or an isocaloric liquid diet containing ethanol at a concentration of 6.6% by volume. High-performance liquid chromatographic analysis of the products of the microsomal metabolism of AOM showed that methylazoxymethanol was the only primary metabolite. While the formation of small (less than 4% of methylazoxymethanol) quantities of methanol and formaldehyde could also be detected in this reaction, these products could be accounted for almost entirely by the spontaneous decomposition of methylazoxymethanol. With NDMA as the substrate in the incubation system, the formation of methylamine, formaldehyde, methanol, and an additional, as yet unidentified metabolite was detected. Liver microsomes obtained from rats fed the ethanol-containing diet up to the time of sacrifice were 12-18 times more active in the metabolism of both AOM and NDMA than liver microsomes obtained from rats fed the control, ethanol-free diet for the same period. When rats fed the ethanol diet for 20.5 days were fed the control diet for 0.5 days and then sacrificed, only a 2- to 3-fold increase in the metabolism of both AOM and NDMA by liver microsomes was observed, indicating that cessation of ethanol intake results in a rapid decrease of the ethanol-induced metabolic enzymes. Hepatocytes isolated from ethanol-fed rats showed a significantly enhanced sensitivity to AOM- as well as to NDMA-induced unscheduled DNA synthesis, indicating that the increased rate of microsomal metabolism induced by ethanol is associated with enhanced carcinogen activation in vitro. The metabolism of AOM and NDMA by liver microsomes was inhibited to similar extents by carbon monoxide, pyrazole, sodium azide, aminoacetonitrile, imidazole, and ethanol. In addition, both ethanol and NDMA were found to inhibit competitively the microsomal metabolism of AOM. These results suggest that AOM and NDMA are metabolized by very similar, indeed perhaps the same rat liver microsomal enzyme(s).

Earlier studies of the methylation status of total genomic DNA and of specific genes have demonstrated, predominantly, hypomethylation in human neoplasms. However, we have recently documented the presence of new sites of methylation in the calcitonin gene in human lymphomas (100%), small cell lung carcinomas (92%), and acute myeloid leukemias (95%). We now report that these same novel calcitonin gene methylation sites are also a feature of DNA from human colonic adenomas (13 out of 14 studied), colon carcinomas (4/13), and established colon carcinoma cell lines (18/19), despite the presence of overall genomic DNA hypomethylation in these neoplasms. The data provide further evidence that regional increases in DNA methylation, like gene hypomethylation, occur in benign colonic neoplasms prior to malignant transformation. The fact that abnormalities of calcitonin gene methylation are less frequent in DNA from human colonic carcinomas than from adenomas and colon carcinoma cell culture lines is of special interest. This finding suggests that a more heterogeneous population of cells is present in the carcinomas and that the calcitonin gene hypermethylation may be inherent to cells which are initially selected for growth in culture or are capable of prolonged survival under culture conditions.

Ornithine decarboxylase, the first regulatory enzyme in polyamine biosynthesis, is inhibited by alpha-difluoromethylornithine (DFMO), a specific enzyme-activated irreversible inhibitor. DFMO has been shown previously to inhibit experimental colonic tumorigenesis in rodents given the large bowel carcinogen azoxymethane or dimethylhydrazine. Therefore, we assessed the effects of DFMO on growth of established autochthonous experimental colonic tumors. Ten-wk-old male F344 rats were given 10 weekly s.c. injections of azoxymethane, 10 mg/kg. Starting 5 wk after the last dose, colonoscopy to the splenic flexure was performed weekly with a pediatric fiberoptic bronchoscope. When a tumor was visualized, its growth was assessed by computer image analysis of weekly colonoscopic photographs which included a scale. After two measurements for baseline tumor growth, the tumor-bearing rats were assigned in predetermined alternating sequence to the DFMO group (n = 26) or control group (n = 28). DFMO, 30 mg/ml (3%), was given in drinking water for 4 wk, resulting in mean weekly intake of 16 +/- 1 (SE) to 18 +/- 1 mg of DFMO/g of body weight. Control rats were pair-fed, resulting in reduced body weights comparable to DFMO rats. DFMO dramatically inhibited tumor growth, beginning in the first week of administration: mean tumor volume of DFMO rats reached only 7.0 +/- 2.0 mm3 compared with 17.4 +/- 3.2 mm3 in controls (P less than 0.02); and tumors in three DFMO rats disappeared. Mean change in tumor volume in DFMO rats was less than controls during all 4 wk of administration, although there was a suggestion of escape from DFMO suppression of tumor growth in the last 2 wk. At necropsy, tumor ornithine decarboxylase activity was 115 +/- 22 pmol/h/mg of protein in DFMO rats as compared with 842 +/- 576 in controls. There was a suggestion of greater tumor desmoplasia in DFMO rats, but tumor differentiation, depth of invasion, inflammation, and labeling index with tritiated thymidine showed no statistically significant differences between the DFMO and control groups. Our findings suggest that (a) ornithine decarboxylase plays a key role in growth of autochthonous experimental colonic tumors, and (b) DFMO may have potential for chemotherapy and chemoprophylaxis of colorectal neoplasms in human beings.

