Molecular genetics improves the management of hereditary non-polyposis colorectal cancer.

BACKGROUND: The syndrome of hereditary non-polyposis colorectal cancer (HNPCC) can be diagnosed fairly accurately using clinical criteria and a family history. Identifying HNPCC helps to prevent large-bowel cancer, or allows cancer to be treated at an early stage. Once the syndrome has been diagnosed a family member's risk can be judged approximately from a family tree, or it can now be predicted accurately if the causative mutation is known. OBJECTIVE: This study involved attempts to improve the management of a family with HNPCC over a period of 10 years. Clinical diagnostic criteria, colonoscopic surveillance, surgical treatment, genetic counselling, molecular genetic research, and finally predictive genetic testing were applied as they evolved during this time. SUBJECTS AND METHODS: A rural general practitioner first noted inherited large-bowel cancer in the family and began screening subjects as they presented, using rigid sigmoidoscopy at the local hospital. At the time that the disorder was recognised as being HNPCC (1987), screening by means of colonoscopy at our university hospital was aimed primarily at first-degree relatives of affected individuals. After realising how many were at risk, screening was brought closer to the family. A team of clinicians and researchers visited the local hospital to identify and counsel those at risk and to perform screening colonoscopy. Family members were recruited for research to find the gene and its mutation that causes the disease, to develop an accurate predictive test and to reduce the number of subjects undergoing surveillance colonoscopies. RESULTS: There are approximately 500 individuals in this family. In the 10 years of this study the number of subjects who have been counselled for increased genetic risk or who have requested colonoscopic surveillance for HNPCC in this kindred has increased from 20 to 140. After the causative mutation was found in the hMLH1 gene on chromosome 3, a test for it has reduced the number of subjects who need screening colonoscopy by over 70%. A protocol has been devised to inform family members, to acquire material for research in order to provide genetic counselling for (pre-test and post-test) risk, and to test for the mutation. Eventually, identifying those with the mutation should focus surveillance accurately. CONCLUSIONS: The benefits of restricting screening to subjects with the mutation that causes colorectal cancer and of performing operations to prevent cancer are hard to measure accurately. However, it is likely that at least half the family members will be able to avoid colonoscopic screening, some deaths from cancer should be prevented, and the cost of preventing and treating cancer in the family should fall substantially.

Histopathology and gene expression changes in rat liver during feeding of fumonisin B1, a carcinogenic mycotoxin produced by Fusarium moniliforme.

Fumonisin B1 (FB1) is a carcinogenic mycotoxin produced by the fungus Fusarium moniliforme in corn. Feeding of FB1 to rats causes acute liver injury, chronic liver injury progressing to cirrhosis, and sometimes terminates in hepatocellular carcinoma or cholangiocarcinoma. This study describes the histolopathology and changes in gene expression in the rat liver during short-term feeding of FB1. Male Fischer rats were fed either FB1 250 mg/kg or control diet, and were killed weekly for 5 weeks. FB1 caused a predominantly zone 3 'toxic' liver injury, with hepatocyte death due to necrosis and apoptosis. Hepatocyte injury and death were mirrored by hepatic stellate cell proliferation and marked fibrosis, with progressive disturbance of architecture and formation of regenerative nodules. Despite ongoing hepatocyte mitotic activity, oval cell proliferation was noted from week 2, glutathione S-transferase pi-positive hepatic foci and nodules developed and, at later time points, oval cells were noted inside some of the 'atypical' nodules. Northern blot (mRNA) analysis of liver specimens from weeks 3 to 5 showed a progressive increase in gene expression for alpha-fetoprotein, hepatocyte growth factor, transforming growth factor alpha (TGF-alpha) and especially TGF-beta1 and c-myc. Immunostaining with LC(1-30) antibody demonstrated a progressive increase in expression of mature TGF-beta1 protein by hepatocytes over the 5 week feeding period. The overexpression of TGF-beta1 may be causally related to the prominent apoptosis and fibrosis seen with FB1-induced liver injury. Increased expression of c-myc may be involved in the cancer promoting effects of FB1.

