A tetranucleotide repeat polymorphism in CYP19 and breast cancer risk

The CYP19 gene codes for aromatase, a key steroidogenic enzyme involved in the conversion of androgens to estrogens, A tetranucleotide (TTTA) repeat p olymorphism is present in intron 4 of CYP19; 2 out of 4 breast cancer case- control studies have reported a greater frequency of 2 specific alleles amo ng affected women. We evaluated associations between CYP19 repeat alleles a nd breast cancer risk in a case-control study nested within the Nurses' Hea lth Study cohort (incident cases: n=462; controls: n=618). We observed seve n different CYP19 alletes (TTTA(7-13)). Compared to controls, cases had a s tatistically significant greater frequency of the 10 (TTTA)(10) repeat alle le (10 allele: 2.3% vs. 0.7%, p = 0.005) and a nonsignificant increase in t he frequency of the 12 (TTTA)(12) allele (12 allele: 3.1% vs. 2.1%, p = 0.1 1). A higher frequency of the 10 allele was observed in more advanced cance r cases defined as four or more involved nodes or distant metastasis [4+ no des: 5/36 (13.9%) vs. 0-3 nodes: 13/330 (3.9%), p = 0.02]. Among controls, we found women with the 7 repeat allele to have decreased levels of estrone sulfate (-16.4%, p = 0.02), estrone (-6.1%, p = 0.22) and estradiol (-9.9% , p = 0.10), and a lower estrone/androstenedione ratio (EI/A) (-10.5%, p = 0.08) compared to non-carriers. A higher EI/A ratio and elevated estrogen l evels were observed among carriers of the 8 repeat allele; EI/A ratio (+21. 0%, p = 0.003), estrone (+7.5%, p = 0.16) and estradiol (+10.8%, p = 0.08). However, we observed no evidence of an association between these alleles a nd breast cancer risk. We were unable to make inferences regarding the effe ct of the 10 allele on hormone levels due to the small number of allele car riers in the subgroup with hormone revels. As this repeat polymorphism is n ot close to the splice sites in intron 4, linkage disequilibrium with other functional polymorphisms in CYP19 may explain the findings of an increased association between breast cancer and the 10 allele variant of CYP19. We d id not detect any sequence variants in the regulatory region or in the adip ose-specific exon 1.4. The lack of an established effect an CYP19 function associated with the 10 allele means that these findings should be interpret ed with caution. Int. J. Cancer 87:204-210, 2000. (C) 2000 Wiley-Liss. Inc.