It has been demonstrated that restriction fragment length polymorphisms of X-chromosome genes can be used in conjunction with methylation patterns to determine the clonal composition of human tumors. In this report, we show that several X-chromosome probes can be used for such analyses. In particular, probes derived from the hypoxanthine phosphoribosyltransferase gene and the phosphoglycerate kinase gene could be used for clonal analysis in over 50% of American females. The X-inactivation patterns observed with these probes were found to accurately reflect clonality in more than 95% of 92 tumors tested.

Epidemiological studies have shown an association between consumption of alcoholic beverages and carcinoma of the large bowel, but studies in experimental models of colonic carcinogenesis have yielded conflicting results. We assessed the effects on azoxymethane-induced colonic carcinogenesis of both timing of chronic dietary ethanol consumption relative to carcinogen administration and quantity of ethanol consumption. Ten-week-old male Fischer 344 rats were given 11%, 22%, or 33% of calories as reagent ethanol or no ethanol by pair feeding with Lieber-DeCarli-type liquid diets providing comparable total carbohydrates, proteins, fats, and calories. Ten weekly s.c. injections of the bowel carcinogen azoxymethane (AOM), 7 mg/kg, were given to all rats in weeks 1-10. Three experimental groups were given their respective ethanol diet during acclimatization and AOM administration (preinduction and induction phases) and then were given the no-ethanol diet from week 11 until sacrifice in week 26 (postinduction phase). Three other groups received the no-ethanol diet during acclimatization and AOM administration and then were changed to their respective ethanol diet until sacrifice. The control AOM group received the no-ethanol diet throughout the study. Suppression of colonic tumorigenesis occurred in the groups with high levels of chronic dietary ethanol consumption during acclimatization and AOM administration: in the 33% and 22% diet groups, the prevalence of colonic tumors was 3% and 20% as compared with 50% in control (P less than 0.001 and P less than 0.02, respectively). Tumorigenesis in the left colon was more affected than in the right colon, as tumor prevalence in the left colon was decreased in both the 33% and 22% diet groups (0% in both versus 24% in control, P less than 0.005), whereas prevalence in the right colon was decreased only in the 33% diet group (3% versus 38%, P less than 0.001). By contrast, prevalence of colonic tumors in the 11% diet group was not significantly different from control. Chronic dietary ethanol consumption after AOM administration had no effect on tumor outcome, regardless of quantity of consumption. In an analogous study of [14C]AOM metabolism in rats fed the 33% diet during acclimatization and AOM administration, 14CO2 was exhaled at a slower rate than in rats fed no-ethanol diet (P = 0.05), indicating suppression of AOM metabolism.(ABSTRACT TRUNCATED AT 400 WORDS)