The effects of dietary iron overload on fumonisin B1-induced cancer promotion in the rat liver.

The present study was performed to determine whether excess hepatic iron modulates the cancer-initiating and promoting properties of FB1. Thirty-eight male F344 rats were divided into four dietary treatment groups: (i) control diet (AIN, n = 8); (ii) FB1 250 mg/kg diet (FB1, n = 10); (iii) 1-2% carbonyl iron (CI, n = 10); or (iv) FB1 plus iron loading (FB1/CI, n = 10) for 5 weeks (2 x 2 factorial design). Hepatic iron concentrations in iron-loaded animals at 5 weeks were 444 +/- 56 (CI) and 479 +/- 80 micromol/g dry weight (FB1/CI) (mean +/- SEM). All the FB1-fed rats, in the presence or absence of CI, developed a toxic hepatitis with a 4-fold rise in serum alanine transaminase (ALT) levels. FB1 appeared to augment iron-induced hepatic lipid peroxidation, as measured by the generation of thiobarbituric acid reacting substances (TBARS) in liver homogenates (P < 0.0001). Morphometric analysis showed that FB1 caused a significantly greater mean +/- SEM number of 'enzyme-altered' foci and nodules per cm2 (5.34 +/- 1.42 vs. 1.50 +/- 0.52, P < 0.05), as well as a greater area (%) of liver occupied by foci and nodules (0.33 +/- 0.12% vs. 0.05 +/- 0.03%, P < 0.001), compared with FB1/CI. The addition of FB1 to dietary iron loading caused a shift in distribution of iron from hepatocytes to Kupffer cells, probably due to phagocytosis of necrotic iron-loaded hepatocytes. In conclusion, (i) FB1 appears to cause toxicity in the liver independently from effects on lipid peroxidation; (ii) FB1 has a potentiating effect on iron-induced lipid peroxidation; and (iii) dietary iron loading appears to protect against the cancer promoting properties of FB1, possibly due to a stimulatory effect of iron on hepatocyte regeneration.

A modified method of continuous venous anastomosis in microsurgery.

The continuous suture technique for end-to-end vascular anastomosis is cautioned against because of the risk of vessel constriction. A modified method of continuous suture for end-to-end venous microanastomosis is presented in which vessel constriction does not occur. This technique was compared with the conventional interrupted suture technique in the rat femoral vein, with each rat serving as its own control. Forty-eight Long-Evans rats were used. The mean time taken to complete the anastomosis was 9.8 minutes (range, 8-14 minutes) for the modified continuous technique and 17.7 minutes (range, 14-24 minutes) for the conventional interrupted technique (p < 0.05, independent t-test). In addition, the veins were examined under the microscope for patency and the milk test was performed on each anastomosis 30 minutes postanastomosis, and 1 week and 1 month postoperatively. Two groups of rats were sacrificed, one at 1 week and one at 1 month, and the two different anastomoses were compared using vessel morphometry in 40 rats and corrosion casts in 8 rats. All veins were patent postoperatively, as well as at 1 week and 1 month postoperatively. Vessel morphometry confirmed a similar luminal surface area in all veins examined at 1 week and 1 month. A two-way analysis of variance of vessel morphometry indicated no significant interaction between the methods used and the postoperative time (p = 0.60). The modified continuous technique is twice as quick as the conventional interrupted technique for end-to-end microvenous anastomosis and does not lead to vessel constriction.

Operative management of papillary cystic neoplasms of the pancreas.

