The main hypotheses of the COS project were:
- That a diet rich in fruit and vegetable, and in antioxidant micronutrients, is more protective in women with genetic predisposition than in women with wild type BRCA genes, because such a diet may protect DNA from oxidative damage. This hypothesis is largely falsified by the C.O.S. results on total and specific fruits and on antioxidant vitamins. A possible limitation of our results is that the widespread recommendation of increasing the consumption of fruit and vegetables to prevent cancer may have induced patients belonging to families with several cancer cases to change their diet and increase fruit and vegetables by a greater extent than patients belonging to families without cancer cases.
- That cruciferous vegetables are more protective in presence of BRCA mutations, because of their high content in indolyl glucosinolates, which upon hydrolysis yield indole-3-carbinol and its dimer diindolylmetane. These substances reduce estrogenic activity by favouring the 2-hydroxilation pathway with respect to the more estrogenic 16-hydroxilation act as androgen antagonists, help detoxifying carcinogenic xenobiotics through the activation of phase-1 and phase-2 enzymes, and upregulate BRCA1 expression. This hypothesis has been fairly strongly corroborated by our results.
- That phytoestrogens are more protective in BRCA mutation carriers. Phytoestrogens are plant polyphenols with week estrogenic activity that may reduce estrogenic stimuli by preventing the link of endogenous estrogens to estrogen receptors (Magee and Rowland, 2002). They are present in soy products (mainly isoflavonoids) and in several other seeds (mainly lignans). Another possible preventive mechanism is that soy isoflavonoids may upregulate BRCA1 expression Upregulation of BRCA expression in women with a germline mutation is important because one functional allele is still present in every cell. We observed a lower interaction OR for soy products consumption, but the proportion of consumers was too low for any firm conclusion.. The hypothesis, therefore, is neither corroborated nor falsified.
- That elevated waist to hip ratio, a marker of insulin resistance, is associated with increasing risk in mutation carriers. This observation was not confirmed by the C.O.S. study, but there was a suggestion that adult obesity, contrary to the general observations in premenopausal women, may be associated with increasing risk in carriers of mutated BRCA genes.
- That foods that increase insulin and IGF-I levels increase breast cancer risk by a greater extent in presence of mutated BRCA genes. Highly refined carbohydrates may stimulate insulin production by causing rapid increases in glycaemia, and saturated fat by increasing insulin resistance. Milk, on the contrary, despite its low glicemic index, directly stimulate insulin production or release. There is evidence that it is the whey fraction of milk protein, which is the fraction that remains in milk serum after cheese production, that contains the predominating insulin secretagogue, and that fermented milk may reduce the insulinemic index of refined cereals. These observations may explain why C.O.S. did not show any interaction with butter, cream, cheese, and even suggested a protective effect of yogurt. Milk consumption, on the other hand, has been consistently found associated with higher plasma levels of IGF-I. High levels of IGF-I may actually be a major determinant of the early occurrence of breast cancer in BRCA mutation carriers, and studies to evaluate this biomarker are urgently needed. C.O.S. results strongly corroborated the hypothesis of an interaction with milk. Results were also consistent, without reaching statistical significance, with an interaction with simple sugars and with the food group including cakes, pies, cookies and biscuits. The latter interaction, however, might be due to the presence of milk in the recipes.
- That tobacco smoking might be protective in mutation carriers, because tobacco may reduce estrogenic stimulation by shifting oestrogen metabolism from 16 to 2-hydroxilation (Bradlow 1996). Our study definitely falsified this hypothesis.
