JAMA: Journal of the American Medical Association 268(12): 1573-1577, 1992.
The review deals first with the epidemiologic model of the fat/fiber hypothesis, briefly covering international correlation studies, metabolism of bile acids, some case-control studies of diet and colon cancer, and the fiber hypothesis of Burkitt (the author quotes a precursor to this concept dating from 1585). The 'high-fat' diet is really a diet high in meat, fat, protein and total energy, while fiber, along with cereals and fruit, is not the most active anticarcinogenic dietary component. Meat/vegetables would be a better name for the hypothesis; other aspects of the diet such calcium (which may operate both to bind fatty and bile acids, and to reduce cell proliferation) and alcohol intake are independently related to colon cancer risk. At the second level of physiological/biochemical models, the 'bile acid/volatile fatty acid' and other hypotheses are discussed. Increased understanding of fermentation in the colon has underlined the importance of volatile fatty acid production and its anticarcinogenic effects, but this itself cannot be considered in a vacuum. Carcinogens produced by cooking at the higher temperatures prevailing in cooking protein (e.g., meat) with fat, stimulation of bile acid production by dietary fat, protective factors, micronutrients such as carotenoids and ascorbic acid, and non-nutrients like flavonoids, phenols, isothiocyanates, and indoles in vegetables, and the action of endocrine factors and neuropeptides on the gut, all add complexity yet also complement one another at different levels. Links between diet, genetic susceptibility and DNA-damaging compounds are illustrated by the arylamines, where genetic polymorphism for high rates of both acetylation and N-oxidation confer an almost three-fold colon cancer risk. Aberrant crypts and microadenomas and the genetic changes associated with progression, in particular the deletion site (DCC) on chromosome 18q which codes for a cell-adhesion molecule that may be important in the accumulation of cells to form an adenoma, are discussed. Also discussed are the FAP gene and increased expression of methyl transferase (implications for DNA hypomethylation), and diacylglycerol (DAG), an important part of the ras activation cascade, that is produced from fat, may mediate in colon carcinogenesis. Excess intraluminal DAG from the diet could mimic endogenous DAG, amplifying cell proliferation signals; this is potentially the first indication of how diet may operate at the molecular level. (66 Refs.)
Rheinische Friedrich- Wilhelms- Universität Bonn