BACKGROUND: Papillary cystic neoplasm (PCN) is a rare malignant tumor of the pancreas that typically occurs in young females and has an excellent prognosis. STUDY DESIGN: We report a retrospective review of 12 patients treated during a 16-year period. Pre-, intra-, and postoperative data were evaluated in all patients to determine optimal management with specific reference to surgical strategy. RESULTS: All 12 tumors occurred in young women (mean age 22 years, range 14-36 years). Six patients presented with an epigastric mass, and three with severe abdominal pain. The correct diagnosis was made preoperatively in only five patients. Incorrect diagnoses included hepatoma, pancreatic pseudocyst, and hydatid cyst. The PCNs had a mean diameter of 12.5 cm (range 8-20 cm), and occurred in the head (four), neck (three), body (three), and tail (two) of the pancreas. All were resected. Operations performed were pylorus-preserving pancreaticoduodenectomy (three), central pancreatectomy with pancreaticogastrostomy (three), distal pancreatectomy (three), and local resection (three). In one patient two liver metastases were resected in addition to the pancreatic primary. One patient presented with tumor rupture and a major bleed into the lesser sac and died of multiple organ failure after resection. Postoperative complications included a stricture at the hepaticojejunostomy after pancreaticoduodenectomy, which resolved after temporary stenting, and a pancreatic duct fistula after local tumor resection, which required a distal pancreatectomy. Eleven patients are well at followup (mean 6.6 years; range 6 months to 15 years). CONCLUSIONS: PCN should be considered in the differential diagnosis of large pancreatic masses, especially in young females. Conservative resection, where technically feasible, is safe and effective and represents the therapy of choice.

Immunohistochemical detection of gastric mucin in normal and disease states.

At least seven human mucin genes have been described, which express glycoproteins MUC1-7 in various tissues. It has been shown that different mucins are expressed in various gastric disease states compared to the normal. In this study we used histochemical and immunohistochemical methods to determine the type and pattern of mucin in 54 patients with a variety of gastric conditions [i.e., normal controls, fetal stomachs, gastritis, low-grade dysplasia, intestinal metaplasia (associated with gastritis, benign ulcers, dysplasia, and cancer), early and advanced intestinal type adenocarcinoma, and diffuse adenocarcinoma]. We report for the first time the use of all seven MUC antibodies in the various conditions. Normal controls were immunoreactive for MUC4, 5, and 6 , and gastritis specimens showed similar results, although the latter showed more MUC1 immunoreactivity. Whereas early fetal stomach showed no MUC immunoreactivity, MUC4, 5, and 6 were present from the early second trimester onwards. There was no significant difference between dysplasia and intestinal metaplasia, both categories showing the presence of MUC2 and 3 predominantly. Early intestinal type adenocarcinomas did not show any mucins in the majority of cases. Advanced intestinal type adenocarcinomas showed immunoreactivity predominantly for MUC1, 5, and 6, as well as MUC2 in some cases. Diffuse adenocarcinomas showed strong positive MUC2 and 6 staining, and in some cases MUC5 and 7. In conclusion, we have shown different patterns of mucin immunoreactivity in various gastric disease states. Specimens with dysplasia, intestinal metaplasia, late intestinal type adenocarcinoma, and diffuse gastric cancer were characterized by increased diversity of mucin types, whereas early intestinal cancer showed loss of mucin immunoreactivity.

Comparison of the significance of three histopathological thresholds of malignancy in experimental colorectal tumours.

The purpose of this experiment was to study sequential histogenesis of colonic epithelial tumours in the dimethylhydrazine (DMH) induced rat colon cancer model. Seventy outbred female Wistar rats treated with DMH (40 mg/kg, subcutaneously weekly for five weeks) were killed and autopsied in batches of 10 every five weeks from the 10th to 40th weeks from first treatment. The resulting 378 colonic lesions were assigned to benign or malignant categories using each of three standard histopathological thresholds of malignancy: alpha, the transition from dysplasia to intraepithelial carcinoma; beta, invasion through the crypt basement membrane; and gamma, invasion through the muscularis mucosae. These comprised 79 'benign' and 299 'malignant' or 273 'benign' and 141 'malignant' lesions depending on the threshold (alpha or gamma) assigned (p less than 0.001). Decreasing ratios of pre-threshold to post-threshold lesions between 15 and 40 weeks (alpha, 2.0 to 0.051; beta, 3.5 to 0.57; gamma, 8.0 to 0.87) provide some support for an 'adenoma-carcinoma' progression for each. Comparison of time-dependent prevalence curves confirms that the alpha threshold (cyto-architecture) is qualitatively different from the beta and gamma thresholds (invasion), showing that the adenoma-carcinoma and de novo hypotheses need not be mutually exclusive. The time-dependent prevalence data support de novo origin of some carcinomas, as well as at least two other modes of accrual for neoplastic lesions.

The relationship between faecal bile acids and the development of experimental colon cancer.

Human metabolic studies have suggested a positive association between dietary intake, faecal bile acid excretion and the development of colon cancer. Similar investigations in experimental models of this disease have also implicated faecal bile acids but in both animals and man the results remain equivocal. This study sequentially examines the outputs and concentrations of faecal bile acids in dimethylhydrazine (DMH) treated groups of rats (n = 10), killed at 5-weekly intervals from the 10th to 40th week following the first injection. The sequential results showed that both the output and concentration of faecal bile acids decreased with time and accompanied an increase in both the incidence and numbers of colonic neoplasms over the 40 weeks of the study. When the animals were grouped according to the histological classification of their tumours, the faecal bile acids did not differ between animals with and without tumours. Further, when the latter group were sub-divided into those with adenomas alone and those with carcinomas, faecal bile acid outputs and concentrations did not differ between them, nor when they were compared with the tumour-free animals. The results of this study do not support a role for faecal bile acids in the dimethylhydrazine-induced model of experimental colon cancer in rats.

Changes in serum lipids related to the presence of experimental colon cancer.

People at risk from coronary heart disease and large bowel cancer are drawn from the same urbanised, industrialised Western populations. Whilst changes in blood lipids are well recognised in heart disease, little is known of their role in large bowel cancer. This study investigates serial alterations in blood lipids in the 1,2-dimethylhydrazine (DMH) rat model of colon cancer. Eighty Wistar rats received a 5 weekly regimen of DMH. At week 10, and at 5 weekly intervals until week 40, random groups of 10 rats were killed and blood taken for total and free cholesterol, phospholipids, triglycerides and liver enzymes. All colonic neoplasms were histologically classified either as adenomas or carcinomas with groups being allocated into tumour-free (n = 16) or tumour-bearing (n = 54), the latter group being further sub-divided into animals with adenoma alone (n = 8) and those with carcinoma (n = 46). Results were considered both sequentially and according to tumour status. Sequential results showed that with increase in colonic neoplasms with time there were accompanying increases in free and % free cholesterol and in phospholipids (P less than 0.001). There were no changes in total cholesterol, triglycerides or liver enzymes. Results according to tumour status showed that whilst there was no difference in total cholesterol or triglycerides between tumour-free and tumour-bearing rats, there was a significant increase in free (P less than 0.01) and % free cholesterol (P less than 0.001) and a decrease in phospholipids in the tumour-bearing animals (P less than 0.001). There was no difference in any serum lipid between tumour-free and adenoma-bearing rats. In animals with carcinoma, while there was no difference in total cholesterol or triglycerides, there was an increase in free (P less than 0.005) and % free cholesterol (P less than 0.001) and a decrease in phospholipids (P less than 0.001) compared to tumour-free rats. The data show for the first time a clear relationship between blood lipids and the presence or absence of large bowel cancer.

Extramedullary haemopoiesis in the small bowel.

A patient with myelofibrosis developed repeated gastrointestinal haemorrhage from the small intestine, which was found to be infiltrated with extramedullary haemopoiesis. Nineteen months later he presented with subacute intestinal obstruction; radiology and laparotomy documented progressive infiltration of the small bowel. Histological examination of the resected terminal ileum showed patchy mucosal ulceration, with underlying submucosal and serosal extramedullary haemopoiesis.

