Fiber and Breast Cancer: Another Piece of the Puzzle--But Still an Incomplete Picture

Open Access
Publisher Website
Google Scholar
Abstract
Cancer is a multifactorial disease in which the factors directly or inversely associated with it (foods and nutrients, for instance) are not only numerous but also complex. It is difficult to demonstrate a simple chain of necessary causes of cancer, and a whole array of studies is required before a factor can be defini- tively labeled as a risk factor or as being protective in nature. This "epidemiological mosaic" (/) is now relatively well codi- fied: Ecological studies generate an hypothesis, which, if backed up with a plausible biological mechanism, is then accepted by the scientific community. Analytical epidemiology aims at es- tablishing etiological links through case-control or prospective studies. Such studies are always disputable and must be repeated under various conditions in order to enlarge the mass of frag- mentary evidence. The hypothesis is tested in animal models so as to reproduce human findings in a simpler design and to pro- vide a better insight into the mechanisms involved. In some cases, in vitro studies and/or molecular biology are part of the process. Once the accumulated evidence concerning an iden- tified protective factor appears convincing enough, an interven- tion study is designed. Most pieces of the puzzle concerning the fiber-breast cancer case (reviewed in (2)) are already in place. So far, however, it has been difficult to fit the pieces for the proposed biological mechanism into the puzzle. According to the hypothesis proposed, the effect of fibers in the bowel would be to reduce the enterohepatic circulation of estrogens (3) via two pathways: 1) a trapping effect exerted by fiber on unconjugated estrogens (the quantity of the fat intake seems to modulate this effect) and 2) a change in the bowel flora characterized by the presence of bacteria with low P-glucuron- idase activity. The latter effect would result in increased fecal excretion, reduced urinary excretion, and lower plasma levels of estrogens, since only unconjugated estrogens can re-enter the circulation. The aim of the study reported by Cohen et al. (4) in this issue of the Journal was to complete the part of the puzzle concerned with the biological mechanism. Cohen et al. designed a scrupu- lously careful protocol to address the hypothesis that tumor in- hibition is the direct consequence of the ability of dietary fiber to alter the enterohepatic circulation of the estrogens. These in- vestigators compared the inhibitory effect of a mixture of a soluble fiber (psyllium) and an insoluble fiber (wheat bran) on the incidence of chemically induced mammary tumors in rats. They assessed P-D-glucuronidase activity in the cecum and es- trogen profiles in the blood, urine, and feces. With the knowledge that psyllium is the best suppressor of P-D-glucuro- nidase activity and that wheat bran is an efficient trapper of un- conjugated estrogens, Cohen et al. observed a dose-response effect by mixing the fibers in different proportions and giving these mixtures to different experimental groups. An equal com- bination of psyllium and wheat bran produced the strongest tumor-inhibiting effect, protecting 33% of the animals at risk for all types of tumor (benign, in situ, or adenocarcinom as) and 70% of the animals at risk for adenocarcinoma. It was better than psyllium alone (17% and 57%, respectively). Wheat bran alone had the weakest inhibiting effect on adenocarcinom as (47%), and weight gain was highest in this group of rats. The difference between the impact of fibers on all types of tumor and that on adenocarcinom as might be caused by the effect of fibers on the estrogens acting at the clonal expansion stage. So far the pieces fit, and we obtain the additional information that both types of fiber are necessary, with the soluble fiber having a slight advantage. In human studies on the association of fiber intake and breast cancer risk demonstrating the protec- tive effect of fiber, it is difficult to discriminate between the ef- fects of soluble and insoluble fibers, since the two are highly correlated. In a study involving 451 breast cancer case patients

This publication has 0 references indexed in Scilit: