Search

A hypothesis has been presented which is compatible with the epidemio- logical data and which is testable. Initial studies derived from the hypothesis have been described and the results are at least compatible with the hypothesis and to date we have no results that are not explicable on the theory that bacteria, in particular Clostridium paraputrificum. produce carcinogens or cocarcinogens from bile acids. If the hypothesis is true, a major implication is that cancer of the colon is readily preventable by modification of the diet and this has been discussed. If the hypothesis is not true at least we are learning a lot about the metabolic activities of the gut flora, the effect of diet on the flora and on fecal loss of bile acids. Am.J. Clin.Nutr. 27: 1475-1480,1974. Colon cancer is one of the major causes of death in North America and North West Europe (and the colon is the commonest site of cancer in the United States) but is relatively rare in Africa, Asia and most of South America. The epidemiology of colon cancer has been investi- gated intensively and it is now generally accepted that a major contributing factor in the etiology of these tumors is diet. There is little agreement concerning the component of the diet responsible, but recent studies of the incidence of colon cancer in 37 countries indicate that it is highly correlated with the intake of bound fat and animal protein (1), while in a controlled case of Japanese immi- grants in Hawaii there was a strong association with beef eating (2). Because the search for preformed carcino- gens in the diet has been unsuccessful in explaining the geographical variations in mci- dence, we postulated (3) that the carcinogen is formed in situ in the colon by the gut bacterial flora from some benign substrate derived from the diet or from some gut secretion produced in response to the diet. In this presentation I have attempted to describe the development of this hypothesis and the evidence in its favor. I have then outlined some of its implications, in particular in the field of prevention of colon cancer by dietary change.

1. 1. A strain of Eschehchia coli has been shown to produce oestradiol from 4androsten-3,17-dione. 2. 2. From the same substrate a strain of Clostridium paraputrificum produced 17-methoxy-1,3,5(10)-estratriene-3-ol, the identity of which was deduced from ultra violet and fluorescence spectra, the mass spectrum of the product and its trifluoracetyl derivative, and by comparison with the products of methylation of oestradiol. 3. 3. The conditions required for the reaction have been investigated.

1. 1. A3-oxo-5β-steroid Δ4-dehydrogenase (EC 1.3.99.5) has been demonstrated in a strain of Clostridium paraputrificum. The enzyme had a pH optimum of 7–8, was only active on 5β steroids and was inactive on steroids with a 3,7-dioxo-5β structure. The presence of a 12-oxo group or the configuration at C-17 had no effect on enzyme activity. 2. 2. A3-oxo-5β-steroid Δ1-dehydrogenase (EC 1.3.99.4) was also produced by the same strain. It had a pH optimum between 8 and 9, and was only active on 5β steroids. 3. 3. Both enzymes were able to use the artificial electron acceptor phenazine methosulphate as well as O2 as the terminal electron acceptor.

The geographical variations in the incidence of carcinoma of the colon seem to be correlated with variation in the fat content of the diet. It was postulated that intestinal bacteria might be able to produce carcinogens from dietary fats or from bile steroids, and that the variations in the incidence of colon cancer might depend partly on differences in the composition of the intestinal bacterial flora brought about by the differences in diet. Samples of faces from six areas were examined for their content of bacteria and steroids. Faces from people in Britain and the U.S.A., where the incidence of colon cancer is high, had higher counts of Bacteroides and lower counts of enterococci and other aerobic bacteria than faeces from people in Uganda, South India, or Japan, where the incidence of the disease is low. Faeces from people in the " Western " countries contained higher concentrations of steroids than those from the African and Eastern countries, and the steroids were also more degraded. The results are consonant with the thesis that the intestinal bacteria may be aetiologically related to cancer of the colon.

The ability of bacteria to catalyse the nitrosation of secondary amines has been investigated. It has been shown that this may be of importance in people with urinary tract infections living in areas where the concentration of nitrate in the drinking water is high.

The consumption of a chemically defined diet altered both the faecal flora and faecal steroids. Enterobacteria increased while enterococci and other lactic acid bacteria decreased in number. No other obvious change occurred. However, the alteration in the degree of degradation of the faecal neutral steroids may reflect some alteration in the anaerobic flora. Total excretion of both acid and neutral steroids was also reduced. On cessation of the diet both faecal flora and faecal steroids reverted to levels before the diet.

The faecal urease-activity of uraemic patients increased in proportion to their blood-urea concentration. The increase was the result of production of urease by a greater proportion of strains of anaerobic bacteria; neither the number of bacteria, the species, nor the mean urease activity per active strain differed from normal. Increased bacterial urease in the colon may fortuitously prevent enzyme availability becoming a limiting factor in the recycling of urea nitrogen by uraemic individuals.

1. 1. Cholanylglycine hydrolase has been studied in cell fractions from two faecal strains of each of the genera: Enterococcus, Clostridium, Bifidobacterium and Bacteroides. 2. 2. Intracellular and extracellular fractions with bile deconjugating ability have been characterized for substrate specificity, product inhibition, pH optimum and for the effect of a range of inhibitors. 3. 3. All preparations showed similar properties in their response to inhibitors and in their pH optimum of 5–6. All were more stable under reducing conditions. We suggest that anaerobic conditions are necessary in vivo for optimal production and stability of cholanylglycine hydrolase from faecal bacteria.

The geographical variations in the incidence of breast cancer seem to be correlated with variation in the fat content of the diet. There is also strong evidence for a direct role for œstrogens in the aetiology of breast cancer. We have already postulated that gut bacteria can produce œstrogens from the biliary steroids present in the colon. Since the amount of biliary steroid found in faeces is correlated with the amount of fat in the diet, this could explain the relation between the amount of dietary fat and the incidence of breast cancer. Gut bacteria have been shown to be able to produce steroidal oestrogens from bile-acids and cholesterol derivatives in vitro.