Enteroendocrine cell hyperplasia, carcinoid tumours and adenocarcinoma in long-standing ulcerative colitis.

Chronic ulcerative colitis may be accompanied by a variety of epithelial changes, including loss of goblet cells, Paneth cell metaplasia, villous metaplasia, and dysplasia. Total colitis is also accompanied by an increased incidence of adenocarcinoma. All these changes are assumed to be secondary to repeated mucosal damage, but how they develop is unknown. Little attention has been paid to the enteroendocrine cell population, despite the postulated role of these cells as producers of trophic hormones. We describe two patients with long-standing ulcerative colitis who developed both adenocarcinoma and carcinoid tumours. In both, there were increased numbers of enteroendocrine cells in the uninvolved colonic mucosa. We suggest that an increased enteroendocrine cell mass may be part of a non-specific reaction to chronic mucosal injury, and by producing an elevated level of trophic hormones may act as a promoter in the development of neoplasia.

Induction of nephroblastoma-like renal tumours in rats by a single dose of 1,2-dimethylhydrazine.

A single injection of 1,2-dimethylhydrazine (200 mg/kg s.c.) resulted in the development of renal tumours in 70 per cent of treated rats within 1 year. The tumours were frequently bilateral (36 per cent) and grew to a large size (up to 100 g) before causing death. Lung metastases were found in one case. Histologically, the majority of the tumours (73 per cent) exhibited the classical triphasic morphological features of nephroblastoma (Wilms' tumour) with varying proportions of epithelial (glomeruloid and tubular structures), stromal (fibrous tissue, smooth muscle, cartilage) and blastematous elements. This pathological resemblance to human nephroblastoma suggests the possibility of using dimethylhydrazine-induced renal tumours as a model of Wilms' tumour.

An investigation into the mechanism of co-carcinogenesis of dietary cholesterol during the induction of colon cancer in rats by 1,2 dimethylhydrazine.

The mechanism by which dietary cholesterol facilities colon carcinogenesis was investigated in the dimethylhydrazine-induced rat colon cancer model. Fifty female Wistar rats received a standard course of dimethylhydrazine (DMH) injections (40 mg/kg/week subcutaneously for ten weeks) while being fed Vivonex, a cholesterol-free elemental diet. Animals were allocated to one of five dietary regimens. One control group received Vivonex with added cholesterol (10 mg/100 ml Vivonex/rat/day) throughout the experiment, while another group received Vivonex alone. The remaining three groups received added cholesterol exclusively before, during or after the ten week DMH induction period. The experiment continued for over 500 days, and was evaluated by comparing, between groups, the time taken for the development of objective signs of colonic disease (time to tumour presentation or TTP). Animals either died spontaneously or were killed and autopsied. Colon cancers were confirmed histologically in every animal. The results showed that cholesterol feeding throughout the experiment or during the DMH induction period reduced the TTP compared to controls (p less than 0.05). Cholesterol prefeeding had no such effect. In the after group, the TTP was correspondingly delayed (p less than 0.05). Cholesterol-fed controls and groups receiving cholesterol during or after the DMH induction had more colon tumours and/or a greater incidence of metastases than cholesterol-free controls or those pre-fed cholesterol. The findings indicate a direct relationship between timing of cholesterol exposure and signs of colon cancer, and demonstrate that dietary cholesterol has promoter-like characteristics.

Combination chemotherapy with 5-fluorouracil (5FU) and 1,3-bis(2-chloro-ethyl)-1-nitrosourea (BCNU) prolongs survival of rats with dimethylhydrazine-induced colon cancer.