We examined the metabolism of 14C-labeled methylazoxymethanol (MAM) in male F344 rats pair-fed for 21.0 days either a liquid control diet, an isocaloric liquid diet containing 6.6% ethanol by volume (continuous ethanol diet), or the ethanol diet for 20.5 days followed by the control diet for 0.5 day (interrupted ethanol diet). Compared to rats fed the control liquid diet, metabolism of [1,2-14C]MAM acetate to exhaled 14CO2 was inhibited by 25 to 42% in rats fed the continuous ethanol diet, but was initially stimulated by 90% in rats given the interrupted ethanol diet. MAM-induced DNA methylation, as reflected in 7-methylguanine and O6-methylguanine content 24 h after carcinogen administration, was inhibited in the colon mucosae of rats fed the interrupted ethanol diet by 52 to 54%, and an even greater inhibition (71 to 86%) of DNA methylation occurred in the colon mucosae of rats fed the continuous ethanol diet. Liver DNA methylation was significantly inhibited (by 32 to 42%) only in those rats fed the continuous ethanol diet. Liver microsomes isolated from rats fed the 3 diets metabolized MAM to formic acid and methanol in vitro, but liver microsomes from rats fed the continuous ethanol diet were 12 to 15 times more active than liver microsomes from rats fed the control diet. Liver microsomes isolated from rats fed the interrupted ethanol diet were only 3 to 5 times more active in MAM metabolism than liver microsomes from rats fed the control diet, indicating very rapid turnover of the ethanol-induced enzyme(s) catalyzing the oxidation of MAM. Although chronic ethanol feeding enhanced the activity of liver microsomes for MAM metabolism, ethanol was found to inhibit the reaction competitively. Hepatocytes isolated from rats fed the continuous ethanol diet were considerably more sensitive to MAM-induced unscheduled DNA synthesis than hepatocytes isolated from rats given the control liquid diet, indicating that the stimulation of MAM metabolism by dietary ethanol results in increased DNA damage, observable in an in vitro system. Thus, the increased metabolic activation of MAM, due to enzyme induction by ethanol which is observed in vitro, is not reflected in increased liver DNA methylation in vivo.(ABSTRACT TRUNCATED AT 400 WORDS)

In the azoxymethane (AOM)-treated rat model of colonic carcinogenesis, serial injections of the carcinogen lead to the eventual development of colonic tumors. However, the precise nature of carcinogenesis events in this commonly used model is not well defined, and the occurrence of classic initiation and promotion phases after serial injections is uncertain. Since increases in ornithine decarboxylase (ODC) activity have been associated with promotion in other tumor models, we measured mucosal ODC during AOM carcinogenesis in the rat. We gave male Fischer 344 rats ten weekly injections of AOM at a dose of 3 mg/kg (Wk 1-10) and measured colonic mucosal ODC activity during the entire 25-wk course. Colonic tumor incidence at Wk 25 was 0% in controls and 48% in AOM animals (15 of 31). Mucosal ODC in AOM animals was significantly increased over controls. The time course of changes in mucosal ODC was similar throughout the entire colon and differed generally in magnitude only. Distinct and prolonged increases in ODC activity occurred within 4 h of a single injection of carcinogen and persisted for at least 14 days. With a second AOM injection on Day 7, there was another distinct and prolonged increase in ODC over the persistently elevated activity. Over the entire 25 wk, the increase in ODC was distinctly biphasic, higher at Wk 2, 11, and 13 than at Wk 6, 17, 21, and 25. The findings indicate that AOM induces an increase in mucosal ODC during colonic carcinogenesis, and they suggest that this carcinogenesis model, with ten weekly injections of carcinogen, has the properties of a multistep process. The early peak in ODC might be associated with the early carcinogenesis (initiation) phase(s), and a subsequent second increase in ODC might be associated with late carcinogenesis (post-initiation or promotion) phase (s). These results strengthen the utility of the rat AOM colonic carcinogenesis model for further studies of the role of ODC and polyamine metabolism in neoplastic transformation.

