Previous reports have suggested an association between circulating IGFBP-3 levels and the risk of premenopausal breast cancer, and a single nucleotide polymorphism (SNP) in the promoter region of IGFBP-3 (nucleotide -202) was shown to influence transcription. There is prior evidence that the action of antiproliferative agents such as retinoids and selective estrogen receptor modulators (SERMs) act in part by upregulating IGFBP3 gene expression. We identified 677 women with incident breast cancer and 834 matched controls from the Nurses' Health Study (NHS) and genotyped them at the -202 locus. For 943 of these women, we had previously measured IGF-I and IGFBP-3 plasma levels, and for 861 of these subjects, plasma retinol levels were available. Age-adjusted mean circulating IGFBP-3 levels were highest in the individuals with the AA genotype and decreased significantly in a stepwise manner in the presence of 1 or 2 copies of the C allele (4,426 ng/ml, 4,060 ng/ml and 3,697 ng/ml, respectively). We found a positive relation between age-adjusted IGFBP-3 levels and plasma retinol (14% difference in IGFBP-3 in top vs. bottom tertiles of retinol, p for trend < 0.001; Spearman correlation coefficient r = 0.25), which was similar across genotypes at the -202 IGFBP3 locus (interaction term, F = 0.10, p = 0.91). Breast cancer risk was not significantly related to genotype at the -202 locus in our prospective analyses. We confirmed a relation between the -202 IGFBP3 polymorphism and IGFBP-3 serum levels and observed a positive correlation between circulating retinol levels and circulating IGFBP-3 levels, providing further evidence that retinoids may influence IGF physiology. Our data do not demonstrate a significant influence of this locus on breast cancer risk, but we cannot exclude a minor influence or an influence confined to subgroups.
Copyright 2003 Wiley-Liss, Inc.