The effects of combination chemotherapy with 5FU and BCNU on rats with dimethylhydrazine (DMH)-induced colon cancer were investigated in a long term survival study. Eighty Wistar rats received a colon cancer producing regimen on DMH (40 mg/kg body weight/week, subcutaneously for 10 weeks). After presenting with signs of colonic disease, all rats underwent diagnostic laparotomy and colonoscopy when colon tumours were located, measured and the extent of the disease staged. Only animals with tumours (n = 63) were included and allocated to one of three tumour stages. Stage A (n = 17), had colonic tumours without serosal involvement; stage B (n = 28) had serosal involvement without metastases; stage C (n = 18) had serosal involvement with lymphadenopathy and/or metastases. Each group was randomly allocated into two subgroups, one serving as untreated controls while the other received 5FU (300 mg/m2 weekly intragastrically for life) together with BCNU (40 mg/m2 intraperitoneally on days 0, 42 and 84). The effect of chemotherapy on tumour growth was measured sequentially by colonoscopy. Animals were observed until death and necropsied, when colon carcinoma was histologically confirmed and survival analysed. The results indicate that chemotherapy significantly prolongs the survival of rats with the least advanced disease (stage A) but was of no benefit to rats with locally advanced or metastatic disease (stages B and C). Furthermore, chemotherapy was associated with a significant reduction in tumour size. Survival analyses in untreated animals show that the laparotomy staging system adopted provides accurate prognostic information. This study shows that DMH-induced colon tumours are chemosensitive, and suggests that this animal model may be a valuable testing ground for new chemotherapeutic agents.

Dietary cholesterol deprivation improves survival and reduces incidence of metastatic colon cancer in dimethylhydrazine-pretreated rats.

Cholesterol feeding of rats with colon cancer induced by dimethylhydrazine results in reduced survival and an increased incidence of metastatic colon cancer. As cholesterol may be implicated in the induction or maintenance of the metastatic process, an experiment was designed to determine whether rats with colon cancer would benefit from the removal of cholesterol from the diet. Female Wistar rats were treated with a colon cancer-inducing regimen of dimethylhydrazine (40 mg/kg/week for 10 weeks) while being fed on a standard cholesterol-containing rat pellet diet. After two rats had died spontaneously of histologically proven adenocarcinoma of the colon at 24 weeks, the remaining rats were randomly allocated in groups of 15 to one of three dietary regimens. Group S continued to receive standard pellet diet, group V were fed on Vivonex alone and group VC were fed Vivonex plus cholesterol (10 mg/100 ml Vivonex). Each group was assessed for survival and incidence of histologically proven metastatic disease. There were no differences in either survival or incidence of metastases when groups S and VC were compared. In the cholesterol deprived group V, however, there was a significant increase in survival compared with groups S and VC (p less than 0.02) and this was due to a significant reduction in the incidence of metastases (p less than 0.05). Cholesterol deprivation therefore benefits rats with established colon cancer induced by dimethylhydrazine by improving survival and reducing the incidence of metastases.

Subtotal colectomy in the dimethylhydrazine-treated rat. A surgical model of colorectal cancer.

The effects of subtotal colectomy on the survival of rats with established colon cancer induced by 1,2-dimethylhydrazine (DMH) have been investigated in an attempt to assess whether it might provide a suitable model of surgical treatment of this disease. 40 female Wistar rats received a regimen of DMH injections (40 mg/kg body weight s.c. every 14 days for 10 weeks) known to produce colon cancer. An additional 10 rats received no DMH, serving as controls (group 1). After presenting with signs of colonic disease at 25 weeks, all DMH-treated rats had diagnostic colonoscopy under general anaesthesia, only those with visible neoplasms (n = 34) being included in the study. These were randomised into two groups: group 2 (n = 13) animals were unoperated controls while group 3 (n = 21) animals had a therapeutic subtotal colectomy with histological confirmation of cancer in the resected colon. The animals were observed until death, the postoperative survival and cause of death at necropsy being compared between groups. The results showed that overall survival (p less than 0.013) and survival from death due to colon cancer (p less than 0.001) were significantly increased in the colectomised group 3 animals compared to unoperated controls (group 2). While 91% of the unoperated controls died of colon cancer, only 8% of the colectomised group died of this cause (p less than 0.001), the remainder dying from unrelated causes, predominantly DMH-induced primary extracolonic cancers. Subtotal colectomy in rats with DMH-induced colon cancer reduces mortality from this disease, providing a suitable model of surgical treatment. However, the high incidence of DMH-induced extracolonic cancers may make the model unsuitable for studies of adjuvant therapy.