Epidemiological studies have shown an association between consumption of alcoholic beverages and increased occurrence of large bowel carcinoma, but studies in experimental models of colonic carcinogenesis have produced conflicting results. We assessed the effects of chronic dietary ethanol consumption during the preinduction and induction phase (period of acclimatization and carcinogen administration) in a high-dose azoxymethane-treated rat model (14 mg/kg/wk for 10 wk). Ten-wk-old male Fischer 344 rats were given 33% of calories as ethanol or no ethanol (controls). Pair-feeding with Lieber-DeCarli-type liquid diets provided comparable total carbohydrates, proteins, fats, and calories. After 3 wk of dietary acclimatization, injections of azoxymethane (AOM) were given s.c. to all rats in Wk 1 to 10. At necropsy in Wk 25, dramatic suppression of gastrointestinal tumorigenesis was evident in the ethanol-fed group: the prevalence of colonic tumors was 5% as compared with 91% in controls; and the prevalence of small bowel tumors was 0% versus 74% (P less than 0.0001). In an analogous study of [14C]AOM metabolism, exhaled 14CO2 was decreased in the ethanol-fed rats, indicating suppression of AOM metabolism. Similarly, in the ethanol-fed rats the levels of the DNA adducts O6-methylguanine and 7-methylguanine 24 h after AOM injection were reduced in the colonic mucosa to 14 +/- 7% and 61 +/- 11% of controls and in the liver to 80 +/- 9% and 86 +/- 6 of controls. By contrast, rats changed from the ethanol diet to no-ethanol diet for 12 h prior to the dose of [14C]AOM metabolized the carcinogen at a faster rate than controls, indicating loss of suppression with cessation of ethanol intake along with induction of metabolizing enzymes; DNA adduct levels were reduced in the colonic mucosa to 90 +/- 13% and 76 +/- 9% of controls and in the liver to 81 +/- 6% and 85 +/- 3% of controls. Our findings indicate that dietary ethanol during the preinduction and induction phase of the AOM model dramatically inhibits tumorigenesis, even with high dosage of carcinogen, and suggest that: (a) inhibition of tumorigenesis may result from suppression of metabolic activation of AOM and the consequent reduced formation of DNA adducts during the induction (initiation) phase of the model; (b) these anti-initiation effects of ethanol are unrelated to the epidemiological association between consumption of alcoholic beverages and large bowel cancer; and (c) mechanisms of action of agents found to modulate carcinogenesis in experimental models should be determined before the results can be generalized to human beings.

Epidemiological studies have shown an association between consumption of alcoholic beverages, particularly beer, and carcinoma of the large bowel, especially the rectum. We studied the effects of chronic dietary beer and ethanol consumption on experimental colonic carcinogenesis, fecal bile acid and neutral sterol levels, fecal bacterial flora, and colonic epithelial DNA synthesis. Ten-week-old male Fischer 344 rats were pair fed throughout the study with Lieber-DeCarli-type liquid diets providing comparable total carbohydrates, proteins, fats, and calories. The diets provided 23 or 12% of calories as alcohol in beer (Hi-Beer and Lo-Beer groups), 18 or 9% of calories as reagent ethanol (Hi-EtOH and Lo-EtOH groups), or no alcohol (control group). After 3 weeks of dietary acclimatization, 10 weekly s.c. injections of the bowel carcinogen azoxymethane, 7 mg/kg, were given (weeks 1-10). At necropsy in week 26, the high alcohol groups (Hi-Beer and Hi-EtOH) showed a significantly reduced incidence of tumors in the right colon (42 and 46% versus 81% in control, P less than 0.01 and P = 0.02) but no effect on left colonic tumorigenesis. By contrast, the low alcohol groups (Lo-Beer and Lo-EtOH) showed a trend toward increased incidence and proportion of tumors in the left colon (incidence of 42 and 35% versus 15% in control, P = 0.06 for Lo-Beer; 28 and 30% of tumors in left colon versus 11%, P = 0.08 and P = 0.07) but no effect on right colonic tumorigenesis. Numbers of right colonic tumors were inversely correlated with alcohol consumption of all rats (r = -0.350, P less than 0.001), but left colonic tumors were not correlated. Fecal bile acid and neutral sterol levels, fecal bacterial counts, and colonic epithelial DNA synthesis did not correlate with the effects of alcohol consumption on colonic tumorigenesis. Our findings suggest that: modulation of experimental colonic tumorigenesis by chronic dietary beer and ethanol consumption was due to alcohol rather than other beverage constituents; tumorigenesis in the right and left colon was affected differentially by the levels of alcohol consumption, reflecting complex interactions among the potential mechanisms for alcohol effects in the model used.