Experimental evidence against the bile salt theory of colon carcinogenesis.

According to the bile salt theory of colon carcinogenesis, therapeutic agents which increase the delivery of bile salts to the large intestine may promote colon cancer. The possibility that aluminium hydroxide (Aludrox), a bile salt binding agent, might facilitate colon carcinogenesis in vivo was tested experimentally, using the dimethylhydrazine (DMH)-induced rat colon cancer model. 48 Wistar rats were randomly allocated to 1 of 3 groups, all fed the same standard diet. Two groups received a course of ten weekly DMH injections. One was allowed fresh drinking water ad libitum whilst the other received Aludrox in their drinking water. A third group received weekly saline injections plus Aludrox in their drinking water. After 1 year's observation, there were no significant differences between the groups of DMH-injected rats given drinking water with or without Aludrox in respect of survival, necropsy incidence of primary or metastatic colon cancer, or in the total number of colon tumours per group. The results provide reassurance that Aludrox does not promote colon cancer and tend to contradict the bile salt theory of colon carcinogenesis.

Experimental colon carcinogenesis is facilitated by endogenous factors in the intestinal contents.

The theory that endogenous factors in the intestinal contents may be pathogenic during large bowel carcinogenesis was tested in the dimethylhydrazine (DMH)-induced rat colon cancer model. Thirty female Wistar rats, each serving as their own control, had a surgical transection of the proximal colon with reanastomosis to the rectum, thereby excluding part of the colon from faecal contact. All rats then received a course of DMH (40 mg/kg body wt/wk s.c. for 10 weeks) while fed on Vivonex. This diet was selected because it lacks known exogenous (dietary) cocarcinogens. It also produces mucosal atrophy in functioning (proximal) colon, to parallel the disuse atrophy induced in the defunctioned (distal) colon. Animals remained on the diet throughout the experiment and were killed when moribund or at 40 weeks. At necropsy, the anatomical distribution, number, size and histological type of colon tumours were compared between functioning and defunctioned colonic segments within the same animal. At autopsy, there were significantly fewer colon tumours in the defunctioned segment (P less than 0.005). Furthermore, there were significantly fewer carcinomas (P less than 0.005) and fewer tumours greater than 1 cm diameter (P less than 0.01) in this segment. The data indicate that endogenous factors in the intestinal contents facilitate chemically-induced colon carcinogenesis. Luminal nutrition may be implicated.

The effects of cholic acid and bile salt binding agents on 1,2-dimethylhydrazine-induced colon carcinogenesis in the rat.

The theory that bile salts may be colon tumour promoters was tested in the dimethylhydrazine (DMH)-induced rat colon cancer model. Fifty Wistar rats were randomly allocated to one of five experimental group (n = 10), all fed the same standard diet. One group served as saline-injected controls, while the other four groups received DMH (20 mg/kg body weight/rat/week s.c.) for 20 weeks. In addition, each of the DMH-injected groups concurrently received 20 weekly i.g. instillations of one of the following: cholic acid (a bile acid); cholestyramine or aluminium hydroxide (both bile acid binding agents), or water. After a years observation period, all the controls were alive and tumour-free, while all the DMH-injected rats had died of histologically proven colon cancer. Irrespective of the type of gastric instillate, there were no significant differences between the groups in terms of time to tumour presentation, survival, in the necropsy incidence of primary or metastatic colon cancer, or in the numbers of colon tumours per group. The data suggest that bile salts and bile salt binding agents are not colon tumour promoters in the rat. The bile salt theory of colon carcinogenesis may need reappraisal.